CN108322754A - Coding method and corresponding coder and decoder - Google Patents

Abstract

An encoding method for encoding a set of reference settings for a picture, the method comprising: judging whether the current coding image is a reference image, if so, setting a reference setting set of the current coding image as an image which does not mark as a reference image any more; and if the current coding image is judged to be the reference image, setting the reference setting set of the current coding image as an image which marks the current coding image not to be used as the reference image. The coding method, the coder and the decoder can ensure that the coded image buffer can avoid the problem that the coded image buffer can not normally move out of the reference image due to image loss, and improve the working efficiency of the coded image buffer.

Description

Encoding method and corresponding encoder and decoder

technical field

The present invention relates to video encoding/decoding, and more particularly, to techniques for coding video data.

Background technique

Digital video capabilities can be incorporated into a wide range of devices including digital televisions, digital live broadcast systems, wireless broadcast systems, personal digital assistants (PDAs), laptop or desktop computers, tablet computers, electronic books Readers, digital video cameras, digital recording devices, digital media players, video game devices, video game consoles, cellular or satellite radiotelephones, so-called "smart phones", video teleconferencing devices, video streaming devices, and similar. Digital video devices implement video compression techniques such as those described in: MPEG-2, MPEG-4, ITU-T H.263, ITU-T H.264/MPEG-4 Part 10 (Advanced Video Coding (AVC)), the High Efficiency Video Coding (HEVC) standard currently in development, and extensions to these standards. Video devices can transmit, receive, encode, decode and/or store digital video information more efficiently by implementing these video compression techniques.

Traditional video coding standards usually adopt block-based coding techniques to apply spatial and temporal redundancy. For example, the basic approach is to partition the entire source frame into blocks, perform intra/inter prediction on each block, perform redundant transformation per block, and perform quantization and entropy coding. In addition, a reconstructed frame is generated in the encoding loop to provide reference pixel data for encoding subsequent blocks. For some video coding standards, in-loop filtering is used to enhance the image quality of reconstructed frames. An inter-coded (P or B) video block of a picture may use spatial prediction with respect to reference samples in neighboring blocks in the same picture or temporal prediction with respect to reference samples in other reference pictures. An image may be referred to as a frame, and a reference image may be referred to as a reference frame.

A video decoder is used to perform the inverse of video encoding performed by a video encoder. For example, an entropy encoder in a video encoder is used to generate an encoded bitstream, and an entropy decoder in a video decoder is used to decode the encoded bitstream. An entropy coder implements arithmetic coding.

Contents of the invention

One of the objectives of the present invention is to provide an encoding method and corresponding encoder and decoder to encode a reference set of images and decode images to ensure the normal operation of the decoded image buffer.

According to the first aspect of the present invention, an encoding method is provided, the encoding method is used for encoding a reference set of images, the method includes: according to whether the currently encoded image is a reference image, setting the decoded reference set of the current encoded image The image buffer holds information.

Wherein, the method judges whether the currently coded picture is a reference picture, and if it is judged that the current coded picture is a non-reference picture, then setting the reference setting set of the current coded picture to not mark pictures that are no longer used as reference pictures; If the picture is a reference picture, then set the reference setting set of the current coded picture to mark a picture that is no longer used as a reference picture.

According to another aspect of the present invention, an encoding method is provided, which is used for a reference setting set of an encoded image, the method comprising: setting the encoding method according to whether the currently encoded image is a non-random access decodable image or a random access decodable image The set of reference settings for the currently encoded image.

Wherein, the encoding method further sets the reference setting set of the current encoded image according to whether the image no longer used as the reference image is a non-random access decodable image after the decoding of the currently encoded image.

Wherein, when the currently coded picture is a non-random access decodable picture, and the picture that is no longer used as a reference picture is a non-random access decodable picture, mark the picture that is no longer used as a reference picture.

Through the encoding method, encoder and decoder of the present invention, the encoded image buffer can avoid the problem that the encoded image buffer cannot normally move out the reference image due to lost images, and improve the working efficiency of the encoded image buffer. .

Other objects of the present invention can be understood by those skilled in the art after reading the following detailed description of the present invention and the accompanying drawings for example.

Description of drawings

FIG. 1 is a schematic diagram showing a reference relationship between images.

2A-2C are schematic diagrams of images stored in the DPB.

3A-3C are another schematic diagrams of images stored in the DPB.

4A-4D are schematic diagrams of image management in the DPB according to an embodiment of the present invention.

Fig. 5 is a flow chart of an encoding method according to an embodiment of the present invention

FIG. 6 is a flowchart of a decoding method according to an embodiment of the present invention.

FIG. 7 is a flowchart of a method for an encoder to set DPB holding information of a reference setting set according to an embodiment of the present invention.

Detailed ways

Throughout the specification and claims, certain terms are used to refer to particular components. As will be understood by those skilled in the art, manufacturers may use different names to refer to a certain component. This document does not attempt to distinguish between those parts that may have the same function but have different names. In the following description and claims, the terms "comprises" and "comprises" are used in an open-ended manner and should therefore be construed as "including, but not limited to...". Likewise, the term "coupled" may mean both an indirect electrical connection and a direct electrical connection. Thus, if a device couples to another device, that connection may be through a direct electrical connection or through an indirect electrical connection through other devices and connections.

The techniques of this disclosure generally relate to the management of reference pictures for inter prediction. For example, a video decoder includes a decoded picture buffer (DPB). The DPB stores decoded pictures and pictures not yet output for display, including reference pictures. A reference picture is a picture that can potentially be used for inter prediction of a picture. In other words, the video coder may predict a picture based on one or more reference pictures stored in the DPB during decoding of the picture. However, if the reference picture stored in the DPB is lost, decoding of the picture referring to the lost reference picture will fail, resulting in a decoding error.

In order to efficiently utilize the DPB, DPB management procedures, such as a storage procedure of decoded pictures in the DPB, a marking procedure of reference pictures, an output and removal procedure of decoded pictures from the DPB, etc., may be specified. Generally speaking, in some current and developing video decoding standards, DPB management may include one or more of the following aspects: image identification and reference image identification, reference image list construction, reference image marking, information from DPB Image output from DPB, image insertion into DPB, and image removal from DPB. The decoder judges which pictures in the current DPB are no longer used as reference pictures for subsequent pictures according to the settings in the reference frame set of the currently processed picture, so as to remark these reference pictures as non-reference pictures. DPB contains two types of images, reference images and unoutput images, where reference images can be unoutput images or images that have been output, unoutput images can be reference images or non-reference images, and output non- Reference images need to be moved out of the DPB. The DPB removal action specified in the present invention is described for the output reference image. The image just decoded is an unoutput image, and according to the output image rules defined by AVS2, after an image is output, the image is marked as an output image. That is, the image stored in the DPB needs to meet at least one of the following two conditions, one of which is a reference image, and the other is that it has not been output for display. Thus when an original reference picture is re-marked as a non-reference picture, and the re-marked non-reference picture is marked as an output picture, it will be removed from the DPB to ensure that there is enough space in the DPB to store subsequent The reference image to ensure the smooth progress of the decoding process.

As described above, in image encoding and decoding procedures, in order to utilize spatial and temporal redundancy, predictive encoding or predictive decoding is usually performed by referring to images. When a picture is predictively coded or decoded with reference to other pictures, the picture is marked as a reference picture, and when a picture is predictively encoded or decoded without being referenced by other pictures, the picture is marked as a non-reference picture. When a picture is decoded, the picture may be a non-reference picture or a reference picture. A reference picture may be a long-term reference picture or a short-term reference picture, and when a reference picture is marked "unused for reference," it may become no longer needed for reference. In some video coding standards, there may be a reference picture marking operation that changes the state of a reference picture. For example, FIG. 1 is a schematic diagram showing a reference relationship between images, where the arrow relationship represents the reference relationship. The obliquely hatched images in FIG. 1 represent images that are referred to when other images are encoded and decoded, that is, reference images, and represent images stored in the DPB. After picture 7 has been decoded, picture 0 is no longer used as a reference picture for subsequent pictures. When picture 0 has been output and displayed, picture 0 can be removed from the DPB, thereby saving space in the DPB. In other words, in the process of image decoding, image 0 needs to be marked as a non-reference image from a reference image. Only by correctly marking an image from a reference image to a non-reference image can the image in the DPB be moved out, thereby ensuring that the DPB provides enough space for subsequent reference images to be stored.

For example, if a marking error occurs during image decoding, so that picture 0 is not correctly marked from a reference picture to a non-reference picture, then picture 0 will always keep the mark of the reference picture, so that picture 0 will always be stored in the DPB , occupying space in the DPB. Since the space of DPB is usually limited, for example, in the AVS2 standard, 5-16 images can usually be stored in DPB. If it continues to accumulate, it will lead to less and less available space in the DPB, reducing the use efficiency of the DPB, and may even cause the DPB to have no available storage space and cause image overflow. When the DPB overflows, usually because the decoder cannot accurately know which image should be selected to be removed from all the images stored in the DPB at this time, the decoder chooses to move out randomly, that is, the decoder randomly selects an image in the DPB to move out . When the random image selected by the decoder is the reference image of the subsequent image to be decoded, it will lead to the problem that the subsequent image to be decoded has no reference image and cannot be decoded normally.

To aid in understanding, a brief description of how reference picture marking may occur according to some video coding standards is provided below. Some of the techniques described in this disclosure address issues that may exist in reference image marking, and DPB image removal and output, in order to improve utilization efficiency of the DPB.

During the image encoding process, the encoder encodes the image according to the image encoding standard (eg, HEVC, AVS2). In the encoding process, a reference configuration set (reference configuration set) corresponding to it is constructed for each image, the reference configuration set is an absolute description about the reference image, and the reference configuration set indicates the reference image of the image corresponding to the reference configuration set, and Indicate which images in the image sequence are no longer used as reference images for subsequent images. In one embodiment, a video encoder may signal a reference setting set in various ways of signaling syntax elements that a video decoder may utilize to derive a reference setting set. For example, a video decoder may signal the syntax elements in a video sequence header, intra picture header/inter picture header, slice header, or any combination thereof. For purposes of illustration, a video encoder may signal syntax elements using sequence headers, picture headers, and slice headers. To derive a reference set set, a video decoder may implement a decoding process to determine identifiers for pictures belonging to the reference set set.

Table 1 is an example of a reference setting set for a picture set by an encoder.

Table 1

where u(1), u(3), u(6), f(1) are descriptors. referred_by_others_flag indicates whether the picture is a reference picture. num_of_reference_picture indicates how many pictures are referenced by the encoding of this picture. And the subsequent syntax indicates which images are the reference images of the image, that is, indicates the image sequence number of the reference image of the image. num_of_removed_picture indicates the number of marking a picture from a reference picture to a non-reference picture. And the subsequent syntax indicates which pictures are removed from the DPB, that is, indicates the sequence number of the picture marked as a non-reference picture from the reference picture. In other words, the reference set of pictures contains the DPB hold message, and the decoder manages the reference pictures stored in the DPB by parsing the DPB hold message.

During the image decoding process, the decoder derives the above-mentioned reference setting set, and decides whether to store the currently decoded image in the DPB according to the above-mentioned reference setting set, and which images in the DPB are marked as non-reference images and removed from the DPB.

2A-2C are schematic diagrams of images stored in the DPB. Wherein, in FIG. 2A , 5 images are stored in the DPB, which are image 0-image 4, wherein R represents a reference image, and O represents an output image. In Figure 2B, the decoder is currently decoding picture 5, and when parsing the reference set set for picture 5, it resolves that picture 5 is a non-reference picture (indicated by N in Figure 2B), but since picture 5 has just finished decoding, it has not yet output (indicated by D in FIG. 2B ), so image 5 remains in the DPB. In addition, when the decoder parses the reference setting set of picture 5, it resolves that picture 1-picture 4 is no longer used as a reference picture of other pictures, so in the DPB shown in Figure 2B, picture 1-picture 4 is marked as N, and In FIG. 2C , the DPB is removed so that the DPB has enough space to store subsequent reference images. And because picture 0 is still referenced by subsequent pictures to be decoded, picture 0 is still kept in the DPB.

3A-3C are another schematic diagrams of images stored in the DPB. In FIG. 3A, image 0-image 4 are stored in the DPB, and image 0-image 4 are reference images that have been output and displayed. If picture 5 loss occurs, the decoder cannot parse the information carried in the reference set set of picture 5, and thus cannot mark pictures 1-picture 4 from reference pictures to non-reference pictures. Therefore, when the picture 6 is decoded, as shown in FIG. 3B , the picture 6 is stored in the DPB as a reference picture that has just been decoded and has not yet been output. Therefore, when the decoder decodes the picture 7, as shown in FIG. 3C , since the picture 7 is a reference picture, it needs to be saved to the DPB. Since there is no free space in the DPB at this time, the decoder cannot tell which picture should be moved out. If the decoder randomly decides to move out picture 0, picture 7 will be saved in the DPB instead of picture 0. If the next picture 8 needs to be decoded with reference to picture 0, picture 8 cannot be decoded correctly due to the lack of picture 0 in the DPB.

In order to solve the above problems, the encoder of the present invention imposes restrictions on the set of reference settings in non-reference images when encoding. That is, according to whether the current coded picture is a reference picture, the decoded picture buffer holding information in the reference setting set of the current coded picture is set. Specifically, when the encoder of the present invention encodes a reference setting set of non-reference pictures, the reference setting set does not mark any reference picture as a non-reference picture. Determine whether the current coded picture is a reference picture, if it is judged that the current coded picture is a non-reference picture, then set the reference setting set of the current coded picture to not mark the picture that is no longer used as a reference picture; if it is judged that the current coded picture is a reference picture , then set the reference setting set of the current coded picture to mark the picture that is no longer used as a reference picture. In other words, num_of_removed_picure is set to 0 in the reference set set for non-reference pictures.

In the examples of Fig. 2A-Fig. 2C and Fig. 3A-Fig. C, the encoder first judges that image 5 is a non-reference image (N), so the encoder sets num_of_removed_picure to 0 when encoding the reference setting set of image 5. Therefore, after the decoder parses the reference setting set of image 5, it will not mark image 1-image 4 from the reference image as a non-reference image, that is, after the decoder decodes image 5, it will not mark image 1-image 4 from the DPB removed. 4A-4D are schematic diagrams of image management in the DPB according to an embodiment of the present invention. The DPB in Figure 4A stores picture 0 - picture 4. When picture 5 is lost, the decoder decodes picture 6 and stores it in the DPB (as shown in Figure 4B). Moreover, when encoding image 6, the encoder judges that image 6 is a reference image, so that it can indicate which images are no longer used as reference images when encoding the reference setting set of image 6, so that the reference setting set of image 6 will indicate that image 1- Image 4 is no longer used as a reference image. As shown in Figure 4C, after decoding image 6, the decoder can mark image 1-image 4 from the reference image as a non-reference image, so that image 1-image 4 is removed from the DPB to ensure that the DPB has enough free space to store the subsequent reference image 7 (as shown in FIG. 4D ), so that the decoding proceeds smoothly.

When encoding a non-reference picture, the encoder of the present invention sets the non-reference picture so that no reference picture can be marked as a non-reference picture. That is, when encoding the reference setting set corresponding to the non-reference picture, it is indicated not to modify any reference picture as the non-reference picture. In other words, the encoder sets num_of_removed_picure to 0 in the reference set set for non-reference pictures. Therefore, by adopting the technical solution of the present invention, it can be ensured that the reference image will not fail to be marked as a non-reference image due to the loss of any non-reference image. It can be guaranteed that the DPB will not be confused and cannot be decoded due to the loss of any non-reference image.

Please refer to FIG. 5 of the accompanying drawings. FIG. 5 is a flowchart of an encoding method according to an embodiment of the present invention. Implemented by an encoder executing software, hardware, and firmware. The encoding method contained in Figure 5 consists of the following steps:

Step 501: encoding the current image;

Step 503: According to the encoding standard adopted by the current encoder, search for an image that is no longer used as a reference image after encoding the current image;

Step 505: Judging whether the current coded picture is a reference picture, if the judgment result is "yes", go to step 507, otherwise go to step 501;

Step 507: In the reference setting set of the current coded picture, set a picture that is no longer used as a reference picture after the current picture is coded.

Those skilled in the art can understand that the execution sequence of the steps in FIG. 5 can be adjusted, and each step can be split or combined, which are all within the scope of the present invention.

Furthermore, the present invention relates to a decoder capable of performing management of DPB by analyzing reference pictures and non-reference pictures set by the encoding method in FIG. 5 .

The present invention further includes a method for encoding a reference set of non-random access decodable pictures. The sequence of images to be decoded includes random access decodable pictures and non-random access decodable pictures. Video coding standards pursue higher coding and compression efficiency, but at the same time must also consider the random access performance of the compressed code stream. Random access performance refers to the ability to decode the bitstream and restore the decoded image from a certain point rather than the starting point of the bitstream, which is directly related to the user experience. Random access performance and coding compression efficiency are contradictory. How to make a compromise and balance between the two is an important issue that must be considered in video coding standards.

The requirements for random access mainly include channel switching, code stream switching, editing and splicing, random positioning of program playback, fast forward and fast rewind, etc. in the broadcasting service. Different services have different requirements for random access performance. For example, for broadcasting services, the DVB (Digital Video Broadcasting) standard stipulates that a random access point should appear every 0.k; Services such as payment services have lower requirements on random access performance.

In order to support random access to video compression streams, various video standards have adopted a series of measures. In the image data stream, it can usually be divided into multiple image sequences (frame sequence). A random access decodable picture means that the picture only refers to pictures in the current picture sequence and does not refer to reference pictures in other picture sequences. A non-random access decodable picture means that the picture refers to reference pictures in other picture sequences. In the encoding process, each random access decodable picture and non-random access decodable picture has a corresponding reference setting set (as shown in Table 1), to indicate whether the random access decodable picture or non-random access decodable picture is is the reference picture, which pictures it itself references, and which reference pictures may be marked as non-reference pictures after decoding the random access decodable picture or the non-random access decodable picture.

For example, when image loss occurs, such as video cropping and splicing, the reference image may be lost, which will cause subsequent images that need to be decoded with reference to the reference image to be unable to be decoded normally, resulting in decoder processing confusion. In order to avoid the above problems, the decoder usually directly discards the images that cannot be decoded normally due to the lack of reference images. FIG. 6 is a flowchart of a decoding method according to an embodiment of the present invention. The decoding method includes the following steps:

Step 601: the decoder obtains the image to be decoded;

Step 603: Judging whether the image to be decoded is a random access decodable image (abbreviated as RAD in the drawings), when the judgment result is "No", then proceed to Step 605, and when the judgment result is "Yes", then proceed to Step 603: 607;

Step 605: Discard the image to be decoded;

Step 607: Parse the reference setting set of the image to be decoded to perform DPB management;

Step 609: Decode the image to be decoded;

Step 611: Judging whether there are subsequent images to be decoded, if the judging result is "yes", return to step 601, if the judging result is "no", then end the decoding method.

For example, when a picture is a non-random access decodable picture, the picture needs to refer to reference pictures in other picture sequences. When performing video clipping or video splicing, usually the reference images in other sequences are lost, so the non-random access decodable image cannot be decoded normally. The decoder can be set to directly discard the non-random access decodable image without performing any decoding. However, when the reference setting set of the non-random access decodable picture contains DPB hold information, that is, when it contains the information that marks the reference picture in the DPB as a non-reference picture, the decoder directly discards the non-random access decodable picture It may lead to the problem that the reference image in the DPB cannot be marked as a non-reference image normally, resulting in DPB confusion.

According to an embodiment of the present invention, the encoder sets the reference setting set of the current coded picture according to whether the current coded picture is a non-random access decodable picture or a random access decodable picture. The encoder further sets the reference setting set of the current coded picture according to whether the picture no longer used as the reference picture is a non-random access decodable picture after the current coded picture is decoded.

The encoder only instructs non-random access decodable pictures to be marked from reference pictures to non-reference pictures when encoding the reference set set of non-random access decodable pictures, and does not mark random access decodable pictures from reference pictures to non-reference pictures . On the other hand, when encoding the reference set of random access decodable pictures, the encoder can either instruct the non-random access decodable pictures to be marked as non-reference pictures from the reference pictures, or instruct the random access decodable pictures to be marked from the reference pictures as non-reference image. In other words, when the encoder encodes a random access decodable image, its reference setting set is normally set, and when the encoder encodes a non-random access decodable image, the setting of the DPB holding information in the reference setting set needs to take into account the indicated Whether the reference image is a non-random access decodable image.

FIG. 7 is a flowchart of a method for an encoder to set DPB holding information of a reference setting set according to an embodiment of the present invention. The method includes the following steps:

Step 701: encode the current image;

Step 703: According to the video standard adopted by the encoder, search for all pictures Pic[i] that need to be marked as non-reference pictures from reference pictures, that is, search for all pictures that are no longer used as reference pictures, and set The count value of the number num_no_longer_ref_pic, to indicate the number of pictures that are no longer used as reference pictures;

Step 705: Set count value i=0;

Step 707: Judging whether the count value i is less than num_no_longer_ref_pic, if the judging result is "yes", proceed to step 709, if the judging result is "no", then the setting of the DPB holding information in the reference setting set of the current coded picture ends;

Step 709: judging whether the current coded image is a random access decodable image (abbreviated as RADPic in FIG. 7), if the judgment result is "yes", then proceed to step 711, if the judgment result is no, then proceed to step 713;

Step 711: In the reference setting set of the current picture, set the indication picture Pic[i] to no longer be a reference picture;

Step 713: judging whether the image Pic[i] is a non-random access decodable image (abbreviated as non-RADPic in FIG. 7), when the judging result is "yes", go to step 711, and when the judging result is no, go to step 715;

Step 715: Count i+1, and return to step 707 to process image Pic[i+1].

Those skilled in the art can understand each step shown in FIG. 7 after reading the above paragraphs, and further description is omitted to simplify the description.

With the above encoding method, the reference set of non-random access decodable pictures will only indicate that the non-random access decodable pictures are no longer reference pictures. Therefore, if the decoder discards a non-random access decodable picture, the decoder cannot obtain the DPB holding information in the reference set related to the discarded non-random access decodable picture, and only the non-random access decodable picture cannot be retrieved from the reference The picture is marked as a non-reference picture without the random-access decodable picture not being marked as a non-reference picture from a reference picture normally. In other words, the sum of the index of picture (DOI) of a non-random access decodable picture and the delta_doi_of_removed_picture shown in Table 1 is not equal to the DOI of a random access decodable picture. The sum of the picture sequence number of the random access decodable picture and the delta_doi_of_removed_picture shown in Table 1 may be equal to the DOI of the random access decodable picture or the DOI of the non-random access decodable picture. That is, it will only happen that the non-random access decodable image cannot be normally removed from the DPB due to the loss of the non-random access decodable image, and it will not happen that the random access decodable image cannot be normally removed from the DPB due to the loss of the non-random access decodable image. In case of removal.

Due to the loss of the previous image sequence during video clipping or video splicing, the decoder will choose to discard all non-random access decodable images in the current video sequence, so even if non-random access decodable images are lost due to non-random access If a decodable image cannot be normally removed from the DPB, the decoder will also discard the image because it is a non-random access decodable image, so that the non-random access decodable image cannot be removed from the DPB, causing DPB confusion.

The method for the reference setting set of an encoder encoding a picture provided by the present invention, by setting the reference setting set related to a non-reference picture to not indicate any picture that is no longer a reference picture, and/or by setting a non-random access decodable picture The relevant reference setting set is set to only indicate that the non-random access decodable picture is no longer used as a reference picture, so that the reference picture in the DPB can be moved out normally, avoiding the problem that the reference picture in the DPB cannot be moved out due to picture loss , to ensure the normal use of the DPB.

Those skilled in the art can understand that the present invention can be implemented using software, hardware, and firmware. Therefore, the present invention also includes a memory embedded in an encoder or a decoder for storing program codes for realizing various functions of the present invention.

The present invention is illustrated by the above examples, but the present invention is not limited to the above examples. The invention should be understood to cover various alternative embodiments and similar arrangements which will occur to those skilled in the art. Accordingly, the claims of the present invention should be construed to cover a wide range of various modified embodiments and similar arrangements that are apparent to those skilled in the art.

Claims (10)

1. a kind of coding method, for the reference setting collection of coded image, this method includes：

Whether it is that reference picture concentrates decoding figure the reference setting of the current encoded image is arranged according to current encoded image As buffering keeps information.

2. coding method according to claim 1, which is characterized in that according to current encoded image whether be reference picture come Be arranged the current encoded image reference setting concentrate decoded picture buffer keep information the step of include：

Judge whether current encoded image is reference picture, if it is determined that the current encoded image is non-reference picture, then should The reference setting collection of current encoded image is set as not marking no longer as the image with reference to image；

If it is determined that the current encoded image is reference picture, then the reference of current encoded image setting collection is set as marking No longer as the image with reference to image.

3. coding method according to claim 2, which is characterized in that wherein if it is determined that the current encoded image is non-ginseng The step of examining image, then setting not marking no longer as the image with reference to image the reference of current encoded image setting collection to Including：

The cue mark that the reference setting concentration of the current encoded image is arranged is that the quantity of non-reference picture is 0.

4. a kind of encoder requires the coding method described in 1-3 for perform claim.

5. a kind of computer readable storage medium, for storing program code, which requires described in 1-3 Coding method.

6. a kind of coding method, for the reference setting collection of coded image, this method includes：

According to whether current encoded image is nonrandom access decodable picture or random access decodable picture, this is set The reference setting collection of current encoded image.

7. according to coding method shown in claim 6, which is characterized in that further include according to coded image solution in this prior After code, no longer as refer to image image whether be nonrandom access decodable picture and present image whether right and wrong with Machine accesses decodable picture and collection is arranged the reference of the current encoded image is arranged.

8. according to coding method shown in claim 7, which is characterized in that can be solved when the current encoded image is nonrandom access Code image, and this is no longer nonrandom access decodable picture as the image with reference to image, marks this no longer as reference The image of image.

9. a kind of encoder requires the coding method described in 6-8 for perform claim.

10. a kind of computer readable storage medium, for storing program code, which requires described in 6-8 Coding method.