CN113807193B - Automatic extraction method and system for traffic road virtual line segments in laser point cloud - Google Patents
Automatic extraction method and system for traffic road virtual line segments in laser point cloud Download PDFInfo
- Publication number
- CN113807193B CN113807193B CN202110973671.6A CN202110973671A CN113807193B CN 113807193 B CN113807193 B CN 113807193B CN 202110973671 A CN202110973671 A CN 202110973671A CN 113807193 B CN113807193 B CN 113807193B
- Authority
- CN
- China
- Prior art keywords
- line segment
- broken line
- point cloud
- virtual line
- line segments
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000605 extraction Methods 0.000 title claims description 61
- 238000001914 filtration Methods 0.000 claims abstract description 55
- 238000000034 method Methods 0.000 claims abstract description 44
- 230000011218 segmentation Effects 0.000 claims abstract description 18
- 238000013507 mapping Methods 0.000 claims abstract description 13
- 238000005520 cutting process Methods 0.000 claims abstract description 12
- 230000008569 process Effects 0.000 claims description 19
- 230000000694 effects Effects 0.000 claims description 13
- 238000003860 storage Methods 0.000 claims description 10
- 238000009499 grossing Methods 0.000 claims description 8
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000012937 correction Methods 0.000 claims description 4
- 230000005764 inhibitory process Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 17
- 238000004590 computer program Methods 0.000 description 11
- 238000004519 manufacturing process Methods 0.000 description 8
- 238000001514 detection method Methods 0.000 description 5
- 238000012545 processing Methods 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000004075 alteration Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013075 data extraction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Landscapes
- Image Processing (AREA)
- Image Analysis (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention relates to a method and a system for automatically extracting virtual line segments of traffic roads in laser point clouds, wherein the method comprises the following steps: cutting laser point cloud data of a road into blocks, acquiring a overlook projection image of a laser point cloud mass, predicting a dotted line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the dotted line segment; extracting the center line of a broken line segment in the mask map and acquiring the width of the broken line segment; filtering the broken line segments in the mask map based on the structural and position characteristics of the broken line segments; according to the mapping relation from the center line and the width of the virtual line segments in the overlook projection image to three dimensions, the three-dimensional coordinate information of each virtual line segment is calculated, and after the operations of connection, filtration and duplication removal are carried out on each virtual line segment, the virtual line segments are submitted to a high-precision map making system, so that the workload of a producer and the high-precision map making cost can be reduced, and the making efficiency is improved. And road lane information is better provided for assisting automatic driving.
Description
Technical Field
The invention relates to the field of high-precision map making, in particular to an automatic extraction method and system for virtual line segments of traffic roads in laser point clouds.
Background
In the process of high-precision map making, laser point cloud data collected by traffic roads are often required to be marked for assisting intelligent driving, and lane line broken line segment information provides lane information for vehicles and is an indispensable part of the high-precision map. However, because the drawing amount of the lane line broken line section on the road is large, and the precision requirement of the high-precision map making element is high, the traditional method adopts a manual drawing method, so that the time and the labor are consumed, and the manufacturing cost is high.
Disclosure of Invention
Aiming at the technical problems in the prior art, the invention provides the automatic extraction method and the system for the virtual line segments of the traffic road in the laser point cloud, which can reduce the workload of the manufacturer and the high-precision map manufacturing cost and improve the manufacturing efficiency. And road lane information is better provided for assisting automatic driving.
According to a first aspect of the present invention, there is provided a method for automatically extracting a virtual line segment of a traffic road in a laser point cloud, including: step 1, cutting laser point cloud data of a road into blocks, acquiring a overlook projection image of the laser point cloud mass, predicting a broken line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the broken line segment;
Step 2, extracting the central line of the virtual line segment in the mask map and acquiring the width of the virtual line segment;
step 3, filtering the broken line segments in the mask map based on the structure and the position characteristics of the broken line segments;
and 4, calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after connecting, filtering and de-duplicating the virtual line segments.
On the basis of the technical scheme, the invention can also make the following improvements.
Optionally, in the step 1, the laser point cloud data is diced at a set fixed distance along the road passing direction, and a mapping relationship between the top projection image and the laser point cloud mass is reserved when the top projection image of the laser point cloud mass is obtained.
Optionally, the process of extracting the center line of the virtual line segment in the mask map and obtaining the width of the virtual line segment in step 2 includes:
And extracting the connected domain outline of the mask map and the skeleton line, extracting the minimum circumscribed rectangle and the opposite side central line of the mask map, and smoothing the skeleton line by taking the opposite side central line as a reference through a local fitting correction method to obtain the central line and the width of a smooth virtual line segment.
Optionally, the filtering the broken line segment in the mask map in step 3 based on the structural and positional characteristics of the broken line segment includes:
According to the characteristics of the area, the length-width ratio and the contour edge smoothness of the connected domain of the mask map, filtering out the non-noise connected domain, extremely unreasonable length and width and poor contour edge non-smoothing effect precision in the virtual line segment;
and filtering and reserving the target virtual line segments with shorter distances from the target virtual line segments with longer distances by taking the average value of the target virtual line segments with other distances from the local area as a reference.
Optionally, the step 3 further includes:
and connecting the parts of the virtual line segments which are closer in distance and similar in slope according to the central line slope, the distance and the direction of the extracted virtual line segments, so as to realize the connection of the cut parts of the same virtual line segment.
Optionally, the process of connecting each virtual line segment in step 4 includes:
and connecting the same broken line segments which are closer in distance and similar in slope in the three-dimensional initial point cloud coordinate system.
Optionally, the filtering and deduplicating process for each dashed segment includes:
filtering out the broken line segments with too short and too long length, and repeatedly extracting the broken line segments by adopting a non-maximum value inhibition method.
According to a second aspect of the present invention, there is provided an automatic extraction system for broken line segments of traffic roads in a laser point cloud, comprising: the device comprises a preprocessing module, a broken line segment extraction module in a top projection view, a broken line segment filtering module and a broken line segment extraction module in point cloud data;
the preprocessing module is used for cutting laser point cloud data of a road into blocks and acquiring overlook projection images of the laser points cloud mass, predicting broken line segments of the overlook projection images based on a semantic segmentation model, and extracting mask images of the broken line segments;
The broken line segment extraction module in the top projection graph is used for extracting the central line of the broken line segment in the mask graph and acquiring the width of the broken line segment;
the broken line segment filtering module is used for filtering broken line segments in the mask map based on the structure and the position characteristics of the broken line segments;
And the dotted line segment extraction module in the point cloud data is used for calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after the operations of connecting, filtering and de-duplicating the virtual line segments.
According to a third aspect of the present invention, there is provided an electronic device including a memory, and a processor for implementing the steps of the method for automatically extracting virtual line segments of a traffic road in a laser point cloud when executing a computer management program stored in the memory.
According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having stored thereon a computer-management-class program which, when executed by a processor, implements the steps of a method for automatically extracting traffic road virtual line segments in a laser point cloud.
According to the automatic extraction method, system, electronic equipment and storage medium for the virtual line segments of the traffic road in the laser point cloud, provided by the invention, the virtual line segments are firstly segmented in advance under the condition of a small number of samples for rough extraction, and then the extraction is subjected to post-treatment for fine extraction, so that the extraction precision is high and the effect is good; the laser point cloud is cut into the laser points cloud mass, and then the overlook projection image is obtained for extraction, so that on one hand, the calculation complexity can be reduced, and meanwhile, the extraction effect of the virtual line segments of the road can be improved; the broken line segment is cut into the laser point cloud mass, and the broken line segment is cut off, and meanwhile, error detection is carried out after the road is cut off, so that broken line segment connection line cutting off and error broken line segment filtering are needed, and operation connection filtering is carried out through the information such as the length, width, area, edge midpoint connection slope and the like of the road traffic broken line segment.
Drawings
FIG. 1 is a flowchart of an embodiment of a method for automatically extracting virtual line segments of traffic roads in a laser point cloud;
fig. 2 (a) is a top projection view during a line extraction process in a broken line segment in a top projection image according to an embodiment of the present invention;
fig. 2 (b) is a schematic drawing of line extraction in a broken line segment in a line extraction process in a broken line segment in a top view projection image according to an embodiment of the present invention;
FIG. 3 (a) is a top view of a top view projection image according to an embodiment of the present invention after a line in a broken line segment is extracted;
fig. 3 (b) is a segmentation mask diagram after line extraction in a broken line segment in a top-view projection image according to an embodiment of the present invention;
FIG. 4 (a) is a diagram of a mask of a broken line segment in a filtering process of a target far from a central area in a top-view projection image according to an embodiment of the present invention;
FIG. 4 (b) is a schematic diagram illustrating the extraction of a broken line segment in the filtering process of a target far from a central region in a top view projection image according to an embodiment of the present invention;
FIG. 5 (a) is a top view of a top view image of a broken line segment in the top view image according to an embodiment of the present invention;
FIG. 5 (b) is a mask diagram of a broken line segment in a connection process of broken line segment cut-off in a top view projection image according to an embodiment of the present invention;
FIG. 5 (c) is a broken line segment truncated connecting diagram during the connection of broken line segments in a top view projection image according to an embodiment of the present invention
Fig. 6 is a schematic hardware structure of one possible electronic device according to the present invention;
fig. 7 is a schematic hardware structure of a possible computer readable storage medium according to the present invention.
Detailed Description
The principles and features of the present invention are described below with reference to the drawings, the examples are illustrated for the purpose of illustrating the invention and are not to be construed as limiting the scope of the invention.
Aiming at the problems of time consumption, labor consumption and higher manufacturing cost of the traditional manual drawing method, the invention provides an automatic extraction method of traffic road virtual line segments in laser point cloud, which comprises the following steps:
Step 1, cutting laser point cloud data of a road for manufacturing a high-precision map into blocks, acquiring a overlook projection image of a laser point cloud mass, predicting a dotted line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the dotted line segment.
The semantic segmentation model can be obtained by training a marked road overlook projection sample, virtual line segments in the overlook projection image of the laser spot cloud mass are predicted based on the semantic segmentation model, and a mask map of the virtual line segments is preliminarily extracted.
And 2, extracting the central line of the broken line segment in the mask map and acquiring the width of the broken line segment.
Step3, filtering the broken line segments in the mask map based on the structure and the position characteristics of the broken line segments; and preliminarily filtering the non-broken line segment area extracted from the predicted top projection image.
And 4, calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlooking projection image to three dimensions, connecting, filtering, removing weight and the like of each virtual line segment, and submitting the virtual line segment to a high-precision map making system to finish automatic extraction operation of the road traffic virtual line segment.
When the broken line section in the road traffic is automatically extracted, the precision can not meet the high precision requirement due to the fact that the interference is large and the sample size is small by adopting a deep learning method, and the automatic extraction effect is poor due to the fact that the traditional method is poor in the general-Chinese capability, complex in road scenes of various cities in the whole country and the like. In order to reduce the high-precision map making cost, the automatic extraction method of the traffic road virtual line segments in the laser point cloud provided by the invention is characterized in that the method is divided in advance under the condition of a small number of samples to carry out rough extraction, and then the extraction is carried out with post-treatment to carry out fine extraction, so that the extraction precision is high and the effect is good.
Because the detection segmentation technology of the current 2d is mature compared with the detection segmentation technology of the current 3d, the method provided by the invention cuts the laser point cloud into the laser points cloud mass and then acquires the overlook projection image for extraction, on one hand, the calculation complexity can be reduced, and meanwhile, the extraction effect of the virtual line segments of the road can be improved.
The workload of the manufacturer and the high-precision map manufacturing cost can be reduced, and the manufacturing efficiency is improved. And road lane information is better provided for assisting automatic driving.
Example 1
Embodiment 1 provided by the present invention is an embodiment of an automatic extraction method for traffic road virtual line segments in a laser point cloud, as shown in fig. 1, which is a flowchart of an embodiment of an automatic extraction method for traffic road virtual line segments in a laser point cloud provided by the present invention, and as can be known from fig. 1, the embodiment includes:
Step 1, cutting laser point cloud data of a road into blocks, acquiring a top-view projection image of a laser point cloud mass, predicting a broken line segment of the top-view projection image based on a semantic segmentation model, and extracting a mask map of the broken line segment.
In one possible embodiment, after the traffic road laser point cloud preprocessed by the high-precision map is obtained, the laser point cloud data is diced at a set fixed distance along the road traffic direction, and the laser point cloud data is diced into a plurality of small laser points cloud mass to preserve the mapping relationship between the overhead projection image and the laser point cloud mass when the overhead projection image of the laser point cloud mass is obtained.
And 2, extracting the central line of the broken line segment in the mask map and acquiring the width of the broken line segment.
In one possible embodiment, the process of extracting the midline of the dashed segment in the mask map and obtaining the width of the dashed segment includes:
Extracting connected domain contours of a binary mask map predicted by semantic segmentation and skeleton lines, extracting the minimum circumscribed rectangle and opposite side central lines of the mask map, smoothing the skeleton lines by taking the opposite side central lines as references through a local fitting correction method, and obtaining the central lines and the widths of smooth virtual line segments.
The direct extraction of the skeleton line as the midline of the virtual line segment may cause the midline to be uneven. The invention adopts the straight line fitting of the local area and the connecting line of the two midpoints of the minimum circumscribed circle as references, and carries out smoothing treatment on the extracted skeleton line, thus obtaining the central line of the smooth virtual line segment and obtaining the width of the lane virtual line segment. The extracted centerline results are respectively a top projection view and a centerline extraction schematic diagram in the dashed line segment centerline extraction process in the top projection image provided by the embodiment of the present invention as shown in fig. 2 (a) and fig. 2 (b).
And 3, filtering the broken line segment in the mask map based on the structure and the position characteristics of the broken line segment.
In one possible embodiment, the filtering the dashed line segment in the mask map based on the structural and positional characteristics of the dashed line segment includes:
And filtering out non-noise connected domains, extremely unreasonable length and width and poor contour edge unsmooth effect precision in the virtual line segments according to the characteristics of the connected domains of the mask map, such as the area, the length-width ratio, the contour edge smoothness and the like.
And filtering the target virtual line segments with the shorter reserved distance by taking the average value of the target virtual line segments which are far away from other targets in the local area as a reference, so as to prevent the lane line virtual line segments of the opposite lane from being extracted into the current lane.
Because the road traffic of the high-precision map making relates to a wide range, the road traffic dotted line section patterns of the whole country are more various and complex, the road is blocked, the scene is complex and the like, under the condition of a small sample size, a large number of problems such as false detection, poor precision and the like can be generated, the predicted dotted line section needs to be subjected to fine extraction processing, and unreasonable dotted line sections in the overlook projection image are filtered through the information such as the length and width of the dotted line section, the position, the contour area, the contour smoothness and the like, as shown in fig. 3 (a) and 3 (b), the overlook projection image and the segmentation mask image after the centerline extraction of the dotted line section in the overlook projection image provided by the embodiment of the invention are respectively shown in fig. 3 (a) and 3 (b), and the mask extraction effect in the examples provided by fig. 3 (a) and 3 (b) has a non-dotted line section part. Fig. 4 (a) and fig. 4 (b) are schematic diagrams of a mask diagram of a broken line segment and an extraction diagram of a broken line segment in a filtering process of a target far from a central area in a top view projection image according to an embodiment of the present invention.
In a possible embodiment, step 3 further comprises:
According to the extracted information such as the central line slope, distance, direction and the like of the virtual line segments, the parts of the virtual line segments which are closer in distance and similar in slope are connected, and the cut parts of the same virtual line segments are connected together.
Due to factors such as point cloud cavities, point cloud shielding, partial extraction effect checking and the like, broken line segments extracted in the semantic segmentation of the overlook projection image are easily cut off and shielded. The parts of the virtual line segments which are closer in same distance and similar in slope are connected through the slope, distance, direction and the like of the extracted virtual line segment, namely the cut parts of the same virtual line segment are connected together. As shown in fig. 5 (a), fig. 5 (b) and fig. 5 (c), in the connection process of broken line segment cut-off in the top view projection image provided by the embodiment of the invention, the top view projection image, the broken line segment mask image and the broken line segment cut-off connection image are respectively, and mask connected domain central lines of the same broken line segment are connected together to form a broken line segment central line.
And 4, calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment to three dimensions in the overlook projection image, connecting, filtering, removing the weight and the like of each virtual line segment, and submitting the virtual line segment to a high-precision map making system.
In a possible embodiment, after the connected information of the broken line segments is back calculated into the original road laser point cloud coordinate system in step 4, because the single small block has a truncated part in the overhead projection drawing, the same broken line segment with a similar slope and a shorter distance in the three-dimensional initial point cloud coordinate system needs to be connected, and the truncated broken line segments are connected.
The process of filtering and de-duplicating each virtual line segment after connection comprises the following steps:
and filtering out the short and long broken line segments, and reserving the broken line segments with proper lengths. And then filtering and repeatedly extracting the virtual line segments by adopting a non-maximum suppression method.
Example 2
The embodiment 2 provided by the invention is an embodiment of an automatic extraction system for a broken line segment of a traffic road in a laser point cloud, which comprises: the device comprises a preprocessing module, a broken line segment extraction module in a top projection view, a broken line segment filtering module and a broken line segment extraction module in point cloud data.
The preprocessing module is used for cutting laser point cloud data of the road into blocks and acquiring overlook projection images of the laser points cloud mass, predicting broken line segments of the overlook projection images based on the semantic segmentation model, and extracting mask images of the broken line segments.
And the broken line segment extraction module in the top projection view is used for extracting the central line of the broken line segment in the mask view and acquiring the width of the broken line segment.
And the virtual line segment filtering module is used for filtering the virtual line segment in the mask map based on the structure and the position characteristics of the virtual line segment.
The point cloud data comprises a point cloud data acquisition module, a point cloud data extraction module and a high-precision map making system, wherein the point cloud data acquisition module is used for acquiring the point cloud data of a virtual line segment, and the point cloud data comprises a point cloud data acquisition module and a point cloud data acquisition module.
It can be understood that the automatic extraction system for the traffic road broken line segment in the laser point cloud provided by the invention corresponds to the automatic extraction method for the traffic road broken line segment in the laser point cloud provided by the foregoing embodiments, and the relevant technical features of the automatic extraction system for the traffic road broken line segment in the laser point cloud can refer to the relevant technical features of the automatic extraction method for the traffic road broken line segment in the laser point cloud, which are not described herein.
Referring to fig. 6, fig. 6 is a schematic diagram of an embodiment of an electronic device according to an embodiment of the invention. As shown in fig. 6, an embodiment of the present invention provides an electronic device, including a memory 1310, a processor 1320, and a computer program 1311 stored on the memory 1320 and executable on the processor 1320, wherein the processor 1320 executes the computer program 1311 to implement the following steps: cutting laser point cloud data of a road into blocks, acquiring a overlook projection image of a laser point cloud mass, predicting a dotted line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the dotted line segment; extracting the center line of a broken line segment in the mask map and acquiring the width of the broken line segment; filtering the broken line segments in the mask map based on the structural and position characteristics of the broken line segments; and calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after connecting, filtering and de-duplicating the virtual line segments.
Referring to fig. 7, fig. 7 is a schematic diagram of an embodiment of a computer readable storage medium according to the present invention. As shown in fig. 7, the present embodiment provides a computer-readable storage medium 1400 having stored thereon a computer program 1411, which computer program 1411, when executed by a processor, performs the steps of:
Cutting laser point cloud data of a road into blocks, acquiring a overlook projection image of a laser point cloud mass, predicting a dotted line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the dotted line segment; extracting the center line of a broken line segment in the mask map and acquiring the width of the broken line segment; filtering the broken line segments in the mask map based on the structural and position characteristics of the broken line segments; and calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after connecting, filtering and de-duplicating the virtual line segments.
According to the automatic extraction method, system and storage medium for the virtual line segments of the traffic road in the laser point cloud, provided by the embodiment of the invention, the rough extraction is carried out by dividing in advance under the condition of a small amount of samples, and then the post-treatment is carried out on the extraction, so that the extraction precision is high and the effect is good; the laser point cloud is cut into the laser points cloud mass, and then the overlook projection image is obtained for extraction, so that on one hand, the calculation complexity can be reduced, and meanwhile, the extraction effect of the virtual line segments of the road can be improved; the broken line segment is cut into the laser point cloud mass, and the broken line segment is cut off, and meanwhile, error detection is carried out after the road is cut off, so that broken line segment connection line cutting off and error broken line segment filtering are needed, and operation connection filtering is carried out through the information such as the length, width, area, edge midpoint connection slope and the like of the road traffic broken line segment.
In the foregoing embodiments, the descriptions of the embodiments are focused on, and for those portions of one embodiment that are not described in detail, reference may be made to the related descriptions of other embodiments.
It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
Claims (4)
1. The automatic extraction method for the virtual line segments of the traffic roads in the laser point cloud is characterized by comprising the following steps of:
Step 1, cutting laser point cloud data of a road into blocks, acquiring a overlook projection image of the laser point cloud mass, predicting a broken line segment of the overlook projection image based on a semantic segmentation model, and extracting a mask map of the broken line segment;
Step 2, extracting the central line of the virtual line segment in the mask map and acquiring the width of the virtual line segment;
step 3, filtering the broken line segments in the mask map based on the structure and the position characteristics of the broken line segments;
Step4, calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after connecting, filtering and de-duplicating the virtual line segments;
the filtering the broken line segment in the mask map based on the structure and the position characteristics of the broken line segment in the step 3 includes:
According to the characteristics of the area, the length-width ratio and the contour edge smoothness of the connected domain of the mask map, filtering out the non-noise connected domain, extremely unreasonable length and width and poor contour edge non-smoothing effect precision in the virtual line segment;
Taking the average value of virtual line segments which are far away from other targets in the local area as a reference, and filtering and reserving the virtual line segments of the targets which are far away and close to each other;
the said step3 and then further comprises:
Connecting the parts of the virtual line segments which are closer in distance and similar in slope according to the central line slope, the distance and the direction of the extracted virtual line segments, so as to realize the connection of the cut parts of the same virtual line segment;
The process of connecting each virtual line segment in the step 4 includes:
connecting the same broken line segments which are closer in distance and similar in slope in the three-dimensional initial point cloud coordinate system;
the process of filtering and de-duplicating each broken line segment includes:
Filtering out the broken line segments with too short and too long length, and repeatedly extracting the broken line segments by adopting a non-maximum value inhibition method;
In the step 1, the laser point cloud data is diced at a set fixed distance along the road passing direction, and the mapping relation between the overlooking projection image and the laser point cloud mass is reserved when the overlooking projection image of the laser point cloud mass is obtained;
The process of extracting the center line of the virtual line segment in the mask map and obtaining the width of the virtual line segment in the step 2 includes:
And extracting the connected domain outline of the mask map and the skeleton line, extracting the minimum circumscribed rectangle and the opposite side central line of the mask map, and smoothing the skeleton line by taking the opposite side central line as a reference through a local fitting correction method to obtain the central line and the width of a smooth virtual line segment.
2. An automatic extraction system for a traffic road broken line section in a laser point cloud is characterized by comprising: the device comprises a preprocessing module, a broken line segment extraction module in a top projection view, a broken line segment filtering module and a broken line segment extraction module in point cloud data;
the preprocessing module is used for cutting laser point cloud data of a road into blocks and acquiring overlook projection images of the laser points cloud mass, predicting broken line segments of the overlook projection images based on a semantic segmentation model, and extracting mask images of the broken line segments;
The broken line segment extraction module in the top projection graph is used for extracting the central line of the broken line segment in the mask graph and acquiring the width of the broken line segment;
the broken line segment filtering module is used for filtering broken line segments in the mask map based on the structure and the position characteristics of the broken line segments;
The dotted line segment extraction module in the point cloud data is used for calculating three-dimensional coordinate information of each virtual line segment according to the mapping relation from the central line and the width of the virtual line segment in the overlook projection image to three dimensions, and submitting the three-dimensional coordinate information to a high-precision map making system after the operations of connecting, filtering and de-duplicating the virtual line segments;
The process of the broken line segment filtering module for filtering the broken line segment in the mask map based on the structure and the position characteristics of the broken line segment comprises the following steps:
According to the characteristics of the area, the length-width ratio and the contour edge smoothness of the connected domain of the mask map, filtering out the non-noise connected domain, extremely unreasonable length and width and poor contour edge non-smoothing effect precision in the virtual line segment;
Taking the average value of virtual line segments which are far away from other targets in the local area as a reference, and filtering and reserving the virtual line segments of the targets which are far away and close to each other;
The extracting of the dotted line segment in the point cloud data further comprises:
Connecting the parts of the virtual line segments which are closer in distance and similar in slope according to the central line slope, the distance and the direction of the extracted virtual line segments, so as to realize the connection of the cut parts of the same virtual line segment;
The process for extracting the broken line segments in the point cloud data and connecting the broken line segments comprises the following steps:
connecting the same broken line segments which are closer in distance and similar in slope in the three-dimensional initial point cloud coordinate system;
the process of filtering and de-duplicating each broken line segment includes:
Filtering out the broken line segments with too short and too long length, and repeatedly extracting the broken line segments by adopting a non-maximum value inhibition method;
Dicing the laser point cloud data along the road passing direction by a set fixed distance in the preprocessing module, and reserving a mapping relation between the overlooking projection image and the laser point cloud mass when acquiring the overlooking projection image of the laser point cloud mass;
the process of extracting the center line of the virtual line segment in the mask graph and obtaining the width of the virtual line segment in the top projection graph in the dashed line segment extracting module comprises the following steps:
And extracting the connected domain outline of the mask map and the skeleton line, extracting the minimum circumscribed rectangle and the opposite side central line of the mask map, and smoothing the skeleton line by taking the opposite side central line as a reference through a local fitting correction method to obtain the central line and the width of a smooth virtual line segment.
3. An electronic device, comprising a memory and a processor, wherein the processor is configured to implement the method for automatically extracting a traffic road segment in a laser point cloud according to claim 1 when executing a computer management program stored in the memory.
4. A computer-readable storage medium, having stored thereon a computer-management-class program which, when executed by a processor, implements the steps of the method for automatically extracting traffic road segments in a laser point cloud as claimed in claim 1.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110973671.6A CN113807193B (en) | 2021-08-23 | 2021-08-23 | Automatic extraction method and system for traffic road virtual line segments in laser point cloud |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110973671.6A CN113807193B (en) | 2021-08-23 | 2021-08-23 | Automatic extraction method and system for traffic road virtual line segments in laser point cloud |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113807193A CN113807193A (en) | 2021-12-17 |
| CN113807193B true CN113807193B (en) | 2024-08-02 |
Family
ID=78893988
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110973671.6A Active CN113807193B (en) | 2021-08-23 | 2021-08-23 | Automatic extraction method and system for traffic road virtual line segments in laser point cloud |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113807193B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115752432B (en) * | 2022-10-29 | 2025-06-17 | 武汉中海庭数据技术有限公司 | Automatic extraction method and system of dashed lane lines in road traffic images collected by UAV |
| EP4475093A1 (en) * | 2023-06-09 | 2024-12-11 | Volkswagen Aktiengesellschaft | Method for identifying sections belonging to one lane boundary of a lane of a road |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110705577A (en) * | 2019-09-29 | 2020-01-17 | 武汉中海庭数据技术有限公司 | Laser point cloud lane line extraction method |
| CN112632201A (en) * | 2020-12-12 | 2021-04-09 | 武汉中海庭数据技术有限公司 | Traffic road topology network construction method and device |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004246554A (en) * | 2003-02-13 | 2004-09-02 | Zenrin Co Ltd | Road data generation method |
| CN104809449B (en) * | 2015-05-14 | 2018-09-21 | 重庆大学 | Track dotted line line of demarcation automatic testing method suitable for highway video monitoring system |
| CN107169464B (en) * | 2017-05-25 | 2019-04-09 | 中国农业科学院农业资源与农业区划研究所 | A road boundary detection method based on laser point cloud |
| CN108981741B (en) * | 2018-08-23 | 2021-02-05 | 武汉中海庭数据技术有限公司 | Path planning device and method based on high-precision map |
| CN110287904B (en) * | 2019-06-27 | 2021-07-16 | 武汉中海庭数据技术有限公司 | A method, device and storage medium for lane line extraction based on crowdsourcing data |
| CN111881790B (en) * | 2020-07-14 | 2023-04-28 | 武汉中海庭数据技术有限公司 | Automatic extraction method and device for road crosswalk in high-precision map making |
| CN112258519B (en) * | 2020-10-12 | 2022-09-20 | 武汉中海庭数据技术有限公司 | Automatic extraction method and device for way-giving line of road in high-precision map making |
| CN112434707B (en) * | 2020-11-14 | 2022-09-09 | 武汉中海庭数据技术有限公司 | Traffic intersection RPP point automatic extraction method and device |
-
2021
- 2021-08-23 CN CN202110973671.6A patent/CN113807193B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110705577A (en) * | 2019-09-29 | 2020-01-17 | 武汉中海庭数据技术有限公司 | Laser point cloud lane line extraction method |
| CN112632201A (en) * | 2020-12-12 | 2021-04-09 | 武汉中海庭数据技术有限公司 | Traffic road topology network construction method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113807193A (en) | 2021-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112200171B (en) | Road point cloud extraction method based on scanning lines | |
| EP3171292B1 (en) | Driving lane data processing method, device, storage medium and apparatus | |
| CN105426864B (en) | One kind being based on the matched Multi-lane Lines Detection method of equidistant marginal point | |
| JP6561199B2 (en) | Urban road recognition method, apparatus, storage medium and equipment based on laser point cloud | |
| CN111881790B (en) | Automatic extraction method and device for road crosswalk in high-precision map making | |
| CN111192311B (en) | Automatic extraction method and device for longitudinal deceleration marked line in high-precision map making | |
| CN110749329B (en) | A lane-level topology construction method and device based on structured road | |
| CN113807193B (en) | Automatic extraction method and system for traffic road virtual line segments in laser point cloud | |
| WO2023216470A1 (en) | Drivable area detection method and apparatus, and device | |
| CN111897906B (en) | Method, apparatus, device and storage medium for processing map data | |
| CN109839119B (en) | Method and device for acquiring bridge floor area of bridge of cross-road bridge | |
| CN112257772B (en) | Road increase and decrease interval segmentation method and device, electronic equipment and storage medium | |
| CN114407919B (en) | Collision detection method and system based on automatic driving | |
| CN115752432B (en) | Automatic extraction method and system of dashed lane lines in road traffic images collected by UAV | |
| CN114779206A (en) | Method and device for identifying road boundary, storage medium and equipment | |
| CN113345015A (en) | Package position detection method, device and equipment and readable storage medium | |
| CN112435336B (en) | Curve type identification method and device, electronic equipment and storage medium | |
| CN117289297A (en) | Point cloud dynamic object searching and removing method based on concentric circle sector division | |
| CN114435402A (en) | Lane line smoothing method and device and electronic equipment | |
| CN113139027B (en) | Method, device and equipment for splitting job task and storage medium | |
| CN118052945A (en) | Road scene modeling method, high-precision map updating method, device and equipment | |
| CN116452616A (en) | Intersecting contour point co-edge method based on image segmentation | |
| CN116052126A (en) | Method and device for detecting drivable area, vehicle and storage medium | |
| CN112258602A (en) | Stop line generation method and device, electronic equipment and storage medium | |
| JP2021018604A (en) | Image processing apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |