Method, device and electronic device for determining quality of crowdsourced maps

Abstract

The embodiment of the present application provides a method, device and electronic device for determining the quality of a crowdsourced map. The method for determining the quality of a crowdsourced map includes: obtaining a crowdsourced map and a true map, wherein the crowdsourced map and the true map are maps of the same road section; semantically matching the crowdsourced map and the true map to obtain a target accuracy value; and determining the quality of the crowdsourced map according to the target accuracy value. The present application determines the quality of the crowdsourced map by matching the semantics of the crowdsourced map and the true map, thereby avoiding manual measurement of the difference between the crowdsourced map and the true map, thereby improving the efficiency of determining the quality of the crowdsourced map, and the determined quality has the technical effect of small error and low cost.

Description

Technical Field

[0001] The present application relates to the technical field of navigation map processing, and in particular to a method, device and electronic device for determining the quality of a crowdsourced map. Background Art

[0002] Crowdsourced maps are generated by road condition information uploaded and shared by end users during driving. Crowdsourced map quality assessment is a means of measuring the effectiveness of the generated crowdsourced maps and the iteration algorithm used in generating crowdsourced maps. Crowdsourced map quality assessment is an indispensable process in crowdsourced map generation technology. In addition, crowdsourced map quality assessment can also serve as important reference data for downstream autonomous driving positioning and navigation.

[0003] Currently, the quality assessment of crowdsourced maps is done manually. Specifically, manually measuring the difference between crowdsourced maps and true maps is time-consuming, error-prone, and costly. Summary of the invention

[0004] The present application provides a method, device and electronic device for determining the quality of a crowdsourced map, which are used to solve the problems of long time consumption, high error and high cost in the existing crowdsourced map quality assessment methods.

[0005] In a first aspect, an embodiment of the present application provides a method for determining the quality of a crowdsourced map, comprising: obtaining a crowdsourced map and a true map, the crowdsourced map and the true map being maps of the same road section; semantically matching the crowdsourced map and the true map to obtain a target accuracy value; and determining the quality of the crowdsourced map based on the target accuracy value.

[0006] In one embodiment of the present application, a crowdsourced map and a true map are semantically matched to obtain a target accuracy value, including: dividing the crowdsourced map into multiple first sub-segments according to trajectory information in the crowdsourced map, and dividing the true map into second sub-segments corresponding to each first sub-segment; for each first sub-segment, semantically matching the first sub-segment with its corresponding second sub-segment to obtain a matching accuracy value corresponding to each first sub-segment; and obtaining a target accuracy value according to the matching accuracy values ​​of each first sub-segment.

[0007] In another embodiment of the present application, for each first sub-segment, semantic matching is performed on the first sub-segment and its corresponding second sub-segment to obtain a matching accuracy value corresponding to each first sub-segment, including: for each first sub-segment, determining a reference point in the first sub-segment; using the current posture information of the reference point as an optimization variable; determining an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment; determining a relational constraint edge formed by the reference point and a matching point in the corresponding second sub-segment, the reference point and the corresponding matching point representing semantic elements at the same position; optimizing the current posture information based on the adjacent constraint edge and the relational constraint edge to obtain the target posture information of the reference point; optimizing the posture information of the first semantic element based on the target posture information; matching the optimized first semantic element with the corresponding second semantic element at the same position to obtain a matching accuracy value corresponding to the first sub-segment.

[0008] In another embodiment of the present application, the optimized first semantic element and the second semantic element corresponding to the same position are matched to obtain a matching accuracy value of the corresponding first sub-road section, including: for each optimized first semantic element, sampling the first semantic element to obtain at least one first discrete point corresponding to the first semantic element; for each second semantic element, sampling the second semantic element to obtain at least one second discrete point corresponding to the second semantic element; for each first sub-road section, matching the first discrete point of the first sub-road section with the corresponding second discrete point to obtain a matching accuracy value of the corresponding first sub-road section.

[0009] In another embodiment of the present application, for each optimized first semantic element, the first semantic element is sampled to obtain at least one first discrete point corresponding to the first semantic element, including: if the first semantic element is continuous, the first semantic element is continuously sampled to obtain multiple sampling points as the first discrete points corresponding to the first semantic element; if the first semantic element is non-continuous, the feature point of the first semantic element is determined to be the first discrete point corresponding to the first semantic element.

[0010] In another embodiment of the present application, the first discrete points of the first sub-segment are matched with the corresponding second discrete points to obtain a matching accuracy value of the corresponding first sub-segment, including: for each first discrete point, determining the second discrete point that is closest to the first discrete point; if the distance between the first discrete point and the second discrete point that is closest to it is less than or equal to a first threshold, determining that the first discrete point and its closest second discrete point have a matching relationship; for each first sub-segment, determining a first number of first discrete points that have a matching relationship, and determining the total number of second discrete points in the corresponding second sub-segment; based on the first number and the total number, determining a completeness rate and a redundancy rate corresponding to the first sub-segment, the completeness rate being a first ratio of the first number to the total number, and the redundancy rate being a second ratio of a first difference between the total number and the first number to the total number; determining the completeness rate and the redundancy rate as the matching accuracy value of the corresponding first sub-segment.

[0011] In another embodiment of the present application, the first discrete point and the corresponding second discrete point are matched to obtain a matching accuracy value of the corresponding first sub-segment, including: for each first discrete point, determining a first distance between the first discrete point and each other first discrete point in the corresponding first sub-segment; determining a second distance between the second discrete point corresponding to the first discrete point and each other second discrete point in the corresponding second sub-segment; and determining a relative accuracy value of the corresponding first sub-segment as the matching accuracy value based on the first distance and the second distance.

[0012] In another embodiment of the present application, a relative accuracy value of the corresponding first sub-segment is determined as a matching accuracy value based on the first distance and the second distance, including: for each first discrete point, a first average value of the relative difference is determined based on the relative differences between the corresponding first distances and the second distances; for each first semantic element, a semantic accuracy value of the corresponding first semantic element is determined based on the first average value corresponding to each first discrete point in the first semantic element and the first number of first discrete points in the first semantic element; for each first sub-segment, a relative accuracy value of the corresponding first sub-segment is determined as a matching accuracy value based on the semantic accuracy values ​​of each first semantic element in the first sub-segment and the second number of first semantic elements in the first sub-segment.

[0013] In another embodiment of the present application, the first discrete point and the corresponding second discrete point are matched to obtain a matching accuracy value of the corresponding first sub-road section, including: for each first sub-road section, determining the absolute value of the position difference between each first discrete point in the first sub-road section and the corresponding second discrete point; determining a second average value of the absolute value as the absolute accuracy value of the first sub-road section; and using the absolute accuracy value as the matching accuracy value.

[0014] In a second aspect, the present application provides a crowdsourced map quality determination device, comprising:

[0015] An acquisition module is used to acquire a crowdsourced map and a true map, where the crowdsourced map and the true map are maps of the same road section;

[0016] The matching module is used to semantically match the crowdsourced map and the true map to obtain the target accuracy value;

[0017] The determination module is used to determine the quality of the crowdsourced map according to the target accuracy value.

[0018] In another embodiment of the present application, the matching module includes:

[0019] A segmentation unit, configured to segment the crowdsourced map into a plurality of first sub-segments according to the trajectory information in the crowdsourced map, and segment the true value map into second sub-segments corresponding to each first sub-segment;

[0020] A matching unit, configured to perform semantic matching on each first sub-segment and its corresponding second sub-segment to obtain a matching accuracy value corresponding to each first sub-segment;

[0021] The determination unit is used to obtain a target accuracy value according to the matching accuracy values ​​of each first sub-road section.

[0022] In another embodiment of the present application, the matching unit is specifically used to: for each first sub-segment, determine a reference point in the first sub-segment; use the current posture information of the reference point as an optimization variable; determine an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment; determine a relationship constraint edge formed by the reference point and a matching point in the corresponding second sub-segment, the reference point and the corresponding matching point representing semantic elements at the same position; optimize the current posture information based on the adjacent constraint edge and the relationship constraint edge to obtain the target posture information of the reference point; optimize the posture information of the first semantic element based on the target posture information; match the optimized first semantic element with the corresponding second semantic element at the same position to obtain a matching accuracy value corresponding to the first sub-segment.

[0023] In another embodiment of the present application, when the matching unit is used to match the optimized first semantic element with the second semantic element corresponding to the same position to obtain the matching accuracy value of the corresponding first sub-segment, it is specifically used to: for each optimized first semantic element, sample the first semantic element to obtain at least one first discrete point corresponding to the first semantic element; for each second semantic element, sample the second semantic element to obtain at least one second discrete point corresponding to the second semantic element; for each first sub-segment, match the first discrete point of the first sub-segment with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-segment.

[0024] In another embodiment of the present application, when the matching unit is used to sample each optimized first semantic element to obtain at least one first discrete point corresponding to the first semantic element, it is specifically used to: if the first semantic element is continuous, continuously sample the first semantic element to obtain multiple sampling points as the first discrete points corresponding to the first semantic element; if the first semantic element is non-continuous, determine the feature point of the first semantic element as the first discrete point corresponding to the first semantic element.

[0025] In another embodiment of the present application, the matching unit is used to match the first discrete point of the first sub-section with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-section, specifically for: for each first discrete point, determining the second discrete point that is closest to the first discrete point; if the distance between the first discrete point and the second discrete point closest to it is less than or equal to the first threshold, then determining that the first discrete point and its closest second discrete point have a matching relationship; for each first sub-section, determining the first number of first discrete points with a matching relationship, and determining the total number of second discrete points in the corresponding second sub-section; based on the first number and the total number, determining the completeness rate and redundancy rate corresponding to the first sub-section, the completeness rate is a first ratio of the first number to the total number, and the redundancy rate is a second ratio of the first difference between the total number and the first number to the total number; determining the completeness rate and the redundancy rate as the matching accuracy value of the corresponding first sub-section.

[0026] In another embodiment of the present application, the matching unit is used to match the first discrete point with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-segment, and is specifically used to: for each first discrete point, determine the first distance between the first discrete point and each other first discrete point in the corresponding first sub-segment; determine the second distance between the second discrete point corresponding to the first discrete point and each other second discrete point in the corresponding second sub-segment; and determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value based on the first distance and the second distance.

[0027] In another embodiment of the present application, the matching unit is used to determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value based on the first distance and the second distance, and is specifically used to: for each first discrete point, determine the first average value of the relative difference according to the relative differences between the corresponding first distances and the second distances; for each first semantic element, determine the semantic accuracy value of the corresponding first semantic element according to the first average value corresponding to each first discrete point in the first semantic element, and the first number of first discrete points in the first semantic element; for each first sub-segment, determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value according to the semantic accuracy value of each first semantic element in the first sub-segment and the second number of first semantic elements in the first sub-segment.

[0028] In another embodiment of the present application, the matching unit is used to match the first discrete point and the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-section, and is specifically used to: for each first sub-section, determine the absolute value of the position difference between each first discrete point in the first sub-section and the corresponding second discrete point; determine the second average value of the absolute value as the absolute accuracy value of the first sub-section; and use the absolute accuracy value as the matching accuracy value.

[0029] In a third aspect, an embodiment of the present application provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to execute the crowdsourced map quality determination method of any one of the above-mentioned first aspects.

[0030] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer execution instructions are stored. When a processor executes the computer execution instructions, a method for determining the quality of a crowdsourced map as described in any one of the first aspects above is implemented.

[0031] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the method for determining the quality of a crowdsourced map according to any one of the first aspects above.

[0032] The embodiment of the present application provides a method, device and electronic device for determining the quality of a crowdsourced map. The method for determining the quality of a crowdsourced map includes: obtaining a crowdsourced map and a true map, wherein the crowdsourced map and the true map are maps of the same road section; semantically matching the crowdsourced map and the true map to obtain a target accuracy value; and determining the quality of the crowdsourced map according to the target accuracy value. The present application determines the quality of the crowdsourced map by matching the semantics of the crowdsourced map and the true map, thereby avoiding manual measurement of the difference between the crowdsourced map and the true map, thereby improving the efficiency of determining the quality of the crowdsourced map, and the determined quality has the technical effect of small error and low cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0034] Figure 1A schematic diagram of a scenario of a method for determining the quality of a crowdsourced map provided in an embodiment of the present application;

[0035] Figure 2 A flowchart of the steps of a method for determining the quality of a crowdsourced map provided in an embodiment of the present application;

[0036] Figure 3 A flowchart of another method for determining the quality of a crowdsourced map provided in an embodiment of the present application;

[0037] Figure 4 A schematic diagram of segmenting another crowdsourced map and a true map provided in an embodiment of the present application;

[0038] Figure 5 A flowchart of the steps for determining a matching accuracy value provided in an embodiment of the present application;

[0039] Figure 6 A schematic diagram of optimizing the first sub-section provided in an embodiment of the present application;

[0040] Figure 7 A schematic diagram of sampling a first semantic element provided in an embodiment of the present application;

[0041] Figure 8 A schematic diagram of sampling a second semantic element provided in an embodiment of the present application;

[0042] Fig. 9 A schematic diagram of matching a first discrete point with a second discrete point provided in an embodiment of the present application;

[0043] Fig.10 A schematic diagram of determining the correspondence between the redundancy rate and the completeness rate provided in an embodiment of the present application;

[0044] Fig.11 A structural block diagram of a crowdsourced map quality determination device provided in an embodiment of the present application;

[0045] Fig.12 A schematic diagram of the hardware structure of an electronic device provided in one embodiment of the present application.

[0046] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0048] The terms "first", "second", "third", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the terms used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0049] In addition, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus that includes a series of steps or elements is not necessarily limited to those steps or elements explicitly listed, but may include other steps or elements not explicitly listed or inherent to such process, method, product, or apparatus.

[0050] For example, Figure 1 A schematic diagram of a scenario of a method for determining the quality of a crowdsourced map provided in an embodiment of the present application, such as Figure 1 The process includes: when a vehicle 11 is traveling on a road 12, a camera installed in the vehicle 11 will collect images of various elements on the road 12, such as a zebra crossing 121, a dividing line 122, a street lamp 123, and buildings beside the road 12. The collected images are uploaded to a server, and the server performs iterative crowdsourcing processing on these images to output a crowdsourcing map. The quality of the output crowdsourcing map needs to be evaluated to determine the effectiveness of the algorithm used for iterative crowdsourcing processing of images.

[0051] Based on the above application scenarios, the present application provides a method for determining the quality of crowdsourced maps, which makes full use of crowdsourced maps and true maps under the same road section, and extracts the first semantic element in the crowdsourced map, extracts the second semantic element in the true map, and matches them through the iterative closest point algorithm (ICP) of the semantic elements. The quality of the crowdsourced map is determined by determining at least one of the corresponding redundancy rate, completeness rate, relative accuracy and absolute accuracy. The present application can automatically determine the quality of the crowdsourced map of any road section covered by the true map after the crowdsourcing vehicle and the cloud system complete the mapping of the specific road section. Compared with the existing method of manually determining the quality of crowdsourced maps, the present application does not require repeated manual measurements and repetitive time investment. The automatic determination of the quality of crowdsourced maps in different scenarios does not require additional investment, which improves the efficiency of crowdsourced map quality determination, and the correctness of quality determination is higher than manual measurement. The present application provides quality indicators for crowdsourced mapping, and provides an important reference direction for the improvement of crowdsourced mapping, so that it can better meet the application of autonomous driving.

[0052] The technical solution of the present application is described in detail below through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0053] Figure 2 A flow chart of a method for determining the quality of a crowdsourced map provided in an embodiment of the present application. The method for determining the quality of a crowdsourced map specifically comprises the following steps:

[0054] S201, obtaining a crowdsourced map and a true map.

[0055] The crowdsourced map and the true map are maps of the same road section. Specifically, the crowdsourced map is generated by the road condition information uploaded and shared by the end users during driving. The true map is a high-precision professional map collected and produced by a professional collection vehicle. This application compares the crowdsourced map and the true map of the same road section to determine the quality of the crowdsourced map.

[0056] S202, semantically matching the crowdsourced map and the true map to obtain a target accuracy value.

[0057] The crowdsourced map includes a plurality of first semantic elements, and the true value map includes a plurality of second semantic elements. The semantic matching between the crowdsourced map and the true value map is essentially matching the first semantic elements with the second semantic elements.

[0058] Specifically, the semantic elements are the shape, size, category and location information of each map element on the corresponding map. The map elements correspond to lane lines, lane markings, poles, stop lines, aerial markings, traffic signs, parking spaces or zebra crossings, etc. Each semantic element is composed of several point clusters.

[0059] Furthermore, the target accuracy value may be represented by one or more of redundancy rate, completeness rate, relative accuracy or absolute accuracy. Specifically, a detailed description is given below.

[0060] S203: Determine the quality of the crowdsourced map according to the target accuracy value.

[0061] In this application, the quality level can be set, and the target accuracy value corresponds to the quality level. For example, if the target accuracy value is less than the first threshold, the quality level of the crowdsourced map is set to primary. If the target accuracy value is greater than or equal to the first threshold and less than the second threshold, the quality level of the crowdsourced map is set to intermediate. If the target accuracy value is greater than or equal to the second threshold, the quality level of the crowdsourced map is set to advanced.

[0062] The embodiment of the present application provides a method, device and electronic device for determining the quality of a crowdsourced map. The method for determining the quality of a crowdsourced map includes: obtaining a crowdsourced map and a true map, wherein the crowdsourced map and the true map are maps of the same road section; semantically matching the crowdsourced map and the true map to obtain a target accuracy value; and determining the quality of the crowdsourced map according to the target accuracy value. The present application determines the quality of the crowdsourced map by matching the semantics of the crowdsourced map and the true map, thereby avoiding manual measurement of the difference between the crowdsourced map and the true map, thereby improving the efficiency of determining the quality of the crowdsourced map, and the determined quality has the technical effect of small error and low cost.

[0063] Reference Figure 3 , is a flowchart of a method for determining the quality of a crowdsourced map provided by another embodiment of the present application, which specifically includes the following steps:

[0064] S301, obtaining a crowdsourced map and a true map.

[0065] The specific implementation process of this step refers to S201 and will not be repeated here.

[0066] S302: Divide the crowdsourced map into a plurality of first sub-segments according to the trajectory information in the crowdsourced map, and divide the true value map into second sub-segments corresponding to each first sub-segment.

[0067] In this application, the crowdsourced map includes trajectory information. The trajectory information represents the path taken by the vehicle during driving, which is the driving mode of the vehicle when collecting the image of the crowdsourced map. Figure 4, including a crowdsourced map 41 and a true map 42. The dotted line A represents the trajectory information, and the arrow of the dotted line A refers to the driving direction of the vehicle during driving.

[0068] Specifically, when the crowdsourced map 41 and the true value map 42 are segmented, the crowdsourced map 41 is segmented into a plurality of first sub-segments L1 according to a random starting point B and a random step length in the crowdsourced map 41 along the driving direction corresponding to the trajectory information. The length of each first sub-segment L1 is the random step length. In addition, in the true value map 42, a point at the same position as the starting point B is selected as the starting point, and the same length as the first sub-segment L1 is selected to segment the true value map 42. The true value map 42 is segmented into a plurality of second sub-segments L2 of the same step length. The first sub-segments and the second sub-segments correspond one to one.

[0069] In this application, since the crowdsourced map may be from a camera with lower pixels, the clarity of the crowdsourced map may be lower, and it is impossible to directly match the crowdsourced map with the true map. After the crowdsourced map and the true map are segmented and matched in sections, the accuracy of the final match can be improved. In addition, the use of random segmentation starting points and step sizes can keep the first and second sub-segments after segmentation random, and will not affect the target accuracy value of the final calculation.

[0070] S303: For each first sub-segment, semantically match the first sub-segment with its corresponding second sub-segment to obtain a matching accuracy value corresponding to each first sub-segment.

[0071] In the present application, semantic matching is first performed on each first sub-segment and the corresponding second sub-segment to obtain a corresponding matching accuracy value; then, the multiple matching accuracy values ​​are averaged to obtain the target accuracy value of the crowdsourcing map.

[0072] Specifically, refer to Figure 5 , S303 includes the following steps:

[0073] S501: For each first sub-section, determine a reference point in the first sub-section.

[0074] The reference point may be any semantic element in the first sub-segment.

[0075] Among them, refer to Figure 6 , including a crowdsourced map 61 and a true map 62. 611 in the crowdsourced map respectively represent reference points in the first sub-segment. Specifically, one first sub-segment corresponds to one reference point.

[0076] S502, using the current posture information of the reference point as an optimization variable.

[0077] The current position information of the reference point includes: the position information and the posture information of the reference point. The posture information refers to the direction.

[0078] Specifically, crowdsourced maps are formed by images collected by end users during driving, so the corresponding current pose information of each semantic element in the crowdsourced map may not be consistent with the true map. For example, the coordinates of the semantic element representing a street lamp in the crowdsourced map may be (30,60), while the coordinates corresponding to the semantic element of the street lamp in the true map may be (60,90). It can be seen that it is meaningless to match the semantic element of the street lamp in the crowdsourced map with the semantic element corresponding to the street lamp in the true map. It is necessary to optimize and align the pose information of the street lamp in the crowdsourced map so that it can match the corresponding semantic element in the true map.

[0079] S503: Determine an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment.

[0080] Specifically, refer to Figure 6 , the adjacent constraint edge 612 may be an edge formed between reference points in adjacent sub-segments.

[0081] S504: Determine a relational constraint edge formed between the reference point and the corresponding matching point in the second sub-segment, where the reference point and the corresponding matching point represent semantic elements at the same position.

[0082] Among them, refer to Figure 6 , the relationship constraint edge 63 is the edge formed between the reference point 611 and the corresponding matching point 621.

[0083] S505, optimizing the current pose information based on the adjacent constraint edges and the relationship constraint edges to obtain the target pose information of the reference point.

[0084] Specifically, the adjacent constraint edges, relationship constraint edges and current pose information are optimized through GTSAM (Georgia Tech Smoothing and Mapping) to obtain the target pose information of the corresponding reference point.

[0085] Among them, the target posture information refers to the adjustment of the current posture information of the reference point, specifically adjusting the position information and posture information of the reference point, so that the position information of the reference point is the same as the position information of the matching point, and the posture information of the reference point is the same as the posture information of the matching point.

[0086] S506: Optimize the posture information of the first semantic element according to the target posture information.

[0087] Among them, since the positional relationship between the reference point and other first semantic elements in the first sub-segment is fixed, the posture information of other first semantic elements can be optimized according to the target posture information of the reference point.

[0088] S507: Match the optimized first semantic element with the second semantic element corresponding to the same position to obtain a matching accuracy value corresponding to the first sub-road segment.

[0089] In the present application, after optimizing the position information of each first semantic element, the crowdsourced map can be matched with the true value map more effectively.

[0090] In addition, in the present application, if a first semantic element in a crowdsourced map does not have a corresponding second semantic element in a true value map, it means that the first semantic element is a redundant semantic element, and the first redundant semantic element is removed.

[0091] Specifically, the optimized first semantic element and the second semantic element corresponding to the same position are matched to obtain the matching accuracy value of the corresponding first sub-road section, including: for each optimized first semantic element, sampling the first semantic element to obtain at least one first discrete point corresponding to the first semantic element; for each second semantic element, sampling the second semantic element to obtain at least one second discrete point corresponding to the second semantic element; for each first sub-road section, matching the first discrete point of the first sub-road section with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-road section.

[0092] Among them, for each optimized first semantic element, the first semantic element is sampled to obtain at least one first discrete point corresponding to the first semantic element, including: if the first semantic element is continuous, the first semantic element is continuously sampled to obtain multiple sampling points as the first discrete points corresponding to the first semantic element; if the first semantic element is discontinuous, the feature point of the first semantic element is determined to be the first discrete point corresponding to the first semantic element.

[0093] For example, refer to Figure 7 , sampling each first semantic element in the first sub-segment 71 of the crowdsourced map, and each first discrete point obtained by sampling is shown as 72. Figure 8 , each second semantic element of the first sub-segment 81 of the truth map is sampled, and each second discrete point obtained by sampling is shown as 82.

[0094] For a continuous first semantic element, for example, a first semantic element representing a continuous road line, sampling with a preset density can be performed, and the obtained multiple sampling points are used as first discrete points to represent the continuous first semantic element. For a discontinuous first semantic element, feature points, such as a center point, a center of gravity point, or a key point, can be sampled as first discrete points to represent the discontinuous first semantic element.

[0095] In the present application, a plurality of first discrete points are obtained by sampling each first semantic element of the first sub-segment, and a plurality of second discrete points are obtained by sampling each second semantic element of the second sub-segment. By matching the first discrete points with the second discrete points, the matching accuracy value between the corresponding first sub-segment and the second sub-segment can be accurately determined.

[0096] Among them, the first method: matching the first discrete point of the first sub-section with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-section, including: for each first discrete point, determining the second discrete point that is closest to the first discrete point; if the distance between the first discrete point and the second discrete point closest to it is less than or equal to the first threshold, then determining that the first discrete point and its closest second discrete point have a matching relationship; for each first sub-section, determining the first number of first discrete points with a matching relationship, and determining the total number of second discrete points in the corresponding second sub-section; based on the first number and the total number, determining the completeness rate and redundancy rate corresponding to the first sub-section, the completeness rate is the first ratio of the first number to the total number, and the redundancy rate is the second ratio of the first difference between the total number and the first number to the total number; determining the completeness rate and the redundancy rate as the matching accuracy value of the corresponding first sub-section.

[0097] In the present application, the matching relationship is specifically determined by a nearest neighbor algorithm (KNN).

[0098] For example, refer to Fig. 9 , the first discrete point 91 in the first sub-segment is matched with the second discrete point 92 in the second sub-segment. Among them, the total number of first discrete points 91 in the first sub-segment is 21. Among them, the number of first discrete points 91 that form a matching relationship is 3 (first discrete point a, first discrete point b and first discrete point c), among which, if the distance between the first discrete point a and the corresponding second discrete point A is greater than the threshold, no matching relationship is formed. If the distance between the first discrete point b and the corresponding second discrete point B is greater than the threshold, no matching relationship is formed. If the distance between the first discrete point c and the corresponding second discrete point C is greater than the threshold, no matching relationship is formed. Then the first number of first discrete points that form a matching relationship in the first sub-segment is 18, the total number of second discrete points 92 in the second sub-segment is 24, then the completion rate is 0.75. Redundancy rate is 0.25.

[0099] In the present application, if the matching accuracy value corresponding to the first sub-road section is expressed by the completion rate and / or redundancy rate, if the target accuracy value of the crowdsourcing map is, it can be based on the average of the completion rate and / or redundancy rate of each first sub-road section to obtain the corresponding target accuracy value.

[0100] In addition, the second method is to match the first discrete point with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-segment, including: for each first discrete point, determining the first distance between the first discrete point and the other first discrete points in the corresponding first sub-segment; determining the second distance between the second discrete point corresponding to the first discrete point and the other second discrete points in the corresponding second sub-segment; based on the first distance and the second distance, determining the relative accuracy value of the corresponding first sub-segment as the matching accuracy value.

[0101] Among them, according to the first distance and the second distance, the relative accuracy value of the corresponding first sub-road section is determined as the matching accuracy value, including: for each first discrete point, a first average value of the relative difference is determined according to the relative differences between the corresponding first distances and the second distances; for each first semantic element, the semantic accuracy value of the corresponding first semantic element is determined according to the first average value corresponding to each first discrete point in the first semantic element and the first number of first discrete points in the first semantic element; for each first sub-road section, the relative accuracy value of the corresponding first sub-road section is determined as the matching accuracy value according to the semantic accuracy value of each first semantic element in the first sub-road section and the second number of first semantic elements in the first sub-road section.

[0102] For example, refer to Fig.10 , if the first sub-segment includes four first semantic elements, each semantic element corresponds to a first discrete point, which are Fig.10 The four first discrete points e, f, g, and h in the figure have second discrete points E, F, G, and H corresponding to the first discrete points. The distances between the first discrete points are as follows: Fig.10 The distances between the second discrete points are as follows: Fig.10D1 to D6 shown. Among them, for the first discrete point e, the distances between the first discrete point e and each of the other first discrete points (f, g and h) are d1, d2 and d3 respectively. The second discrete point corresponding to the first discrete point e is E, and the distances between the second discrete point and each of the other second discrete points (F, G and H) are D1, D2 and D3 respectively. Then the first average value of the relative difference of the first discrete point e is R1 = {(d1-D1) / d1+(d2-D2) / d2+(d3-D3) / d3} / 3, and the first average values ​​of the relative differences of the first discrete points (f, g and h) are calculated in the same way as R2, R3 and R4 respectively. For example, the relative difference average value of the second discrete point f is R2 = {(d1-D1) / d1+(d4-D4) / d4+(d5-D5) / d5} / 3. Then the relative accuracy value of the first sub-section is RE = (R1+R2+R3+R4) / 4.

[0103] In addition, in the present application, the semantic accuracy value corresponding to each first semantic element can also be calculated first in the above manner. Among them, if a first semantic element corresponds to only one first discrete point, the first average value of the relative difference corresponding to the first discrete point is the semantic accuracy value of the first semantic element. If a first semantic element corresponds to N first discrete points, the first average values ​​of the relative differences corresponding to the N first discrete points are averaged to obtain the semantic accuracy value of the first semantic element. Then, the semantic accuracy values ​​of each first semantic element in the first sub-segment are averaged again to obtain the relative accuracy value of the first sub-segment as the matching accuracy value.

[0104] Furthermore, a third method matches the first discrete point with the corresponding second discrete point to obtain a matching accuracy value of the corresponding first sub-section, including: for each first sub-section, determining the absolute value of the position difference between each first discrete point in the first sub-section and the corresponding second discrete point; determining the second average value of the absolute value as the absolute accuracy value of the first sub-section; and using the absolute accuracy value as the matching accuracy value.

[0105] Each first discrete point has position information in the crowdsourced map, such as p1, p2, p3, ..., pn. Each second discrete point corresponding to the first discrete point has position information in the true map, such as P1, P2, P3, ..., Pn. Where n is the number of first discrete points in the first sub-segment. Then the absolute accuracy value ab of the first sub-segment = {|p1-P1|+|p2-P2|+|p3-P3|+...+|pn-Pn|} / n

[0106] S304: Obtain a target accuracy value according to the matching accuracy values ​​of the first sub-road sections.

[0107] The matching accuracy values ​​of the first sub-sections of the crowdsourced map are averaged to obtain the target accuracy value.

[0108] S305: Determine the quality of the crowdsourced map according to the target accuracy value.

[0109] The specific implementation process of this step refers to S203 and will not be repeated here.

[0110] The embodiment of the present application divides the crowdsourced map into a plurality of first sub-segments, divides the true value map into a plurality of second sub-segments, then performs registration optimization on the first sub-segments, samples each first semantic element of the first sub-segment after registration optimization to obtain a first discrete point, samples each second semantic element of the second sub-segment to obtain a second discrete point, performs matching calculation on the first discrete point and the second discrete point, and can accurately determine the matching accuracy value of each first sub-segment, and then determine the target progress value of the crowdsourced map, thereby accurately determining the quality of the crowdsourced map. And the quality determination method of the crowdsourced map has the technical effect of small error and low cost.

[0111] Fig.11 This is a structural block diagram of a crowdsourced map quality determination device 110 provided in an embodiment of the present application. The crowdsourced map quality determination device 110 provided in an embodiment of the present application includes:

[0112] An acquisition module 111 is used to acquire a crowdsourced map and a true map, where the crowdsourced map and the true map are maps of the same road section;

[0113] A matching module 112 is used to perform semantic matching between the crowdsourced map and the true value map to obtain a target accuracy value;

[0114] The determination module 113 is used to determine the quality of the crowdsourced map according to the target accuracy value.

[0115] In another embodiment of the present application, the matching module 112 includes:

[0116] A segmentation unit, configured to segment the crowdsourced map into a plurality of first sub-segments according to the trajectory information in the crowdsourced map, and segment the true value map into second sub-segments corresponding to each first sub-segment;

[0117] A matching unit, configured to perform semantic matching on each first sub-segment and its corresponding second sub-segment to obtain a matching accuracy value corresponding to each first sub-segment;

[0118] The determination unit is used to obtain a target accuracy value according to the matching accuracy values ​​of each first sub-road section.

[0119] In another embodiment of the present application, the matching unit is specifically used to: for each first sub-segment, determine a reference point in the first sub-segment; use the current posture information of the reference point as an optimization variable; determine an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment; determine a relationship constraint edge formed by the reference point and a matching point in the corresponding second sub-segment, the reference point and the corresponding matching point representing semantic elements at the same position; optimize the current posture information based on the adjacent constraint edge and the relationship constraint edge to obtain the target posture information of the reference point; optimize the posture information of the first semantic element based on the target posture information; match the optimized first semantic element with the corresponding second semantic element at the same position to obtain a matching accuracy value corresponding to the first sub-segment.

[0120] In another embodiment of the present application, when the matching unit is used to match the optimized first semantic element with the second semantic element corresponding to the same position to obtain the matching accuracy value of the corresponding first sub-segment, it is specifically used to: for each optimized first semantic element, sample the first semantic element to obtain at least one first discrete point corresponding to the first semantic element; for each second semantic element, sample the second semantic element to obtain at least one second discrete point corresponding to the second semantic element; for each first sub-segment, match the first discrete point of the first sub-segment with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-segment.

[0121] In another embodiment of the present application, when the matching unit is used to sample each optimized first semantic element to obtain at least one first discrete point corresponding to the first semantic element, it is specifically used to: if the first semantic element is continuous, continuously sample the first semantic element to obtain multiple sampling points as the first discrete points corresponding to the first semantic element; if the first semantic element is non-continuous, determine the feature point of the first semantic element as the first discrete point corresponding to the first semantic element.

[0122] In another embodiment of the present application, the matching unit is used to match the first discrete point of the first sub-section with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-section, specifically for: for each first discrete point, determining the second discrete point that is closest to the first discrete point; if the distance between the first discrete point and the second discrete point closest to it is less than or equal to the first threshold, then determining that the first discrete point and its closest second discrete point have a matching relationship; for each first sub-section, determining the first number of first discrete points with a matching relationship, and determining the total number of second discrete points in the corresponding second sub-section; based on the first number and the total number, determining the completeness rate and redundancy rate corresponding to the first sub-section, the completeness rate is a first ratio of the first number to the total number, and the redundancy rate is a second ratio of the first difference between the total number and the first number to the total number; determining the completeness rate and the redundancy rate as the matching accuracy value of the corresponding first sub-section.

[0123] In another embodiment of the present application, the matching unit is used to match the first discrete point with the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-segment, and is specifically used to: for each first discrete point, determine the first distance between the first discrete point and each other first discrete point in the corresponding first sub-segment; determine the second distance between the second discrete point corresponding to the first discrete point and each other second discrete point in the corresponding second sub-segment; and determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value based on the first distance and the second distance.

[0124] In another embodiment of the present application, the matching unit is used to determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value based on the first distance and the second distance, and is specifically used to: for each first discrete point, determine the first average value of the relative difference according to the relative differences between the corresponding first distances and the second distances; for each first semantic element, determine the semantic accuracy value of the corresponding first semantic element according to the first average value corresponding to each first discrete point in the first semantic element, and the first number of first discrete points in the first semantic element; for each first sub-segment, determine the relative accuracy value of the corresponding first sub-segment as the matching accuracy value according to the semantic accuracy value of each first semantic element in the first sub-segment and the second number of first semantic elements in the first sub-segment.

[0125] In another embodiment of the present application, the matching unit is used to match the first discrete point and the corresponding second discrete point to obtain the matching accuracy value of the corresponding first sub-section, and is specifically used to: for each first sub-section, determine the absolute value of the position difference between each first discrete point in the first sub-section and the corresponding second discrete point; determine the second average value of the absolute value as the absolute accuracy value of the first sub-section; and use the absolute accuracy value as the matching accuracy value.

[0126] The quality determination device of the crowdsourced map provided in the embodiment of the present application is used to perform Figure 2 , Figure 3 or Figure 4 The technical solutions in the corresponding method embodiments have similar implementation principles and technical effects, which will not be repeated here.

[0127] Fig.12 The hardware structure diagram of the electronic device provided in the embodiment of the present application is shown in FIG. Fig.12 As shown, the electronic device 120 of the embodiment of the present application may include: at least one processor 1201 ( Fig.12 Only one processor is shown); and a memory 1202 in communication with at least one processor. The memory 1202 stores instructions that can be executed by at least one processor 1201, and the instructions are executed by at least one processor 1201 so that the electronic device 120 can execute the technical solution in any of the aforementioned method embodiments.

[0128] Optionally, the memory 1202 may be independent or integrated with the processor 1201 .

[0129] When the memory 1202 is a device independent of the processor 1201 , the electronic device 120 further includes: a bus 1203 for connecting the memory 1202 and the processor 1201 .

[0130] The electronic device provided in the embodiment of the present application can execute the technical solution of any of the aforementioned method embodiments, and its implementation principles and technical effects are similar, which will not be repeated here.

[0131] An embodiment of the present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the technical solution in any of the aforementioned method embodiments is implemented.

[0132] An embodiment of the present application provides a computer program product, including a computer program, which implements the technical solution in any of the aforementioned method embodiments when executed by a processor.

[0133] The embodiment of the present application also provides a chip, including: a processing module and a communication interface, and the processing module can execute the technical solution in the aforementioned method embodiment.

[0134] Furthermore, the chip also includes a storage module (such as a memory), the storage module is used to store instructions, the processing module is used to execute the instructions stored in the storage module, and the execution of the instructions stored in the storage module enables the processing module to execute the technical solution in the aforementioned method embodiment.

[0135] It should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the invention may be directly implemented by a hardware processor or implemented by a combination of hardware and software modules in the processor.

[0136] The memory may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and may also be a USB flash drive, a mobile hard disk, a read-only memory, a magnetic disk or an optical disk, etc.

[0137] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0138] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0139] An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic device.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for determining the quality of a crowdsourced map, characterized in that: include: Obtaining a crowdsourced map and a true map, wherein the crowdsourced map and the true map are maps of the same road section; Performing semantic matching on the crowdsourced map and the true value map to obtain a target accuracy value; wherein the crowdsourced map includes a plurality of first semantic elements, and the true value map includes a plurality of second semantic elements; Determining the quality of the crowdsourced map according to the target accuracy value; The semantic matching of the crowdsourced map and the true value map to obtain a target accuracy value includes: Dividing the crowdsourced map into a plurality of first sub-segments according to the trajectory information in the crowdsourced map, and dividing the true value map into second sub-segments corresponding to each of the first sub-segments; For each of the first sub-road sections, semantically matching the first sub-road section with its corresponding second sub-road section to obtain a matching accuracy value corresponding to each of the first sub-road sections; Obtaining the target accuracy value according to the matching accuracy values ​​of each of the first sub-road sections; The step of performing semantic matching between each of the first sub-segments and its corresponding second sub-segment to obtain a matching accuracy value corresponding to each of the first sub-segments includes: For each of the first sub-sections, determining a reference point in the first sub-section; Using the current pose information of the reference point as an optimization variable; Determine an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment; Determine a relational constraint edge formed by the reference point and a matching point in the corresponding second sub-segment, wherein the reference point and the corresponding matching point represent semantic elements at the same position; the semantic elements are the shape, size, category and position information of each map element on the corresponding map; the reference point is any semantic element in the first sub-segment; Optimizing the current pose information based on the adjacent constraint edges and the relationship constraint edges to obtain target pose information of the reference point; Optimizing the position information of the first semantic element according to the target position information; The optimized first semantic element is matched with the second semantic element corresponding to the same position to obtain a matching accuracy value corresponding to the first sub-road segment.

2. The method for determining the quality of a crowdsourced map according to claim 1, characterized in that: The step of matching the optimized first semantic element with the second semantic element corresponding to the same position to obtain a matching accuracy value corresponding to the first sub-segment includes: For each optimized first semantic element, sampling the first semantic element to obtain at least one first discrete point corresponding to the first semantic element; For each of the second semantic elements, sampling the second semantic element to obtain at least one second discrete point corresponding to the second semantic element; For each of the first sub-road sections, the first discrete points of the first sub-road section are matched with the corresponding second discrete points to obtain a matching accuracy value of the corresponding first sub-road section.

3. The method for determining the quality of a crowdsourced map according to claim 2, characterized in that: The step of sampling each optimized first semantic element to obtain at least one first discrete point corresponding to the first semantic element includes: If the first semantic element is continuous, continuously sampling the first semantic element to obtain a plurality of sampling points as the first discrete points corresponding to the first semantic element; If the first semantic element is discontinuous, the feature point of the first semantic element is determined to be the first discrete point corresponding to the first semantic element.

4. The method for determining the quality of a crowdsourced map according to claim 2 or 3, characterized in that: The matching of the first discrete point of the first sub-segment with the corresponding second discrete point to obtain a matching accuracy value of the corresponding first sub-segment includes: For each of the first discrete points, determine a second discrete point that is closest to the first discrete point; If the distance between the first discrete point and the second discrete point closest to it is less than or equal to a first threshold, determining that the first discrete point and the second discrete point closest to it have a matching relationship; For each of the first sub-segments, determining a first number of the first discrete points having the matching relationship, and determining a total number of second discrete points in the corresponding second sub-segment; Determine, according to the first number and the total number, a completion rate and a redundancy rate corresponding to the first sub-segment, the completion rate being a first ratio of the first number to the total number, and the redundancy rate being a second ratio of a first difference between the total number and the first number to the total number; The completion rate and the redundancy rate are determined as a matching accuracy value corresponding to the first sub-segment.

5. The method for determining the quality of a crowdsourced map according to claim 2 or 3, characterized in that: The matching the first discrete point with the corresponding second discrete point to obtain a matching accuracy value of the corresponding first sub-segment includes: For each of the first discrete points, determining first distances between the first discrete point and other first discrete points in the corresponding first sub-segment; Determine a second distance between the second discrete point corresponding to the first discrete point and other second discrete points in the corresponding second sub-segment; According to the first distance and the second distance, a relative accuracy value corresponding to the first sub-segment is determined as a matching accuracy value.

6. The method for determining the quality of a crowdsourced map according to claim 5, characterized in that: The determining, according to the first distance and the second distance, a relative accuracy value of the corresponding first sub-segment as a matching accuracy value includes: For each of the first discrete points, determine a first average value of the relative differences according to the relative differences between the corresponding first distances and the corresponding second distances; For each of the first semantic elements, determine a semantic accuracy value corresponding to the first semantic element according to the first average value corresponding to each first discrete point in the first semantic element and the first number of the first discrete points in the first semantic element; the semantic accuracy value is equal to the quotient of the sum of the first average values ​​corresponding to the first discrete points divided by the first number; For each of the first sub-road sections, a relative accuracy value of the corresponding first sub-road section is determined as a matching accuracy value based on the semantic accuracy values ​​of each first semantic element in the first sub-road section and the second number of first semantic elements in the first sub-road section; the relative accuracy value is equal to the quotient of the sum of the semantic accuracy values ​​of the first semantic elements divided by the second number.

7. The method for determining the quality of a crowdsourced map according to claim 2 or 3, characterized in that: The matching the first discrete point with the corresponding second discrete point to obtain a matching accuracy value of the corresponding first sub-segment includes: For each of the first sub-sections, determining an absolute value of a position difference between each first discrete point in the first sub-section and a corresponding second discrete point; Determine a second average value of the absolute values ​​as the absolute accuracy value of the first sub-section; The absolute accuracy value is used as the matching accuracy value.

8. A device for determining the quality of a crowdsourced map, characterized in that: include: An acquisition module, used to acquire a crowdsourced map and a true map, wherein the crowdsourced map and the true map are maps of the same road section; A matching module, used for performing semantic matching between the crowdsourced map and the true value map to obtain a target accuracy value; wherein the crowdsourced map includes a plurality of first semantic elements, and the true value map includes a plurality of second semantic elements; A determination module, used to determine the quality of the crowdsourced map according to the target accuracy value; The matching module comprises: A segmentation unit, configured to segment the crowdsourced map into a plurality of first sub-segments according to the trajectory information in the crowdsourced map, and segment the true value map into second sub-segments corresponding to each of the first sub-segments; a matching unit, configured to perform semantic matching on each of the first sub-segments and its corresponding second sub-segment, to obtain a matching accuracy value corresponding to each of the first sub-segments; a determination unit, configured to obtain the target accuracy value according to the matching accuracy values ​​of the first sub-road sections; The matching unit is specifically used to determine a reference point in each of the first sub-segments; use the current posture information of the reference point as an optimization variable; determine an adjacent constraint edge formed by a first sub-segment adjacent to the first sub-segment; determine a relationship constraint edge formed by the reference point and a matching point in the corresponding second sub-segment, wherein the reference point and the corresponding matching point represent semantic elements at the same position; the semantic element is the shape, size, category and position information of each map element on the corresponding map; the reference point is any semantic element in the first sub-segment; based on the adjacent constraint edge and the relationship constraint edge, the current posture information is optimized to obtain the target posture information of the reference point; according to the target posture information, the posture information of the first semantic element is optimized; the optimized first semantic element is matched with the corresponding second semantic element at the same position to obtain the matching accuracy value of the corresponding first sub-segment.

9. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to perform the quality determination method of the crowdsourced map according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions. When the processor executes the computer-executable instructions, the method for determining the quality of a crowdsourced map as described in any one of claims 1 to 7 is implemented.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for determining the quality of a crowdsourced map according to any one of claims 1 to 7 is implemented.

Images