CN116472556B - Prediction method, encoder, decoder, and storage medium - Google Patents

Abstract

The embodiment of the present application discloses a prediction method, an encoder, a decoder, and a storage medium, the method comprising: determining the geometric information of a point cloud to be encoded/decoded; determining a set of repeated points based on the geometric information of the point cloud to be encoded/decoded; wherein the set of repeated points includes at least two points in the point cloud to be encoded/decoded; sorting the attribute information of the repeated points in the set of repeated points to determine the order of the repeated points; and predicting the attributes of the repeated points based on the order of the repeated points to determine the attribute prediction results of the repeated points. In this way, by sorting the attribute information of the repeated points, repeated points with continuous attribute information are obtained, and the prediction reference points are determined using the sorted order of the repeated points, which is conducive to reducing the output code stream and improving the encoding and decoding performance.

Description

Prediction method, encoder, decoder and storage medium

Technical Field

The embodiments of the present application relate to the field of coding and decoding technology, and in particular to a prediction method, an encoder, a decoder, and a storage medium.

Background Art

In the geometry-based point cloud compression (G-PCC) encoder framework, the geometric information of the point cloud and the attribute information corresponding to each point cloud are encoded separately. After the geometry encoding is completed, the geometry information will be reconstructed, and the encoding of the attribute information will depend on the reconstructed geometry information. Among them, the attribute information encoding is for the encoding of color and reflectivity, generating an attribute code stream.

However, in the current related technologies, when predicting attributes, since there are sequences containing repeated points, that is, the Morton order or Hilbert order of the repeated points is equal (the geometric coordinates are equal), when predicting the repeated points, the prediction reference points will be selected according to the original point cloud order, but the selected prediction reference points are not the best prediction reference points, which affects the prediction performance of the repeated points.

Summary of the invention

The embodiments of the present application provide a prediction method, an encoder, a decoder, and a storage medium, which use the sorted repeated point sequence to determine the prediction reference point, which is beneficial to reducing the output bit stream and improving the encoding and decoding performance.

The technical solution of the embodiment of the present application can be implemented as follows:

In a first aspect, an embodiment of the present application provides a prediction method, which is applied to an encoder, and the method includes:

Determine the geometric information of the point cloud to be encoded;

Determine a set of repeated points based on the geometric information of the point cloud to be encoded; wherein the set of repeated points includes at least two points in the point cloud to be encoded;

Sorting the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

Attribute prediction is performed on the repeated points based on the repeated point sequence to determine the attribute prediction results of the repeated points.

In a second aspect, an embodiment of the present application provides a prediction method, which is applied to a decoder, and the method includes:

Parse the bitstream to determine the geometric information of the point cloud to be decoded;

Determine a set of repeated points based on the geometric information of the point cloud to be decoded; wherein the set of repeated points includes at least two points in the point cloud to be decoded;

Sorting the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

Attribute prediction is performed on the repeated points based on the repeated point sequence to determine the attribute prediction results of the repeated points.

In a third aspect, an embodiment of the present application provides an encoder, the encoder comprising:

A first determination unit is configured to determine geometric information of a point cloud to be encoded; based on the geometric information of the point cloud to be encoded, determine a set of repeated points; wherein the set of repeated points includes at least two points in the point cloud to be encoded;

A first sorting unit is configured to sort the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

The first prediction unit is configured to perform attribute prediction on the repeated points based on the repeated point sequence, and determine the attribute prediction result of the repeated points.

In a fourth aspect, an embodiment of the present application provides an encoder, the encoder comprising:

A second determination unit is configured to parse the bitstream to determine the reconstructed geometric information of the point cloud to be decoded; based on the geometric information of the point cloud to be decoded, determine a set of repeated points; wherein the set of repeated points includes at least two points in the point cloud to be decoded;

A second sorting unit is configured to sort the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

The second prediction unit is configured to perform attribute prediction on the repeated points based on the repeated point sequence, and determine the attribute prediction result of the repeated points.

In a fifth aspect, an embodiment of the present application further provides an encoder, comprising: a first memory and a first processor; the first memory stores a computer program that can be run on the first processor, and the first processor implements the prediction method of the encoder when executing the program.

In a sixth aspect, an embodiment of the present application further provides a decoder, comprising: a second memory and a second processor; the second memory stores a computer program that can be run on the second processor, and the second processor implements the prediction method of the decoder when executing the program.

In the seventh aspect, an embodiment of the present application provides a storage medium, comprising: a computer program stored thereon, which, when executed by a first processor, implements the prediction method of the encoder; or, when executed by a second processor, implements the prediction method of the decoder.

The embodiment of the present application provides a prediction method, an encoder, a decoder, and a storage medium, the method comprising: determining the geometric information of a point cloud to be encoded/decoded; determining a set of repeated points based on the geometric information of the point cloud to be encoded/decoded; wherein the set of repeated points includes at least two points in the point cloud to be encoded/decoded; sorting the attribute information of the repeated points in the set of repeated points to determine the order of the repeated points; and predicting the attributes of the repeated points based on the order of the repeated points to determine the attribute prediction results of the repeated points. In this way, by sorting the attribute information of the repeated points, repeated points with continuous attribute information are obtained, and the prediction reference points are determined using the sorted order of the repeated points, which is conducive to reducing the output code stream and improving the encoding and decoding performance.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flowchart of an exemplary encoding process provided by an embodiment of the present application;

FIG2 is a flowchart of an exemplary decoding process provided by an embodiment of the present application;

FIG3 is a schematic diagram of a first flow chart of a prediction method provided in an embodiment of the present application;

FIG4 is a schematic diagram of a second flow chart of the prediction method provided in an embodiment of the present application;

FIG5 is a schematic diagram of a spatial relationship between a current node and neighboring nodes provided by the related art;

FIG6 is a schematic diagram of a Morton code relationship between neighboring nodes coplanar with the current node and encoded and decoded within a preset neighbor range provided by the related art;

FIG7 is a schematic diagram of a Morton code relationship between neighboring nodes that are co-linear with the current node and have been encoded and decoded within a preset neighbor range provided by the related art;

FIG8 is a schematic diagram of a third flow chart of the prediction method provided in an embodiment of the present application;

FIG9 is a schematic diagram of a first component structure of an encoder in an embodiment of the present application;

FIG10 is a schematic diagram of a second component structure of an encoder in an embodiment of the present application;

FIG11 is a schematic diagram of a first component structure of a decoder in an embodiment of the present application;

FIG. 12 is a schematic diagram of a second component structure of a decoder in an embodiment of the present application.

DETAILED DESCRIPTION

In order to enable a more detailed understanding of the features and technical contents of the embodiments of the present application, the implementation of the embodiments of the present application is described in detail below in conjunction with the accompanying drawings. The attached drawings are for reference only and are not used to limit the embodiments of the present application.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

It should be pointed out that the terms "first\second\third" involved in the embodiments of the present application are merely to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second\third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of the present application described here can be implemented in an order other than that illustrated or described here.

Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application are explained. The nouns and terms involved in the embodiments of the present application are applicable to the following explanations: Point Cloud Compression (PCC), Point Cloud Reference Model (PCRM), Morton code (Morton code, Morton), bounding box, octree, Red-Green-Blue (RGB), Luminance-Chrominance (YUV), Look Up Table (LUT), Peak Signal to Noise Ratio (PSNR), Audio Video Coding Standard (AVS).

Point Cloud refers to a collection of massive three-dimensional points. The points in the point cloud may include the location information of the points and the attribute information of the points. For example, the location information of the points may be the three-dimensional coordinate information of the points. The location information of the points may also be referred to as the geometric information of the points. For example, the attribute information of the points may include color information and/or reflectivity, etc. For example, the color information may be information on any color space. For example, the color information may be RGB information. Among them, R represents red (Red, R), G represents green (Green, G), and B represents blue (Blue, B). For another example, the color information may be luminance and chrominance (YcbCr, YUV) information. Among them, Y represents brightness (Luma), Cb (U) represents blue color difference, and Cr (V) represents red color difference.

For a point cloud obtained according to the principle of laser measurement, the points in the point cloud may include the three-dimensional coordinate information of the points and the laser reflection intensity (reflectance) of the points. For another example, for a point cloud obtained according to the principle of photogrammetry, the points in the point cloud may include the three-dimensional coordinate information of the points and the color information of the points. For another example, a point cloud obtained by combining the principles of laser measurement and photogrammetry may include the three-dimensional coordinate information of the points, the laser reflection intensity (reflectance) of the points, and the color information of the points.

Point clouds can be divided into the following categories according to the way they are obtained:

The first type of static point cloud: the object is stationary, and the device that obtains the point cloud is also stationary;

The second type of dynamic point cloud: the object is moving, but the device that obtains the point cloud is stationary;

The third type of dynamic point cloud acquisition: the device that acquires the point cloud is moving.

For example, point clouds can be divided into two categories according to their usage:

Category 1: Machine perception point cloud, which can be used in autonomous navigation systems, real-time inspection systems, geographic information systems, visual sorting robots, disaster relief robots, etc.

Category 2: Point cloud perceived by the human eye, which can be used in point cloud application scenarios such as digital cultural heritage, free viewpoint broadcasting, 3D immersive communication, and 3D immersive interaction.

Since point clouds are a collection of massive points, storing point clouds not only consumes a lot of memory, but is also not conducive to transmission. There is also not enough bandwidth to support direct transmission of point clouds at the network layer without compression. Therefore, point clouds need to be compressed.

Up to now, the point cloud coding framework that can compress point clouds can be the geometry-based point cloud compression (G-PCC) codec framework or the video-based point cloud compression (V-PCC) codec framework provided by the Moving Picture Experts Group (MPEG), or the AVS-PCC codec framework provided by the Audio Video Standard (AVS). The G-PCC codec framework can be used to compress the first type of static point clouds and the third type of dynamically acquired point clouds, and the V-PCC codec framework can be used to compress the second type of dynamic point clouds. The G-PCC codec framework is also called the point cloud codec TMC13, and the V-PCC codec framework is also called the point cloud codec TMC2.

It can be understood that in the point cloud G-PCC encoder framework, the input point cloud is sliced and then the slices are independently encoded.

FIG1 is an exemplary encoding flow chart provided by an embodiment of the present application. As shown in FIG1, in the encoding process of the point cloud, the geometric information and the attribute information corresponding to each point cloud are encoded separately. In the encoding process of the geometric information, the coordinates of the geometric position are transformed so that all the point clouds are contained in a bounding box, and then quantized. This step of quantization mainly plays a role in scaling. Due to the quantization rounding, the geometric information of a part of the point cloud is the same, so it is determined based on the parameters whether to remove duplicate points. The process of quantization and removal of duplicate points is also called voxelization. Then the bounding box is divided into octrees. In the octree-based geometric position encoding process, the bounding box is divided into 8 sub-cubes, and the non-empty (containing the points in the point cloud) sub-cubes are divided into 8 equal parts until the leaf node obtained by the division is a 1x1x1 unit cube. The division is stopped, and the points in the leaf node are entropy encoded to generate a geometric code stream.

In the attribute encoding process, the geometric information is reconstructed. At present, attribute encoding is mainly performed on color and reflectivity information. First, determine whether to perform color space conversion. If color space conversion is performed, the color information is converted from RGB color space to YUV color space. Then, the reconstructed point cloud is recolored using the original point cloud so that the unencoded attribute information corresponds to the reconstructed geometric information. In color information encoding, after sorting the point cloud using Morton code, the post-term differential prediction is directly performed, and finally the prediction residual is quantized and encoded to generate the attribute code stream.

The high-level syntax elements, the geometric code stream and the attribute code stream constitute a binary code stream of the three-dimensional image, and the encoder sends the binary code stream to the decoder.

FIG2 is a flowchart of an exemplary decoding process provided by an embodiment of the present application; as shown in FIG2, the decoder obtains a binary code stream, and independently decodes the geometric code stream and the attribute code stream in the binary code stream. When decoding the geometric code stream, the geometric information of the point cloud is obtained through entropy decoding-octree reconstruction-inverse coordinate quantization-inverse coordinate translation; when decoding the attribute code stream, the attribute information of the point cloud is obtained through entropy decoding-inverse quantization-attribute reconstruction-inverse space conversion, and the three-dimensional image model of the point cloud data is restored based on the geometric information and attribute information.

A point cloud encoding and decoding method provided in an embodiment of the present application mainly acts on the attribute prediction part and the attribute reconstruction part surrounded by the dotted boxes in Figures 1 and 2.

Based on the above-introduced background, the point cloud encoding method and point cloud decoding method provided in the embodiments of the present application are respectively introduced below.

The present application embodiment provides a point cloud encoding method, which is applied to an encoder. FIG3 is a schematic diagram of a first flow chart of a prediction method provided by the present application embodiment. As shown in FIG3 , the method may include:

Step 101: Determine the geometric information of the point cloud to be encoded;

Here, the geometric information specifically refers to the three-dimensional geometric coordinates of the point.

In some embodiments, the geometric information is reconstructed geometric information obtained during the encoding process. After the geometric encoding is completed, the geometric information is reconstructed to guide the attribute encoding.

Step 102: determining a set of repeated points based on the geometric information of the point cloud to be encoded; wherein the set of repeated points includes at least two points in the point cloud to be encoded;

Here, duplicate points refer to points with the same or similar positions in the point cloud. In practical applications, the distance between two points can be used to determine whether the two points coincide.

In some embodiments, determining a set of repeated points based on the geometric information of the point cloud to be encoded includes: based on the geometric information of the point cloud to be encoded, determining points in the point cloud to be encoded whose distance from a target point is less than a distance threshold to constitute the set of repeated points.

Here, the distance threshold is greater than or equal to 0. When the distance threshold is greater than 0, the repeated points include points with the same or similar geometric coordinates; when the distance threshold is equal to 0, the repeated points only include points with the same geometric coordinates.

In some embodiments, determining the set of repeated points based on the geometric information of the point cloud to be encoded includes: determining, based on the geometric information of the point cloud to be encoded, that points with the same geometric information constitute the set of repeated points. That is, when the distance threshold is equal to 0, determining that points with completely equal geometric information coordinates constitute the set of repeated points.

In some embodiments, determining a set of repeated points based on the geometric information of the point cloud to be encoded includes: generating a Morton code or a Hilbert code of the point cloud to be encoded based on the geometric information; and determining that points with equal Morton codes or equal Hilbert codes constitute the set of repeated points.

That is to say, when determining repeated points, they can be determined directly by geometric coordinates, or the geometric coordinates can be converted into Morton codes or Hilbert codes to determine that points with equal Morton codes or equal Hilbert codes constitute a set of repeated points.

It should be noted that the point cloud to be encoded includes at least one repeated point set, each repeated point set corresponds to the same geometric information, and the prediction method of each repeated point set is the same.

Step 103: sorting the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

In the embodiment of the present application, attribute information and attribute value are two different expressions of the same encoded data.

In some embodiments, the sorting of attribute information of the duplicate points in the duplicate point set to determine the order of the duplicate points includes: sorting the duplicate points in the duplicate point set in order of attribute information (i.e., attribute values) from small to large to determine the order of the duplicate points; or sorting the duplicate points in the duplicate point set in order of attribute information from large to small to determine the order of the duplicate points.

It should be noted that the existing repeated points contain multiple attribute values, and these attribute values are distributed discontinuously according to the order of the original point cloud. When selecting the prediction reference point, if the attribute values of the repeated points are discontinuous, it will affect the prediction performance of the repeated points. For example, the order of repeated points is: 0, 0, 25, 50, 0, 0, 50, 25, 50, 50. If the first N encoded repeated points are selected as prediction reference points for the current repeated point in this order, it is not easy to select the best prediction reference point. The present application sorts the attribute values of the repeated points according to the attribute values, and obtains the order of repeated points: 0, 0, 0, 0, 25, 25, 50, 50, 50. When the first N encoded repeated points are selected as prediction reference points for the current repeated point in this order, the best prediction reference point will be selected, thereby improving the coding performance.

Step 104: performing attribute prediction on the repeated points based on the repeated point sequence, and determining attribute prediction results of the repeated points.

Specifically, the predicted reference point of the current repeated point to be encoded is determined based on the repeated point sequence, the predicted reference point is the encoded repeated point, and the predicted attribute information of the current repeated point is determined according to the reconstructed attribute information of the predicted reference point, and the predicted attribute information is the prediction result of the current repeated point.

It should be noted that if there is only one prediction reference point, the reconstructed attribute information of the prediction reference point is directly used as the prediction attribute information; when there are at least two prediction reference points, the reconstructed attribute information can be averaged or weighted to obtain the reconstructed attribute information of the current repeated point.

In the point cloud coding method, after obtaining the predicted attribute information of the current repeated point, it also includes: making a residual of the real attribute information and the predicted attribute information, quantizing the residual, entropy coding and writing it into the bit stream.

By adopting the above technical solution, the attribute information of the repeated points is sorted to obtain repeated points with continuous attribute information, and the prediction reference points are determined using the sorted order of the repeated points, which is beneficial to reducing the output code stream and improving the encoding performance.

The following is a further example of a method for predicting duplicate points. FIG4 is a schematic diagram of a second flow chart of a prediction method provided in an embodiment of the present application. As shown in FIG4 , the method includes:

Step 201: when the current repeated point is the first repeated point in the repeated point sequence, based on the point cloud sequence of the point cloud to be encoded, determining the encoded neighboring points of the current repeated point; and using the neighboring points as prediction reference points for the current repeated point;

Here, the neighbor point includes: K neighbor points with the smallest distance to the current repeated point; wherein K is a positive integer. The value of K can be 3, 6, or even 8 or other values. In the embodiment of the present application, the value of K is specifically set according to the actual situation and is not limited here.

Here, the distance may be Euclidean distance, Manhattan distance, Earth Mover's Distance (EMD), etc. In order to reduce the amount of calculation and improve the coding efficiency, Manhattan distance is preferred.

It should be noted that for the first repeated point, since it is in the first place in the repeated point sequence, there is no coded repeated point in front of it, so it can only be regarded as a non-repeated point for prediction. Specifically, according to the point cloud sequence of the non-repeated point, the nearest K neighboring points are selected as prediction reference points.

Step 202: when the current repeated point is not the first repeated point in the repeated point sequence, determining a predicted reference point of the current repeated point based on the repeated point sequence;

In some embodiments, determining the predicted reference point of the current repeated point based on the repeated point sequence includes: determining the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer; and using the first N repeated points as the predicted reference point of the current repeated point.

Here, N can be an integer such as 1, 2, 3, 4, 5, etc. When N is 1, the reconstructed attribute information of the previous repeated point of the current repeated point is used for prediction. For example, the reconstructed attribute information of the previous repeated point is used as the predicted attribute information of the current repeated point.

When N is greater than 1, the reconstructed attribute information of the previous multiple repeated points of the current repeated point is used for prediction. For example, the reconstructed attribute information of the previous multiple repeated points is used as the predicted attribute information of the current repeated point. In other words, when N is greater than 1, starting from the N+1th repeated point, the previous N repeated points are used as prediction reference points. For the 2nd to Nth repeated points, the nearest neighbor point and the previous one or more repeated points can be used as prediction reference points at the same time.

In practical applications, the number of prediction reference points can be fixed or not. When the number of prediction reference points is fixed, the number of neighbor points will increase as the number of repeated points decreases, and the number of neighbor points will decrease as the number of repeated points increases. When the number of prediction reference points is not fixed, when the first three repeated points are predicted, the neighbor point only takes one nearest neighbor point, and the repeated points take all the previously encoded points, 0, 1, or 2.

In some embodiments, determining the predicted reference point of the current repeated point based on the repeated point sequence includes: determining the encoded neighbor points of the current repeated point based on the point cloud sequence of the point cloud to be encoded; determining the first M repeated points of the current repeated point based on the repeated point sequence; wherein M is a positive integer; and using the neighbor points and the first M repeated points as predicted reference points for the current repeated point.

In some embodiments, the point cloud reordering based on the geometric information of the point cloud to be encoded to determine the point cloud order of the point cloud to be encoded includes: when the point cloud to be encoded is a dense point cloud, Morton reordering based on the geometric information of the point cloud to be encoded to obtain a first Morton order; Morton reordering after adding an offset to the geometric information of the point cloud to be encoded to obtain a second Morton order;

Correspondingly, the point cloud order based on the point cloud to be encoded is used to determine the encoded neighbor points of the current repeated point, including: determining the first P neighbor points of the current repeated point from the first Morton order; determining the first Q neighbor points of the current repeated point from the second Morton order; determining the K neighbor points with the smallest distance to the current repeated point from the first P neighbor points and the first Q neighbor points; wherein P and Q are both positive integers, and K is less than the sum of P+Q.

In PCRM, P=Q=4, K=3.

In some embodiments, the reordering of the point cloud based on the reconstructed geometric information of the point cloud to be encoded to determine the point cloud order of the point cloud to be encoded includes: when the point cloud to be encoded is a sparse point cloud, performing Hilbert reordering based on the geometric information of the point cloud to be encoded to obtain a Hilbert order;

Correspondingly, the method of determining the encoded neighbor points of the current repeated point based on the point cloud order of the point cloud to be encoded includes: determining the first L neighbor points of the current repeated point from the Hilbert order; determining the K neighbor points with the smallest distance to the current repeated point from the first L neighbor points; wherein L is a positive integer and K is less than L.

In the Hilbert order, the K points closest to the current repeated point are searched among the first L (L can be the first maximum number of neighbors (maxNumOfNeighbours)) points of the current repeated point to be encoded. In some embodiments, maxNumOfNeighbours is 128, K is 3, and the distance calculation method is the Manhattan distance calculation method.

Step 203: Based on the reconstructed attribute information of the prediction reference point, perform attribute prediction on the current repeated point.

In some embodiments, the attribute prediction of the current repeated point is performed based on the reconstructed attribute information of the predicted reference point, including: when the number of the predicted reference points is greater than 1, weighted operation is performed on the reconstructed attribute information of the predicted reference point to obtain the predicted attribute information of the current repeated point.

In some embodiments, the method further includes: determining a weight value of each predicted reference point based on a distance between each predicted reference point and the current repeated point.

It should be noted that the closer the distance to the current repeated point is, the higher the accuracy of prediction is. Therefore, determining the weight value of the reference point according to the distance can improve the prediction performance. For example, the smaller the distance, the larger the weight value, and the larger the distance, the smaller the weight value.

In some embodiments, the method further includes: when the predicted reference point includes at least two repeated points, determining weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence.

It should be noted that the distances of all duplicate points in the duplicate point set are 0, so the weight value cannot be set using the distance. Therefore, the weight value can be set based on the distance between the selected duplicate point and the current duplicate point according to the order of the duplicate points. For example, the closer the point is to the current duplicate point, the greater the weight value is, and the farther the point is from the current duplicate point, the smaller the weight value is.

Exemplarily, each repeated point is weightedly predicted using the first N points, and the weight is related to the sorting position of the predicted reference point to the current repeated point, that is, the weight of the first N repeated points of the current repeated point is 1/N, the larger N is, the smaller the weight is, and the smaller N is, the larger the weight is.

The prediction of duplicate points is further explained in detail below.

(1) First, obtain a point cloud arranged according to the Morton code or Hilbert code. For the repeated point sets (point sets with the same geometric coordinates) _Pi {i=1, 2, ...}, arrange the attribute values in _Pi in ascending order.

(2) The attributes in the point cloud are predicted according to the existing method, and the prediction method for repeated points is as follows:

Prediction of repeated points _Pi : when i<=H (for example, H=1, the H value is variable in actual applications), the prediction method of _Pi is the same as the prediction method of non-repeated points; when H<i<=J, the predicted attribute value of _Pi is the average or weighted value of the attribute values of the first N repeated points and K neighboring points of _Pi ; when J<i<=I, the predicted attribute value of _Pi is the average or weighted value of the attribute values of the first N (for example, N=1, the N value is variable in actual applications) repeated points of _Pi .

(3) The residual between the true attribute value and the predicted attribute value of each point in the point cloud is quantized and encoded.

The following is a description of the distance for different attribute information.

When the attribute information is color, the attribute prediction method may specifically include the following:

When the point to be encoded is a non-repeated point, the Morton code is first used to find the spatial neighbor points of the current point, and then the attributes of the current point are predicted based on the found neighbor points.

Morton code can also be called z-order code because its encoding order follows the spatial z order. The specific method of calculating Morton code is described as follows. For each component of the three-dimensional coordinate (x, y, z) represented by a d-bit binary number, the representation of its three components is achieved by:

Among them, x _l ,y _l ,z _l ∈{0,1} are the binary values corresponding to the highest bit (l=1) to the lowest bit (l=d) of x, y, z respectively. The Morton code M is x, y, z starting from the highest bit, and then arranged in sequence from x _l ,y _l ,z _l to the lowest bit. The calculation formula of M is as follows:

Among them, m _l′ ∈{0,1} are the values of the highest bit (l′＝1) to the lowest bit (l′＝3d) of M. After obtaining the Morton code M of each point in the point cloud, the points in the point cloud are arranged in order from small to large Morton codes, and the weight w of each point is set to 1. Expressed in computer language, it is similar to the combination of z|(y<<1)|(x<<2).

It should be noted that the Morton code of the current node is first calculated, and then the Morton code of the reference point is located according to the Morton code of the current node. Here, the Morton code of the reference point is the minimum Morton code among the neighboring nodes of the current node. Exemplarily, if the three-dimensional coordinates of the current node are (x, y, z), then the three-dimensional coordinates of the reference point are (x-1, y-1, z-1), and the calculated Morton code is the minimum at this time.

After obtaining the reference point, the coordinate difference (offset) between the reference point and the N neighbor nodes to be searched can be calculated, and the N coordinate differences can be converted into N Morton codes, which are then stored in a preset lookup table.

The spatial range of the neighboring points is shown in Figure 5. The current point to be encoded is the middle gray marked block, and its corresponding preset neighborhood range is 3×3×3 in size; wherein X represents the horizontal coordinate axis, Y represents the vertical coordinate axis, and Z represents the z coordinate axis. Here, the number of neighbor nodes of the current node is 26 in total; among them, the number of neighbor nodes coplanar with the current node is 6, the number of neighbor nodes colinear with the current node is 12, and the number of neighbor nodes copointing with the current node is 8. It should be noted that the preset neighborhood range is not limited to 3×3×3 in size, but can also be 4×4×4 in size, 5×5×5 in size, 6×6×6 in size, or even larger than 6×6×6 in size, etc., and the embodiments of the present application do not make specific limitations.

In conjunction with the example of FIG. 5 , consider that the search range of neighbor nodes is a preset neighborhood range of 3×3×3 size of the current node to be encoded. First, the Morton code of the current point node is used to obtain the block with the smallest Morton code value in the neighborhood, and the block is used as the reference block. Then, the reference block is used to find the encoded and decoded neighbor nodes that are coplanar and colinear with the current point node to be encoded. The Morton code relationship between the neighbor nodes that are coplanar with the current node to be encoded and have been encoded/decoded (can be referred to as "encoded and decoded") within the preset neighborhood range is shown in FIG. 6, and the Morton code relationship between the neighbor nodes that are colinear with the current node within the preset neighborhood range and have been encoded and decoded is shown in FIG. 7. Here, in FIG. 6, the block filled with 7 represents the current node, and the blocks filled with 3, 5, and 6 are neighbor nodes that are coplanar with the current node. In FIG. 7, the block filled with 7 still represents the current node, and the blocks filled with 1, 2, and 4 are neighbor nodes that are colinear with the current node.

As shown in Figures 6 and 7, the six neighbor nodes filled with 1, 2, 3, 4, 5, and 6 must complete the attribute encoding and decoding before the current node, but in addition to these neighbor nodes, some collinear neighbor nodes may also complete the attribute encoding and decoding before the current node. However, in the relevant technology, only three coplanar and three collinear neighbor nodes as shown in Figures 6 and 7 are currently considered, which will miss some collinear neighbor nodes that have been encoded and decoded. Then, after the six coplanar or collinear neighbor nodes cannot be found, directly using the reconstructed attribute value of the previous encoded and decoded node whose Morton code is smaller than the current node as the predicted attribute value is inaccurate in space, because the Morton code has periodic jump points, even if the Morton codes are adjacent, but the spatial position cannot be guaranteed to be adjacent. In this way, for attribute prediction based on geometric spatial relationships, the spatial correlation of adjacent nodes is not fully utilized, thereby reducing the encoding and decoding efficiency of point clouds.

When the point to be encoded is a repeated point, the method also includes: performing Morton reordering based on the geometric information of the point cloud to be encoded to obtain a Morton order (i.e., a point cloud order); determining a set of repeated points according to the Morton order; sorting the attribute information of the repeated points in the set of repeated points to determine the order of the repeated points; and predicting the attributes of the repeated points based on the repeated point order to determine the attribute prediction results of the repeated points.

When the current repeated point is the first repeated point in the repeated point sequence, the encoded neighbor points of the current repeated point are determined based on the point cloud sequence of the point cloud to be encoded; the neighbor points are used as prediction reference points for the current repeated point; and based on the reconstructed attribute information of the predicted reference points, the attributes of the current repeated point are predicted.

When the current repeated point is not the first repeated point in the repeated point sequence, a predicted reference point of the current repeated point is determined based on the repeated point sequence; and attributes of the current repeated point are predicted based on the reconstructed attribute information of the predicted reference point.

Specifically, the method for determining the prediction reference point includes: based on the repeated point sequence, determining the first N repeated points of the current repeated point; wherein N is a positive integer; and using the first N repeated points as the prediction reference point of the current repeated point. Based on the point cloud sequence of the point cloud to be encoded, determining the encoded neighbor points of the current repeated point; based on the repeated point sequence, determining the first M repeated points of the current repeated point; wherein M is a positive integer; and using the neighbor points and the first M repeated points as the prediction reference points of the current repeated point.

When the attribute information is reflectivity, the attribute prediction method may specifically include the following:

There are two methods for sorting point clouds according to different types of sequences: Morton reordering and Hilbert reordering. Specifically, Morton reordering is used for single-frame dense point clouds, and Hilbert reordering is used for multi-frame sparse point clouds.

That is to say, when predicting the reflectivity, both repeated points and non-repeated points need to be reordered first to obtain the point cloud order (ie, the Morton order or the Hilbert order).

Morton reordering: Generate the corresponding Morton code according to the geometric coordinates of the point cloud. After reordering the Morton code to obtain Morton order 1 (i.e., the first Morton order), add a fixed value to all point coordinates to obtain new coordinates, use the new coordinates to generate the Morton code corresponding to the point cloud, and obtain Morton order 2 (i.e., the second Morton order) according to Morton order 1 decoding. Find the nearest neighbor point of the point to be decoded, select the first P points of the current point as candidates in Morton order 1, and select the first Q points of the current point as candidates in Morton order 2. Among the above P and Q points, calculate the Manhattan distance d from each point to the current point, and select the K decoded points with the smallest distance from these P+Q points as the predicted points of the current point. In PCRM, P=Q=4, K=3.

Hilbert reordering: Generate the corresponding Hilbert code according to the geometric coordinates of the point cloud, sort it to get the Hilbert order, and encode it according to the Hilbert order. In the Hilbert order, find the K points closest to the current point among the maximum number of neighbors (maxNumOfNeighbours) of the current point to be encoded, where maxNumOfNeighbours defaults to 128, K is 3, and the distance calculation method is Manhattan distance.

For these two prediction methods of reflectivity, prediction is performed according to geometric position, which is called "prediction method based on geometric position". If the residual generated by the prediction method based on geometric position is large, the prediction method based on attribute value can usually reduce the prediction residual and improve the coding efficiency.

The prediction method based on attribute values is divided into the following steps:

(1) Save the most recently encoded 32 different attribute prediction values in the candidate prediction value table;

(2) Select the point closest to the current point attribute and use its attribute value as the attribute prediction value of the current point;

(3) Encode the serial number of the selected point in the candidate prediction value table.

In PCRM, the prediction method based on geometric position is first used to predict the attributes, and the residual values obtained are counted, that is, the residual values greater than or equal to 3 are accumulated and their number is recorded. After encoding 512 points, the average value of the residual values greater than or equal to 3 is calculated and compared with the pre-set threshold. When the calculated residual average exceeds the threshold, switch to the prediction method based on attribute value. In PCRM, the lossless coding threshold is 50, and the lossy coding threshold is 25.

The prediction method provided in the embodiment of the present application is applied to the existing point cloud prediction. By sorting the attribute information of the repeated points, repeated points with continuous attribute information are obtained, and the prediction reference points are determined according to the order of the sorted repeated points, which is beneficial to reducing the output code stream and improving the encoding and decoding performance.

Methods for encoding attribute information of non-repeated points include the following:

Method 1: For the attribute information of the first point, directly encode, quantize, entropy encode and write it into the bitstream; for the attribute information of the second and subsequent points, refer to the attribute information of the previously encoded point, make predictions, quantize and entropy encode the residual and write it into the bitstream;

Method 2: For the attribute information of the first point, directly encode it, quantize it, entropy encode it and write it into the bitstream; for the attribute information of the second point, refer to the attribute information of the first encoded point, make a prediction, quantize the residual, entropy encode it and write it into the bitstream; for the third point and subsequent points, select the attribute information of a specified number of points in the previously encoded points to construct a predicted value, make a prediction, quantize the residual, entropy encode it and write it into the bitstream.

The present application embodiment further provides another prediction method, which is applied to a decoder. FIG8 is a schematic diagram of a third flow chart of the prediction method provided by the present application embodiment. As shown in FIG8 , the method includes:

Step 301: Parse the bitstream to determine the reconstructed geometric information of the point cloud to be decoded;

Here, after obtaining the code stream, the geometric code stream is parsed and the geometric information is predicted to obtain the reconstructed geometric information, and the reconstructed geometric information is used to guide the attribute decoding, and the geometric information specifically points to the three-dimensional geometric coordinates.

Specifically, the parsing process can perform entropy decoding and inverse quantization on the bitstream.

Step 302: Determine a set of repeated points based on the geometric information of the point cloud to be decoded; wherein the set of repeated points includes at least two points in the point cloud to be decoded;

Here, duplicate points refer to points with the same or similar positions in the point cloud. In practical applications, the distance between two points can be used to determine whether the two points coincide.

In some embodiments, determining a set of repeated points based on the geometric information of the point cloud to be decoded includes: based on the geometric information of the point cloud to be decoded, determining points in the point cloud to be decoded whose distance from the target point is less than a distance threshold to constitute the set of repeated points.

Here, the distance threshold is greater than or equal to 0. When the distance threshold is greater than 0, the repeated points include points with the same or similar geometric coordinates; when the distance threshold is equal to 0, the repeated points only include points with the same geometric coordinates.

In some embodiments, determining the repeated point set based on the geometric information of the point cloud to be decoded includes: determining, based on the geometric information of the point cloud to be decoded, that points with the same geometric information constitute the repeated point set. That is, when the distance threshold is equal to 0, determining that points with completely equal geometric information coordinates constitute the repeated point set.

In some embodiments, determining a set of repeated points based on the geometric information of the point cloud to be decoded includes: generating a Morton code or a Hilbert code of the point cloud to be decoded based on the geometric information; and determining that points with equal Morton codes or equal Hilbert codes constitute the set of repeated points.

That is to say, when determining repeated points, they can be determined directly by geometric coordinates, or the geometric coordinates can be converted into Morton codes or Hilbert codes to determine that points with equal Morton codes or equal Hilbert codes constitute a set of repeated points.

It should be noted that the point cloud to be decoded includes at least one repeated point set, each repeated point set corresponds to the same geometric information, and the prediction method of each repeated point set is the same.

Step 303: sorting the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

In the embodiment of the present application, attribute information and attribute value are two different expressions of the same decoded data.

In some embodiments, the sorting of attribute information of the duplicate points in the duplicate point set to determine the order of the duplicate points includes: sorting the duplicate points in the duplicate point set in order of attribute information (i.e., attribute values) from small to large to determine the order of the duplicate points; or sorting the duplicate points in the duplicate point set in order of attribute information from large to small to determine the order of the duplicate points.

It should be noted that the existing repeated points contain multiple attribute values, and these attribute values are distributed discontinuously according to the order of the original point cloud. When selecting the prediction reference point, if the attribute values of the repeated points are discontinuous, it will affect the prediction performance of the repeated points. For example, the order of repeated points is: 0, 0, 25, 50, 0, 0, 50, 25, 50, 50. If the first N decoded repeated points are selected as prediction reference points for the current repeated point in this order, it is not easy to select the best prediction reference point. The present application sorts the attribute values of the repeated points according to the attribute values, and obtains the order of repeated points: 0, 0, 0, 0, 25, 25, 50, 50, 50. When the first N decoded repeated points are selected as prediction reference points for the current repeated point in this order, the best prediction reference point will be selected, thereby improving the decoding performance.

Step 304: performing attribute prediction on the repeated points based on the repeated point sequence, and determining attribute prediction results of the repeated points.

Specifically, the predicted reference point of the current repeated point to be decoded is determined based on the repeated point sequence, the predicted reference point is the decoded repeated point, and the predicted attribute information of the current repeated point is determined according to the reconstructed attribute information of the predicted reference point, and the predicted attribute information is the prediction result of the current repeated point.

It should be noted that if there is only one prediction reference point, the reconstructed attribute information of the prediction reference point is directly used as the prediction attribute information; when there are at least two prediction reference points, the reconstructed attribute information can be averaged or weighted to obtain the reconstructed attribute information of the current repeated point.

By adopting the above technical solution, the attribute information of the repeated points is sorted to obtain the repeated points with continuous attribute information, and the prediction reference points are determined by using the sorted order of the repeated points to improve the decoding performance.

The specific prediction method for repeated points is shown in FIG4 , and the method includes:

Step 201: when the current repeated point is the first repeated point in the repeated point sequence, based on the point cloud sequence of the point cloud to be decoded, determine the decoded neighbor point of the current repeated point; and use the neighbor point as a prediction reference point for the current repeated point;

Here, the neighbor points include: K neighbor points with the shortest distance from the current repeated point; wherein K is a positive integer.

Here, the distance may be Euclidean distance, Manhattan distance, or Earth Mover's Distance (EMD). In order to reduce the amount of calculation and improve decoding efficiency, Manhattan distance is preferred.

It should be noted that for the first repeated point, since it is in the first place in the repeated point sequence, there is no decoded repeated point in front of it, so it can only be regarded as a non-repeated point for prediction. Specifically, according to the point cloud sequence of the non-repeated point, the nearest K neighboring points are selected as prediction reference points.

Step 202: when the current repeated point is not the first repeated point in the repeated point sequence, determining a predicted reference point of the current repeated point based on the repeated point sequence;

In some embodiments, determining the predicted reference point of the current repeated point based on the repeated point sequence includes: determining the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer; and using the first N repeated points as the predicted reference point of the current repeated point.

Here, N can be an integer such as 1, 2, 3, 4, 5, etc. When N is 1, the reconstructed attribute information of the previous repeated point of the current repeated point is used for prediction. For example, the reconstructed attribute information of the previous repeated point is used as the predicted attribute information of the current repeated point.

When N is greater than 1, the reconstructed attribute information of the previous multiple repeated points of the current repeated point is used for prediction. For example, the reconstructed attribute information of the previous multiple repeated points is used as the predicted attribute information of the current repeated point. In other words, when N is greater than 1, starting from the N+1th repeated point, the previous N repeated points are used as prediction reference points. For the 2nd to Nth repeated points, the nearest neighbor point and the previous one or more repeated points can be used as prediction reference points at the same time.

In practical applications, the number of prediction reference points can be fixed or not. When the number of prediction reference points is fixed, the number of neighbor points will increase as the number of repeated points decreases, and the number of neighbor points will decrease as the number of repeated points increases. When the number of prediction reference points is not fixed, when the first three repeated points are predicted, the neighbor point only takes one nearest neighbor point, and the repeated points take all previously decoded points, 0, 1, or 2.

In some embodiments, determining the predicted reference point of the current repeated point based on the repeated point sequence includes: determining the decoded neighbor points of the current repeated point based on the point cloud sequence of the point cloud to be decoded; determining the first M repeated points of the current repeated point based on the repeated point sequence; wherein M is a positive integer; and using the neighbor points and the first M repeated points as predicted reference points for the current repeated point.

In some embodiments, the point cloud reordering based on the geometric information of the point cloud to be decoded to determine the point cloud order of the point cloud to be decoded includes: when the point cloud to be decoded is a dense point cloud, Morton reordering based on the geometric information of the point cloud to be decoded to obtain a first Morton order; adding an offset to the geometric information of the point cloud to be decoded and performing Morton reordering to obtain a second Morton order;

Correspondingly, the decoded neighbor points of the current repeated point are determined based on the point cloud order of the point cloud to be decoded, including: determining the first P neighbor points of the current repeated point from the first Morton order; determining the first Q neighbor points of the current repeated point from the second Morton order; determining the K neighbor points with the smallest distance to the current repeated point from the first P neighbor points and the first Q neighbor points; wherein P and Q are both positive integers, and K is less than the sum of P+Q.

In PCRM, P=Q=4, K=3.

In some embodiments, the reordering of the point cloud based on the reconstructed geometric information of the point cloud to be decoded to determine the point cloud order of the point cloud to be decoded includes: when the point cloud to be decoded is a sparse point cloud, performing Hilbert reordering based on the geometric information of the point cloud to be decoded to obtain a Hilbert order;

Correspondingly, the decoded neighbor points of the current repeated point are determined based on the point cloud order of the point cloud to be decoded, including: determining the first L neighbor points of the current repeated point from the Hilbert order; determining the K neighbor points with the smallest distance to the current repeated point from the first L neighbor points; wherein L is a positive integer and K is less than L.

In the Hilbert order, the K points closest to the current repeated point are searched among the first L (L can be the first maximum number of neighbors (maxNumOfNeighbours)) points of the current repeated point to be decoded. In some embodiments, maxNumOfNeighbours is 128, K is 3, and the distance calculation method is the Manhattan distance calculation method.

Step 203: Based on the reconstructed attribute information of the predicted reference point, perform attribute prediction on the current repeated point.

In some embodiments, the attribute prediction of the current repeated point is performed based on the reconstructed attribute information of the predicted reference point, including: when the number of the predicted reference points is greater than 1, weighted operation is performed on the reconstructed attribute information of the predicted reference point to obtain the predicted attribute information of the current repeated point.

In some embodiments, the method further includes: determining a weight value of each predicted reference point based on a distance between each predicted reference point and the current repeated point.

It should be noted that the closer the distance to the current repeated point is, the higher the accuracy of prediction is. Therefore, determining the weight value of the reference point according to the distance can improve the prediction performance. For example, the smaller the distance, the larger the weight value, and the larger the distance, the smaller the weight value.

In some embodiments, the method further includes: when the predicted reference point includes at least two repeated points, determining weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence.

It should be noted that the distances of all duplicate points in the duplicate point set are 0, so the weight value cannot be set using the distance. Therefore, the weight value can be set based on the distance between the selected duplicate point and the current duplicate point according to the order of the duplicate points. For example, the closer the point is to the current duplicate point, the greater the weight value is, and the farther the point is from the current duplicate point, the smaller the weight value is.

Exemplarily, each repeated point is weightedly predicted using the first N points, and the weight is related to the sorting position of the predicted reference point to the current repeated point, that is, the weight of the first N repeated points of the current repeated point is 1/N, the larger N is, the smaller the weight is, and the smaller N is, the larger the weight is.

The prediction of duplicate points is further explained in detail below.

(1) First, obtain a point cloud arranged according to the Morton code or Hilbert code. For the repeated point sets (point sets with the same geometric coordinates) _Pi {i=1, 2, ...}, arrange the attribute values in _Pi in ascending order.

(2) The attributes in the point cloud are predicted according to the existing method, and the prediction method for repeated points is as follows:

Prediction of repeated points _Pi : when i<=H (for example, H=1, the H value is variable in actual applications), the prediction method of _Pi is the same as the prediction method of non-repeated points; when H<i<=J, the predicted attribute value of _Pi is the average or weighted value of the attribute values of the first N repeated points and K neighboring points of _Pi ; when J<i<=I, the predicted attribute value of _Pi is the average or weighted value of the attribute values of the first N (for example, N=1, the N value is variable in actual applications) repeated points of _Pi .

(3) The residual between the true attribute value and the predicted attribute value of each point in the point cloud is quantized and decoded.

The following is a description of the distance for different attribute information.

When the attribute information is color, the attribute prediction method may specifically include the following:

When the point to be decoded is a non-repeated point, the Morton code is first used to find the spatial neighbor points of the current point, and then the attributes of the current point are predicted based on the found neighbor points.

The spatial range of neighbor points is shown in Figure 5, the neighbor points coplanar with the current point are shown in Figure 6, and the neighbor points colinear with the current point are shown in Figure 7. In practical applications, neighbor points co-pointed with the current point can also be selected.

When the point to be decoded is a repeated point, the method also includes: performing Morton reordering based on the geometric information of the point cloud to be decoded to obtain a Morton order (i.e., a point cloud order); determining a set of repeated points according to the Morton order; sorting the attribute information of the repeated points in the set of repeated points to determine the order of the repeated points; and predicting the attributes of the repeated points based on the order of the repeated points to determine the attribute prediction results of the repeated points.

When the current repeated point is the first repeated point in the repeated point sequence, the decoded neighbor points of the current repeated point are determined based on the point cloud sequence of the point cloud to be decoded; the neighbor points are used as prediction reference points for the current repeated point; and attributes of the current repeated point are predicted based on the reconstructed attribute information of the predicted reference points.

When the current repeated point is not the first repeated point in the repeated point sequence, a predicted reference point of the current repeated point is determined based on the repeated point sequence; and attributes of the current repeated point are predicted based on the reconstructed attribute information of the predicted reference point.

Specifically, the method for determining the prediction reference point includes: based on the repeated point sequence, determining the first N repeated points of the current repeated point; wherein N is a positive integer; and using the first N repeated points as the prediction reference point of the current repeated point. Based on the point cloud sequence of the point cloud to be decoded, determining the decoded neighbor points of the current repeated point; based on the repeated point sequence, determining the first M repeated points of the current repeated point; wherein M is a positive integer; and using the neighbor points and the first M repeated points as the prediction reference points of the current repeated point.

When the attribute information is reflectivity, the attribute prediction method may specifically include the following:

There are two methods for sorting point clouds according to different types of sequences: Morton reordering and Hilbert reordering. Specifically, Morton reordering is used for single-frame dense point clouds, and Hilbert reordering is used for multi-frame sparse point clouds.

That is to say, when predicting the reflectivity, both repeated points and non-repeated points need to be reordered first to obtain the point cloud order (ie, the Morton order or the Hilbert order).

Morton reordering: Generate the corresponding Morton code according to the geometric coordinates of the point cloud. After reordering the Morton code to obtain Morton order 1 (i.e., the first Morton order), add a fixed value to all point coordinates to obtain new coordinates, use the new coordinates to generate the Morton code corresponding to the point cloud, and obtain Morton order 2 (i.e., the second Morton order) according to Morton order 1 decoding. Find the nearest neighbor point of the point to be decoded, select the first P points of the current point as candidates in Morton order 1, and select the first Q points of the current point as candidates in Morton order 2. Among the above P and Q points, calculate the Manhattan distance d from each point to the current point, and select the K decoded points with the smallest distance from these P+Q points as the predicted points of the current point. In PCRM, P=Q=4, K=3.

Hilbert reordering: Generate the corresponding Hilbert code according to the geometric coordinates of the point cloud, sort it to get the Hilbert order, and decode it according to the Hilbert order. In the Hilbert order, find the K points closest to the current point among the maximum number of neighbors (maxNumOfNeighbours) of the current point to be decoded, where maxNumOfNeighbours defaults to 128, K is 3, and the distance calculation method is Manhattan distance.

For these two prediction methods of reflectivity, prediction is performed according to geometric position, which is called "prediction method based on geometric position". If the residual generated by the prediction method based on geometric position is large, the prediction method based on attribute value can usually reduce the prediction residual and improve decoding efficiency.

The prediction method based on attribute values is divided into the following steps:

(1) Save the most recently decoded 32 different attribute prediction values in the candidate prediction value table;

(2) Select the point closest to the current point attribute and use its attribute value as the attribute prediction value of the current point;

(3) Decode the sequence number of the selected point in the candidate prediction value table.

In PCRM, the prediction method based on geometric position is first used to predict the attributes, and the residual values obtained are counted, that is, the residual values greater than or equal to 3 are accumulated and their number is recorded. After decoding 512 points, the average value of the residual values greater than or equal to 3 is calculated and compared with the pre-set threshold. When the calculated residual average exceeds the threshold, switch to the prediction method based on attribute value. In PCRM, the lossless decoding threshold is 50, and the lossy decoding threshold is 25.

The prediction method provided in the embodiment of the present application is applied to the above-mentioned point cloud prediction. By sorting the attribute information of the repeated points, repeated points with continuous attribute information are obtained, and the prediction reference points are determined using the order of the sorted repeated points to improve the decoding performance.

The embodiment of the present application provides an encoder. FIG9 is a schematic diagram of a first component structure of the encoder in the embodiment of the present application. As shown in FIG9 , the encoder 90 includes:

The first determining unit 901 is configured to determine geometric information of a point cloud to be encoded; based on the geometric information of the point cloud to be encoded, determine a set of repeated points; wherein the set of repeated points includes at least two points in the point cloud to be encoded;

A first sorting unit 902 is configured to sort the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

The first prediction unit 903 is configured to perform attribute prediction on the repeated points based on the repeated point sequence, and determine attribute prediction results of the repeated points.

In some embodiments, the first determining unit 901 is configured to determine points in the point cloud to be encoded whose distance from the target point is less than a distance threshold based on geometric information of the point cloud to be encoded, to form the repeated point set.

In some embodiments, the first determination unit 901 is configured to determine, based on the geometric information of the point cloud to be encoded, that points with the same geometric information constitute the repeated point set.

In some embodiments, the first determination unit 901 is configured to generate a Morton code or a Hilbert code of the point cloud to be encoded based on the geometric information; and determine that points with equal Morton codes or equal Hilbert codes constitute the repeated point set.

In some embodiments, the first sorting unit 902 is configured to sort the repeated points in the repeated point set in order of attribute information from small to large to determine the order of the repeated points; or, to sort the repeated points in the repeated point set in order of attribute information from large to small to determine the order of the repeated points.

In some embodiments, the first prediction unit 903 is configured to determine a predicted reference point for the current repeated point based on the repeated point sequence when the current repeated point is not the first repeated point in the repeated point sequence; and perform attribute prediction on the current repeated point based on the reconstructed attribute information of the predicted reference point.

In some embodiments, the first prediction unit 903 is configured to determine the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer; and use the first N repeated points as prediction reference points for the current repeated point.

In some embodiments, the first prediction unit 903 is configured to determine the encoded neighbor points of the current repeated point based on the point cloud order of the point cloud to be encoded; determine the first M repeated points of the current repeated point based on the repeated point order; wherein M is a positive integer; and use the neighbor points and the first M repeated points as prediction reference points for the current repeated point.

In some embodiments, the first prediction unit 903 is configured to determine the encoded neighbor points of the current repeated point based on the point cloud order of the point cloud to be encoded when the current repeated point is the first repeated point in the repeated point sequence; use the neighbor points as prediction reference points for the current repeated point; and perform attribute prediction on the current repeated point based on the reconstructed attribute information of the predicted reference points.

In some embodiments, the neighbor points include: K neighbor points with the shortest distance from the current repeated point; wherein K is a positive integer.

In some embodiments, the distance is Manhattan distance.

In some embodiments, the first sorting unit 902 is configured to reorder the point cloud based on the geometric information of the point cloud to be encoded, and determine the point cloud order of the point cloud to be encoded.

In some embodiments, the first sorting unit 902 is configured to, when the point cloud to be encoded is a dense point cloud, perform Morton reordering based on the geometric information of the point cloud to be encoded to obtain a first Morton order; perform Morton reordering after adding an offset to the geometric information of the point cloud to be encoded to obtain a second Morton order;

Correspondingly, the first prediction unit 903 is configured to determine the first P neighbor points of the current repeated point from the first Morton order; determine the first Q neighbor points of the current repeated point from the second Morton order; determine the K neighbor points with the smallest distance to the current repeated point from the first P neighbor points and the first Q neighbor points; wherein P and Q are both positive integers, and K is less than the sum of P+Q.

In some embodiments, the first sorting unit 902 is configured to perform Hilbert reordering based on geometric information of the point cloud to be encoded to obtain a Hilbert order when the point cloud to be encoded is a sparse point cloud;

Correspondingly, the first prediction unit 903 is configured to determine the first L neighbor points of the current repeated point from the Hilbert order; determine the K neighbor points with the smallest distance to the current repeated point from the first L neighbor points; wherein L is a positive integer and K is less than L.

In some embodiments, the first prediction unit 903 is configured to perform a weighted operation on the reconstruction attribute information of the prediction reference points when the number of the prediction reference points is greater than 1, so as to obtain the prediction attribute information of the current repeated point.

In some embodiments, the first prediction unit 903 is configured to determine a weight value of each predicted reference point based on a distance between each predicted reference point and the current repeated point.

In some embodiments, the first prediction unit 903 is configured to determine weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence when the predicted reference points include at least two repeated points.

In some embodiments, the first determining unit 901 is configured to determine the reconstructed geometric information of the point cloud to be encoded.

In practical applications, an embodiment of the present application also provides an encoder. Figure 10 is a schematic diagram of the second component structure of the encoder in the embodiment of the present application. The encoder 100 includes: a first memory 1001 and a first processor 1002; the first memory 1001 stores a computer program that can be run on the first processor 1002, and the point cloud encoding method on the encoder side is executed by the first processor 1002.

By using the above encoder, repeated points are sorted so that repeated points with the same attribute information are arranged continuously. When the sorting result of the repeated points is used to predict the attributes of the repeated points, the output code stream can be reduced and the encoding performance can be improved.

The present application embodiment provides a decoder. FIG11 is a schematic diagram of a first component structure of a decoder in the present application embodiment. As shown in FIG11 , the decoder 110 includes:

The second determining unit 1101 is configured to parse the bitstream to determine the geometric information of the point cloud to be decoded; based on the geometric information of the point cloud to be decoded, determine a set of repeated points; wherein the set of repeated points includes at least two points in the point cloud to be decoded;

A second sorting unit 1102 is configured to sort the attribute information of the repeated points in the repeated point set to determine the order of the repeated points;

The second prediction unit 1103 is configured to perform attribute prediction on the repeated points based on the repeated point sequence, and determine attribute prediction results of the repeated points.

In some embodiments, the second determining unit 1101 is configured to determine points in the point cloud to be decoded whose distances from the target point are less than a distance threshold based on the geometric information of the point cloud to be decoded, to form the repeated point set.

In some embodiments, the second determination unit 1101 is configured to determine, based on the geometric information of the point cloud to be decoded, that points with the same geometric information constitute the repeated point set.

In some embodiments, the second determination unit 1101 is configured to generate a Morton code or a Hilbert code of the point cloud to be decoded based on the geometric information; and determine that points with equal Morton codes or equal Hilbert codes constitute the repeated point set.

In some embodiments, the second sorting unit 1102 is configured to sort the repeated points in the repeated point set in order of attribute information from small to large to determine the order of the repeated points; or, to sort the repeated points in the repeated point set in order of attribute information from large to small to determine the order of the repeated points.

In some embodiments, the second prediction unit 1103 is configured to determine a predicted reference point for the current repeated point based on the repeated point sequence when the current repeated point is not the second repeated point in the repeated point sequence; and perform attribute prediction on the current repeated point based on the reconstructed attribute information of the predicted reference point.

In some embodiments, the second prediction unit 1103 is configured to determine the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer; and use the first N repeated points as prediction reference points for the current repeated point.

In some embodiments, the second prediction unit 1103 is configured to determine the decoded neighbor points of the current repeated point based on the point cloud order of the point cloud to be decoded; determine the first M repeated points of the current repeated point based on the repeated point order; where M is a positive integer; and use the neighbor points and the first M repeated points as prediction reference points for the current repeated point.

In some embodiments, the second prediction unit 1103 is configured to determine the decoded neighbor points of the current repeated point based on the point cloud order of the point cloud to be decoded when the current repeated point is the second repeated point in the repeated point sequence; use the neighbor points as prediction reference points for the current repeated point; and perform attribute prediction on the current repeated point based on the reconstructed attribute information of the predicted reference points.

In some embodiments, the neighbor points include: K neighbor points with the shortest distance from the current repeated point; wherein K is a positive integer.

In some embodiments, the distance is Manhattan distance.

In some embodiments, the second sorting unit 1102 is configured to reorder the point clouds based on the geometric information of the point clouds to be decoded, and determine the point cloud order of the point clouds to be decoded.

In some embodiments, the second sorting unit 1102 is configured to, when the point cloud to be decoded is a dense point cloud, perform Morton reordering based on the geometric information of the point cloud to be decoded to obtain a second Morton order; perform Morton reordering after adding an offset to the geometric information of the point cloud to be decoded to obtain the second Morton order;

Correspondingly, the second prediction unit 1103 is configured to determine the first P neighbor points of the current repeated point from the second Morton order; determine the first Q neighbor points of the current repeated point from the second Morton order; determine the K neighbor points with the smallest distance to the current repeated point from the first P neighbor points and the first Q neighbor points; wherein P and Q are both positive integers, and K is less than the sum of P+Q.

In some embodiments, the second sorting unit 1102 is configured to perform Hilbert reordering based on geometric information of the point cloud to be decoded to obtain a Hilbert order when the point cloud to be decoded is a sparse point cloud;

Correspondingly, the second prediction unit 1103 is configured to determine the first L neighbor points of the current repeated point from the Hilbert order; determine the K neighbor points with the smallest distance to the current repeated point from the first L neighbor points; wherein L is a positive integer and K is less than L.

In some embodiments, the second prediction unit 1103 is configured to perform a weighted operation on the reconstruction attribute information of the prediction reference points when the number of the prediction reference points is greater than 1, so as to obtain the prediction attribute information of the current repeated point.

In some embodiments, the second prediction unit 1103 is configured to determine a weight value of each predicted reference point based on a distance between each predicted reference point and the current repeated point.

In some embodiments, the second prediction unit 1103 is configured to determine weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence when the predicted reference points include at least two repeated points.

In some embodiments, the second determining unit 1101 is configured to determine the reconstructed geometric information of the point cloud to be decoded.

In practical applications, an embodiment of the present application also provides a decoder. Figure 12 is a schematic diagram of the second component structure of the decoder in the embodiment of the present application. The decoder 120 includes: a second memory 1201 and a second processor 1202; the second memory 1201 stores a computer program that can be run on the second processor 1202, and a point cloud decoding method on the decoder side when the second processor 1202 executes the program.

By using the above decoder, the repeated points are sorted so that the repeated points with the same attribute information are arranged continuously. When the repeated point attribute prediction is performed using the sorting result of the repeated points, the decoding performance is improved.

It is understandable that in the embodiments of the present application, a "unit" may be a part of a circuit, a part of a processor, a part of a program or software, etc., and of course, it may be a module, or it may be non-modular. Moreover, the components in the present embodiment may be integrated into a processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional module.

If the integrated unit is implemented in the form of a software function module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor to perform all or part of the steps of the method described in this embodiment. The aforementioned storage medium includes: U disk, mobile hard disk, read only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the first processor 1803. The above-mentioned first processor 1803 can be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to execute, or the hardware and software modules in the decoding processor are combined and executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc.

Accordingly, an embodiment of the present application provides a computer storage medium on which a computer program is stored. When the computer program is executed by a first processor, it implements the point prediction method on the encoder side; or, when the computer program is executed by a second processor, it implements the prediction method on the decoder side.

It should be noted here that the description of the above storage medium and device embodiments is similar to the description of the above method embodiments, and has similar beneficial effects as the method embodiments. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

It should be noted that, in this application, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element.

The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

The methods disclosed in several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

The features disclosed in several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

The features disclosed in several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Industrial Applicability

The embodiment of the present application provides a prediction method, an encoder, a decoder, and a storage medium, the method comprising: determining the geometric information of a point cloud to be encoded/decoded; determining a set of repeated points based on the geometric information of the point cloud to be encoded/decoded; wherein the set of repeated points includes at least two points in the point cloud to be encoded/decoded; sorting the attribute information of the repeated points in the set of repeated points to determine the order of the repeated points; and predicting the attributes of the repeated points based on the order of the repeated points to determine the attribute prediction results of the repeated points. In this way, by sorting the attribute information of the repeated points, repeated points with continuous attribute information are obtained, and the prediction reference points are determined using the sorted order of the repeated points, which is conducive to reducing the output code stream and improving the encoding and decoding performance.

Claims (40)

1. A prediction method applied to an encoder, the method comprising:

Determining geometric information of a point cloud to be encoded;

determining a repeated point set based on the geometric information of the point cloud to be encoded; wherein the repeated point set comprises at least two points in the point cloud to be encoded;

sorting attribute information of the repeated points in the repeated point set, and determining the sequence of the repeated points;

and carrying out attribute prediction on the repeated points based on the repeated point sequence, and determining an attribute prediction result of the repeated points.

2. The method of claim 1, wherein the determining a set of duplicate points based on geometric information of the point cloud to be encoded comprises:

And determining points, of which the distances between the points to be coded and the target points are smaller than a distance threshold, in the point cloud to be coded based on the geometric information of the point cloud to be coded, so as to form the repeated point set.

3. The method of claim 2, wherein the determining a set of duplicate points based on geometric information of the point cloud to be encoded comprises:

And determining that the points with the same geometric information form the repeated point set based on the geometric information of the point cloud to be encoded.

4. A method according to claim 3, wherein said determining a set of repetition points based on geometrical information of the point cloud to be encoded comprises:

generating Morton codes or Hilbert codes of the point cloud to be coded based on the geometric information;

points for which the Morton codes are equal or the Hilbert codes are equal are determined to constitute the set of repetition points.

5. The method of claim 1, wherein the ordering the attribute information of the repeat points in the set of repeat points, determining a repeat point order, comprises:

sequencing the repeated points in the repeated point set according to the sequence from small to large of the attribute information, and determining the sequence of the repeated points;

Or sorting the repeated points in the repeated point set according to the order of the attribute information from large to small, and determining the repeated point order.

6. The method of claim 1, wherein the predicting the repeat points for the attributes based on the order of repeat points comprises:

when the current repetition point is not the first repetition point in the repetition point sequence, determining a prediction reference point of the current repetition point based on the repetition point sequence;

and carrying out attribute prediction on the current repeated point based on the reconstruction attribute information of the prediction reference point.

7. The method of claim 6, wherein the determining a predicted reference point for a current repetition point based on the repetition point order comprises:

determining the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer;

and taking the first N repeated points as prediction reference points of the current repeated point.

8. The method of claim 6, wherein the determining a predicted reference point for a current repetition point based on the repetition point order comprises:

Determining the coded neighbor points of the current repeated point based on the point cloud sequence of the point cloud to be coded;

determining the first M repeated points of the current repeated point based on the repeated point sequence; wherein M is a positive integer;

and taking the neighbor point and the first M repeated points as prediction reference points of the current repeated point.

9. The method of claim 6, wherein the method further comprises:

when the current repetition point is the first repetition point in the repetition point sequence, determining the coded neighbor point of the current repetition point based on the point cloud sequence of the point cloud to be coded;

taking the neighbor point as a prediction reference point of the current repetition point;

and carrying out attribute prediction on the current repeated point based on the reconstruction attribute information of the prediction reference point.

10. The method of claim 8 or 9, wherein the neighbor point comprises: k neighbor points with minimum distances from the current repeated point; wherein K is a positive integer.

11. The method of claim 10, wherein the distance is a manhattan distance.

12. The method according to claim 8 or 9, wherein the method further comprises: and carrying out point cloud reordering based on the geometric information of the point cloud to be encoded, and determining the point cloud sequence of the point cloud to be encoded.

13. The method of claim 12, wherein the determining the order of the point clouds to be encoded based on the point cloud reordering based on the geometric information of the point clouds to be encoded comprises:

when the point cloud to be encoded is dense point cloud, morton reordering is carried out based on geometric information of the point cloud to be encoded, and a first Morton order is obtained;

adding an offset to the geometric information of the point cloud to be encoded, and then performing Morton reordering to obtain a second Morton order;

the determining the coded neighbor point of the current repeated point based on the point cloud sequence of the point cloud to be coded comprises the following steps:

determining the first P neighbor points of the current repeated point from the first Morton order;

determining the first Q neighbor points of the current repetition point from the second Morton order;

Determining K neighbor points with the minimum distance from the current repeated point from the front P neighbor points and the front Q neighbor points;

Wherein, P and Q are positive integers, K is smaller than the sum of P+Q.

14. The method of claim 12, wherein the determining the order of the point clouds to be encoded based on the re-establishing geometry information of the point clouds to be encoded comprises:

When the point cloud to be encoded is a sparse point cloud, hilbert reordering is performed based on geometric information of the point cloud to be encoded, and a Hilbert sequence is obtained;

the determining the coded neighbor point of the current repeated point based on the point cloud sequence of the point cloud to be coded comprises the following steps:

determining the first L neighbor points of the current repeated point from the Hilbert sequence;

from the first L neighbor points, determining K neighbor points with the minimum distance from the current repeated point;

wherein L is a positive integer, and K is smaller than L.

15. The method of claim 6, wherein said predicting the attribute of the current repetition point based on the reconstructed attribute information of the prediction reference point comprises:

And when the number of the prediction reference points is greater than 1, carrying out weighted operation on the reconstruction attribute information of the prediction reference points to obtain the prediction attribute information of the current repetition point.

16. The method of claim 15, wherein the method further comprises:

A weight value for each prediction reference point is determined based on the distance of each prediction reference point from the current repetition point.

17. The method of claim 16, wherein the method further comprises:

When the prediction reference point comprises at least two repeated points, determining weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence.

18. The method of claim 1, wherein the determining geometric information of the point cloud to be encoded comprises: and determining reconstruction geometric information of the point cloud to be encoded.

19. A prediction method applied to a decoder, the method comprising:

analyzing the code stream and determining geometric information of the point cloud to be decoded;

determining a repeated point set based on the geometric information of the point cloud to be decoded; wherein the repeated point set comprises at least two points in the point cloud to be decoded;

sorting attribute information of the repeated points in the repeated point set, and determining the sequence of the repeated points;

and carrying out attribute prediction on the repeated points based on the repeated point sequence, and determining an attribute prediction result of the repeated points.

20. The method of claim 19, wherein the determining a set of duplicate points based on geometric information of the point cloud to be decoded comprises:

And determining points, of which the distances between the points to be decoded and the target points are smaller than a distance threshold, in the point cloud to be decoded based on the geometric information of the point cloud to be decoded, so as to form the repeated point set.

21. The method of claim 20, wherein the determining a set of duplicate points based on geometric information of the point cloud to be decoded comprises:

and determining that the points with the same geometric information form the repeated point set based on the geometric information of the point cloud to be decoded.

22. The method of claim 21, wherein the determining a set of duplicate points based on geometric information of the point cloud to be decoded comprises:

generating Morton codes or Hilbert codes of the point cloud to be decoded based on the geometric information;

points for which the Morton codes are equal or the Hilbert codes are equal are determined to constitute the set of repetition points.

23. The method of claim 19, wherein the ordering the attribute information of the repeat points in the set of repeat points, determining a repeat point order, comprises:

sequencing the repeated points in the repeated point set according to the sequence from small to large of the attribute information, and determining the sequence of the repeated points;

Or sorting the repeated points in the repeated point set according to the order of the attribute information from large to small, and determining the repeated point order.

24. The method of claim 19, wherein the predicting the repeat points based on the repeat point order comprises:

when the current repetition point is not the first repetition point in the repetition point sequence, determining a prediction reference point of the current repetition point based on the repetition point sequence;

and carrying out attribute prediction on the current repeated point based on the reconstruction attribute information of the prediction reference point.

25. The method of claim 24, wherein the determining a predicted reference point for a current repetition point based on the repetition point order comprises:

determining the first N repeated points of the current repeated point based on the repeated point sequence; wherein N is a positive integer;

and taking the first N repeated points as prediction reference points of the current repeated point.

26. The method of claim 24, wherein the determining a predicted reference point for a current repetition point based on the repetition point order comprises:

determining decoded neighbor points of the current repeated point based on the point cloud sequence of the point cloud to be decoded;

determining the first M repeated points of the current repeated point based on the repeated point sequence; wherein M is a positive integer;

and taking the neighbor point and the first M repeated points as prediction reference points of the current repeated point.

27. The method of claim 24, wherein the method further comprises:

when the current repetition point is the first repetition point in the repetition point sequence, determining decoded neighbor points of the current repetition point based on the point cloud sequence of the point cloud to be decoded;

taking the neighbor point as a prediction reference point of the current repetition point;

and carrying out attribute prediction on the current repeated point based on the reconstruction attribute information of the prediction reference point.

28. The method of claim 26 or 27, wherein the neighbor point comprises: k neighbor points with minimum distances from the current repeated point; wherein K is a positive integer.

29. The method of claim 28, wherein the distance is a manhattan distance.

30. The method of claim 26 or 27, wherein the method further comprises: and carrying out point cloud reordering based on the geometric information of the point cloud to be decoded, and determining the point cloud sequence of the point cloud to be decoded.

31. The method of claim 30, wherein the determining the order of the point clouds to be decoded based on the point cloud reordering based on the geometric information of the point clouds to be decoded comprises:

when the point cloud to be decoded is dense point cloud, morton reordering is carried out based on geometric information of the point cloud to be decoded, and a first Morton order is obtained;

adding an offset to the geometric information of the point cloud to be decoded, and then performing Morton reordering to obtain a second Morton order;

the determining the decoded neighbor point of the current repeated point based on the point cloud sequence of the point cloud to be decoded comprises the following steps:

determining the first P neighbor points of the current repeated point from the first Morton order;

determining the first Q neighbor points of the current repetition point from the second Morton order;

Determining K neighbor points with the minimum distance from the current repeated point from the front P neighbor points and the front Q neighbor points;

Wherein, P and Q are positive integers, K is smaller than the sum of P+Q.

32. The method of claim 30, wherein the determining the order of the point clouds to be decoded based on the reconstructed geometry information of the point clouds to be decoded comprises:

When the point cloud to be decoded is a sparse point cloud, hilbert reordering is performed based on geometric information of the point cloud to be decoded, and a Hilbert sequence is obtained;

the determining the decoded neighbor point of the current repeated point based on the point cloud sequence of the point cloud to be decoded comprises the following steps:

determining the first L neighbor points of the current repeated point from the Hilbert sequence;

from the first L neighbor points, determining K neighbor points with the minimum distance from the current repeated point;

wherein L is a positive integer, and K is smaller than L.

33. The method of claim 24, wherein said predicting the attribute of the current repetition point based on the reconstructed attribute information of the prediction reference point comprises:

And when the number of the prediction reference points is greater than 1, carrying out weighted operation on the reconstruction attribute information of the prediction reference points to obtain the prediction attribute information of the current repetition point.

34. The method of claim 33, wherein the method further comprises:

A weight value for each prediction reference point is determined based on the distance of each prediction reference point from the current repetition point.

35. The method of claim 34, wherein the method further comprises:

When the prediction reference point comprises at least two repeated points, determining weight values of the at least two repeated points based on positions of the at least two repeated points in the repeated point sequence.

36. An encoder, wherein the encoder comprises:

the first determining unit is configured to determine geometric information of the point cloud to be encoded; determining a repeated point set based on the geometric information of the point cloud to be encoded; wherein the repeated point set comprises at least two points in the point cloud to be encoded;

the first ordering unit is configured to order the attribute information of the repeated points in the repeated point set and determine the sequence of the repeated points;

and the first prediction unit is configured to predict the attribute of the repeated points based on the sequence of the repeated points and determine the attribute prediction result of the repeated points.

37. A decoder, wherein the decoder comprises:

the second determining unit is configured to analyze the code stream and determine reconstruction geometric information of the point cloud to be decoded; determining a repeated point set based on the geometric information of the point cloud to be decoded; wherein the repeated point set comprises at least two points in the point cloud to be decoded;

the second ordering unit is configured to order the attribute information of the repeated points in the repeated point set and determine the sequence of the repeated points;

and the second prediction unit is configured to predict the attribute of the repeated points based on the sequence of the repeated points and determine the attribute prediction result of the repeated points.

38. An encoder, the encoder comprising a first memory and a first processor; wherein,

The first memory is used for storing a computer program capable of running on the first processor;

the first processor being configured to perform the method of any one of claims 1 to 18 when the computer program is run.

39. A decoder comprising a second memory and a second processor; wherein,

The second memory is used for storing a computer program capable of running on the second processor;

the second processor being adapted to perform the method of any of claims 19 to 35 when the computer program is run.

40. A computer storage medium storing a computer program which when executed by a first processor implements the method of any one of claims 1 to 18 or when executed by a second processor implements the method of any one of claims 19 to 35.