Method, device, medium, equipment and vehicle for evaluating map matching algorithm module

Abstract

This application discloses a method, apparatus, medium, device, and vehicle for evaluating a map matching algorithm module. The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number and match the sub-maps in the two target maps. The method includes: obtaining the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of successfully matched sub-maps; based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding relative transformation relationship truth value, counting the number of first sub-map matching pairs that meet a preset difference condition; determining whether the number of targets is the same as the number of sub-maps of the target map; and determining that the two target maps are successfully matched if the number of targets is the same as the number of sub-maps of the target map. This scheme can evaluate the map matching effect of the map matching algorithm module.

Description

Technical Field

[0001] This application relates to the field of map processing technology, and more specifically, to a method, apparatus, medium, equipment, and vehicle for evaluating a map matching algorithm module. Background Technology

[0002] To improve the accuracy of semantic maps during generation, a map matching algorithm module can first match road elements from multiple initial semantic maps generated based on crowdsourced data. Then, based on the relative positions of the same road element, the two successfully matched initial semantic maps are shifted to reduce the positional offset between them. Finally, the matched and shifted initial semantic maps are clustered to obtain the final semantic map. The map matching algorithm module includes a map matching model and a post-processing module. The map matching model performs pairwise matching of the multiple initial semantic maps, and the post-processing module filters out abnormal matching results obtained by the map matching model. However, currently, there is no method to accurately and quickly evaluate the map matching performance of this algorithm module. Summary of the Invention

[0003] This application provides a method, apparatus, medium, equipment, and vehicle for evaluating a map matching algorithm module, which can accurately and quickly evaluate the map matching effect of the map matching algorithm module.

[0004] The specific technical solution is as follows:

[0005] In a first aspect, embodiments of this application provide an evaluation method for a map matching algorithm module. The map matching algorithm module is used to divide two target maps into a plurality of sub-maps of the same number, and to match the sub-maps in the two target maps. The method includes:

[0006] Obtain the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of sub-maps that have been successfully matched;

[0007] Based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding relative transformation relationship truth value, the target number of first subgraph matching pairs that satisfy the preset difference condition is counted.

[0008] Determine whether the number of targets is the same as the number of sub-maps of the target map;

[0009] If the number of targets is the same as the number of sub-maps of the target map, then the two target maps are determined to be successfully matched.

[0010] As can be seen from the above scheme, this application implements matching by dividing two target maps into multiple sub-maps of equal number, and obtaining the relative transformation relationship and its corresponding truth value between each first sub-map matching pair (i.e., a successfully matched pair of sub-maps) output by the map matching algorithm module. Based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding truth value, the number of targets in the first sub-map matching pair that meet the preset difference condition is counted. Finally, by judging whether the number of targets is the same as the number of sub-maps of the target map, it is determined whether the two target maps are successfully matched. Therefore, it can be seen that the method of determining whether the complete map is successfully matched by summarizing the sub-map matching differences can not only ensure the evaluation accuracy of the map matching effect of the map matching algorithm module, but also improve the evaluation efficiency.

[0011] In a first possible implementation of the first aspect, determining whether the number of targets is the same as the number of sub-maps of the target map includes:

[0012] If the ratio of the number of target maps to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold, it is determined whether the number of target maps is the same as the number of sub-map maps of the target map.

[0013] As can be seen from the above scheme, in this embodiment of the application, the number of targets and the number of first sub-map matching pairs output by the map matching algorithm module are greater than or equal to a preset percentage threshold before it is determined whether the number of targets and the number of sub-maps of the target map are the same. This ensures that the matching effect at the sub-map level meets the requirements before the evaluation of the complete map is carried out, thereby further improving the evaluation accuracy of the map matching effect of the map matching algorithm module.

[0014] In a second possible implementation of the first aspect, obtaining the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding truth value includes:

[0015] If the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements, obtain the relative transformation relationship between each first sub-map matching pair and its corresponding truth value; or,

[0016] If the precision and / or recall of the road element matching pairs in each of the first sub-graph matching pairs output by the map matching algorithm module meet the requirements, the relative transformation relationship between each of the first sub-graph matching pairs and the corresponding truth value of the relative transformation relationship are obtained.

[0017] As can be seen from the above scheme, the embodiments of this application can obtain the relative transformation relationship between each first sub-map matching pair and its corresponding true value only when the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements, or when the precision and / or recall of road element matching pairs in each first sub-map matching pair output by the map matching algorithm module meet the requirements. This ensures that the matching effect at the road element level meets the requirements before performing sub-map level or full map level evaluation, thereby further improving the evaluation accuracy of the map matching effect of the map matching algorithm module.

[0018] In a third possible implementation of the first aspect, the map matching algorithm module includes a map matching model and a post-processing module, the post-processing module including:

[0019] For each second sub-map matching pair output by the map matching model, obtain the relative transformation relationship of each first road element matching pair contained in the second sub-map matching pair, wherein the first road element matching pair includes road element matching pairs that are not lane lines;

[0020] The optimal relative transformation relationship is determined based on the relative transformation relationship of each first road element matching pair in the second subgraph matching pair, and the target transformation range within the error range and including the optimal relative transformation relationship is determined.

[0021] The first road element matching pairs that are not within the target conversion range are filtered out, and the second road element matching pairs that are not within the target conversion range are filtered out, wherein the second road element matching pairs include matching pairs of each dashed line segment in lane lines where the line presentation is dashed;

[0022] At least one target point is selected from any road element in the third road element matching pair, and each target point is transformed according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point. The third road element matching pair includes lane line matching pairs with solid lines.

[0023] Based on the target distance from each of the transition points to the other road element in the third road element matching pair and the error range, abnormal third road element matching pairs are filtered out.

[0024] As can be seen from the above scheme, the embodiments of this application first determine the optimal relative transformation relationship from the relative transformation relationship of each first road element matching pair of each second sub-map matching pair output by the map matching model, and determine the target transformation range within the error range and including the optimal relative transformation relationship. For non-lane line road element matching pairs and each dashed line segment matching pair in the dashed lane lines, anomaly filtering can be performed directly through the target transformation range. For solid line lane line matching pairs (i.e., third road element matching pairs), at least one target point can be selected from any road element in the third road element matching pair, and each target point can be transformed according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point. Then, based on the target distance from each transformation point to the other road element in the third road element matching pair and the error range, abnormal third road element matching pairs are filtered. Thus, the embodiments of this application can use different methods to perform anomaly filtering according to different road elements, and the filtering range is each second sub-map matching pair, thereby realizing the local consistency verification of map matching results.

[0025] In a fourth possible implementation of the first aspect, filtering out anomalous third road element matching pairs based on the target distance from each of the transition points to the other road element in the third road element matching pair and the error range includes:

[0026] Calculate the target distance from each of the conversion points to the other road element in the third road element matching pair, and filter the third road element matching pair if there is a target distance greater than or equal to a preset proportion outside the error range in the target distances corresponding to the at least one target point.

[0027] As can be seen from the above scheme, the embodiments of this application only determine that the third road element matching pair is abnormal and filter it when the target distance greater than or equal to the preset ratio is outside the error range. This can avoid the third road element matching pair being mistakenly considered abnormal because the target distances corresponding to a small number of target points are outside the error range, thereby improving the accuracy of abnormal detection of the third road element matching pair.

[0028] In a fifth possible implementation of the first aspect, the post-processing module further includes:

[0029] After performing anomaly filtering for road element matching pairs on each of the second subgraph matching pairs output by the map matching model, the second subgraph matching pairs corresponding to the optimal relative transformation relationship that do not satisfy the linear distribution are filtered.

[0030] As can be seen from the above scheme, after completing the local consistency verification for each second subgraph matching pair, the embodiments of this application can achieve the global consistency verification of the complete map by judging whether the optimal relative transformation relationship corresponding to each second subgraph matching pair conforms to the linear distribution, thereby improving the accuracy of the post-processing module.

[0031] Secondly, embodiments of this application provide an evaluation device for a map matching algorithm module. The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number and to match the sub-maps in the two target maps. The device includes:

[0032] The acquisition unit is used to acquire the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of sub-maps that have been successfully matched;

[0033] The statistical unit is used to count the target number of first subgraph matching pairs that satisfy a preset difference condition based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding relative transformation relationship truth value.

[0034] The judgment unit is used to determine whether the number of targets is the same as the number of sub-maps of the target map;

[0035] The determining unit is used to determine that the two target maps are successfully matched when the number of targets is the same as the number of sub-maps of the target map.

[0036] In a first possible implementation of the second aspect, the judging unit is used to determine whether the number of targets is the same as the number of sub-maps of the target map when the ratio of the number of targets to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold.

[0037] In a second possible implementation of the third aspect, the acquisition unit is configured to acquire the relative transformation relationship between each first sub-map matching pair and its corresponding truth value, provided that the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements; or, provided that the precision and / or recall of all road element matching pairs in each first sub-map matching pair output by the map matching algorithm module meet the requirements, acquire the relative transformation relationship between each first sub-map matching pair and its corresponding truth value.

[0038] In a third possible implementation of the third aspect, the map matching algorithm module includes a map matching model and a post-processing module. The post-processing module includes: for each second sub-map matching pair output by the map matching model, obtaining the relative transformation relationship of each first road element matching pair contained in the second sub-map matching pair, wherein the first road element matching pair includes road element matching pairs that are not lane lines; determining the optimal relative transformation relationship based on the relative transformation relationship of each first road element matching pair in the second sub-map matching pair, and determining a target transformation range within the error range and containing the optimal relative transformation relationship; filtering out first roads whose relative transformation relationship is not within the target transformation range. The system includes: matching pairs of road elements, and filtering second road element matching pairs whose relative transformation relationship is not within the target transformation range, wherein the second road element matching pair includes matching pairs of each dashed line segment in lane lines where the line presentation is dashed; selecting at least one target point from any road element in the third road element matching pair, and transforming each target point according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point, wherein the third road element matching pair includes matching pairs of lane lines where the line presentation is solid; filtering out abnormal third road element matching pairs based on the target distance from each transformation point to another road element in the third road element matching pair and the error range.

[0039] In a fourth possible implementation of the third aspect, the post-processing module includes: calculating the target distance from each of the conversion points to another road element in the third road element matching pair, and filtering the third road element matching pair if there is a target distance greater than or equal to a preset proportion outside the error range among the target distances corresponding to the at least one target point.

[0040] In a fifth possible implementation of the third aspect, the post-processing module further includes: after performing anomaly filtering for road element matching pairs on each of the second subgraph matching pairs output by the map matching model, filtering the second subgraph matching pairs corresponding to the optimal relative transformation relationship that does not satisfy the linear distribution.

[0041] As can be seen from the above scheme, this application implements matching by dividing two target maps into multiple sub-maps of equal number, and obtaining the relative transformation relationship and its corresponding truth value between each first sub-map matching pair (i.e., a successfully matched pair of sub-maps) output by the map matching algorithm module. Based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding truth value, the number of targets in the first sub-map matching pair that meet the preset difference condition is counted. Finally, by judging whether the number of targets is the same as the number of sub-maps of the target map, it is determined whether the two target maps are successfully matched. Therefore, it can be seen that the method of determining whether the complete map is successfully matched by summarizing the sub-map matching differences can not only ensure the evaluation accuracy of the map matching effect of the map matching algorithm module, but also improve the evaluation efficiency.

[0042] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method as described in any possible implementation of the first aspect.

[0043] Fourthly, embodiments of this application provide an electronic device, which includes:

[0044] One or more processors;

[0045] The processor is coupled to a storage device for storing one or more programs;

[0046] When one or more programs are executed by one or more processors, the electronic device performs the method as described in any possible implementation of the first aspect.

[0047] Fifthly, embodiments of this application provide a vehicle that includes the means as described in any possible implementation of the second aspect, or includes electronic equipment as described in the fourth aspect.

[0048] In a sixth aspect, embodiments of this application provide a computer program product containing instructions that, when executed on a computer or processor, cause the computer or processor to perform the method described in any possible implementation of the first aspect. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0050] Figure 1A flowchart illustrating an evaluation method for a map matching algorithm module provided in an embodiment of this application;

[0051] Figure 2 A block diagram illustrating the composition of a map matching algorithm module provided in an embodiment of this application;

[0052] Figure 3 This is a block diagram of an evaluation device for a map matching algorithm module provided in an embodiment of this application. Detailed Implementation

[0053] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0054] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The terms "comprising" and "having," and any variations thereof, in the embodiments and drawings of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0055] Figure 1 This is a flowchart illustrating an evaluation method for a map matching algorithm module. This method can be applied to electronic or computer equipment, specifically vehicles or servers, and may include the following steps:

[0056] S110: Obtain the relative transformation relationship between each first subgraph matching pair output by the map matching algorithm module and its corresponding true value.

[0057] The map matching algorithm module is used to divide two target maps into multiple sub-maps of equal number and then match these sub-maps. The two target maps are the maps to be matched. When dividing a target map into multiple sub-maps, each target map can be directly divided into multiple equal sub-maps, or keyframes of each target map can be obtained first (e.g., image frames extracted at certain distance intervals), and then each road element can be assigned to the nearest keyframe, thus forming a sub-map with multiple road elements near a keyframe. The distance between a road element and a keyframe is determined primarily based on the road element's position on the target map and the camera's position when the keyframe was captured, and is unrelated to the image content of the keyframe itself.

[0058] The first subgraph matching pair includes a pair of successfully matched submaps. The relative transformation relationship between the first subgraph matching pairs includes a transformation matrix between them, which includes a translation matrix. The ground truth of the relative transformation relationship corresponding to the first subgraph matching pair includes the transformation matrix between any submap in the first subgraph matching pair and its truly matched submap, or the transformation matrix between a specified submap in the first subgraph matching pair and its truly matched submap. The specified submap can be a submap within a specified target map; for example, the more important target map among two target maps can be used as the specified target map. The ground truth of the relative transformation relationship can be pre-calculated based on the truly matched first subgraph matching pairs and stored in a preset space, or it can be calculated in real-time based on the ground truth annotations in the training samples of the map matching model in the map matching algorithm module. The ground truth annotations contain the annotation information of the truly matched road elements. The map matching model can be a GAT (Graph Attention Network) or other types of neural networks.

[0059] To improve the evaluation accuracy of map matching performance by the map matching algorithm module, the relative transformation relationship between each first sub-map matching pair and its corresponding truth value can be obtained, provided that the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements. Alternatively, the relative transformation relationship between each first sub-map matching pair and its corresponding truth value can be obtained, provided that the precision and / or recall of all road element matching pairs in each first sub-map matching pair output by the map matching algorithm module meet the requirements. In other words, this embodiment of the application can perform evaluation at the next level (i.e., a level lower than the road element granularity) only after ensuring that the matching performance at the road element level meets the requirements, thereby achieving a layered evaluation effect and improving the evaluation accuracy of map matching performance by the map matching algorithm module.

[0060] The precision and / or recall of road element matching pairs must meet the requirements if: the precision of the road element matching pairs is greater than or equal to a first precision threshold, and / or the recall of the road element matching pairs is greater than or equal to a first recall threshold. The first precision threshold and the first recall threshold may be the same or different.

[0061] Precision is relative to the prediction result; it represents the percentage of samples that were predicted as positive. Recall is relative to the sample size; it represents the percentage of positive samples that were correctly predicted.

[0062] The formulas for calculating precision and recall include:

[0063] Precision = TP total / (TP total +FP total )

[0064] Precall = TP total / (TP total +FN total )

[0065] TP total FP represents the number of positive samples that are judged as positive. total FN represents the number of negative samples misclassified as positive. total This represents the number of positive samples that were misclassified as negative. Positive samples represent correct road element matches or correct subgraph matches, while negative samples represent incorrect road element matches or incorrect subgraph matches.

[0066] S120: Based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding relative transformation relationship truth value, count the target number of first subgraph matching pairs that satisfy the preset difference condition.

[0067] The preset difference condition includes that the difference between the relative transformation relationship between the matching pairs of the first subgraphs and the true value of their corresponding relative transformation relationship is less than or equal to a preset difference threshold. The preset difference condition, which includes the difference between the relative transformation relationship between the matching pairs of the first subgraphs and the true value of their corresponding relative transformation relationship, can be the absolute value of the difference between the relative transformation relationship between the matching pairs of the first subgraphs and the true value of their relative transformation relationship. The preset difference threshold can be determined based on practical experience.

[0068] The preset difference condition can be expressed by the following formula:

[0069] |TT gt |≤ΔT target

[0070] Where T represents the relative transformation relationship between matching pairs in the first subgraph, T gt Represents the truth value of the relative transformation relationship, ΔT target This indicates the preset difference threshold.

[0071] S130: Determine whether the number of targets is the same as the number of submaps on the target map.

[0072] If the ratio of the number of target maps to the number of matching pairs in the first sub-map output by the map matching algorithm module is greater than or equal to a preset percentage threshold, it is determined whether the number of target maps is the same as the number of sub-maps in the target map. If the ratio is less than the preset percentage threshold, it is not determined whether the number of target maps is the same as the number of sub-maps in the target map, and the two target maps are directly determined to have failed to match. The preset percentage threshold can be determined based on practical experience.

[0073] In this embodiment, the determination of whether the number of targets and the number of sub-maps of the target map are the same is only made when the ratio of the number of targets to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold. This ensures that the matching effect at the sub-map level meets the requirements before the evaluation of the complete map is performed, thereby further improving the evaluation accuracy of the map matching effect of the map matching algorithm module.

[0074] S140: If the number of targets is the same as the number of sub-maps of the target map, the two target maps are determined to be successfully matched.

[0075] If the number of targets is the same as the number of submaps of the target map, it means that each first submap matching pair satisfies the preset difference condition, thus it can be determined that the two target maps are successfully matched; if the number of targets is different from the number of submaps of the target map, it means that at least one first submap matching pair satisfies the non-preset difference condition, thus it can be determined that the two target maps are not matched.

[0076] The evaluation method for the map matching algorithm module provided in this application involves dividing two target maps into multiple sub-maps of equal number for matching. The method obtains the relative transformation relationship and its corresponding ground truth value between each first sub-map matching pair (i.e., a successfully matched pair of sub-maps) output by the map matching algorithm module. Based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding ground truth value, the number of first sub-map matching pairs that meet a preset difference condition is counted. Finally, the matching success of the two target maps is determined by whether the number of targets is the same as the number of sub-maps of the target map. Therefore, determining the success of a complete map match by summarizing sub-map matching differences not only ensures the accuracy of the map matching algorithm module's map matching performance evaluation but also improves evaluation efficiency.

[0077] In one implementation, such as Figure 2 As shown, the map matching algorithm module includes a map matching model and a post-processing module. The map matching model is used to divide two target maps into multiple sub-maps of the same number and match the sub-maps in the two target maps to obtain multiple second sub-map matching pairs. The post-processing module is used to filter out abnormal matching results in the matching results obtained by the map matching model. This embodiment of the application can evaluate not only the map matching algorithm module but also the map matching model.

[0078] The evaluation method for map matching models includes: calculating the precision and / or recall of road element matching pairs for each type of road element output by the map matching model; when the precision of road element matching pairs for all road element types is greater than or equal to a second precision threshold, and / or the recall of road element matching pairs for all road element types is greater than or equal to a second recall threshold, the map matching model is determined to meet the quality requirements.

[0079] The second precision thresholds for different road element types can be the same or different; similarly, the second recall thresholds for different road element types can be the same or different. Road element types include road signs, traffic lights, streetlights, lane markings, and non-lane markings on the road surface, among others.

[0080] In one implementation, the post-processing module includes steps A1-A5:

[0081] A1. For each second sub-graph matching pair output by the map matching model, obtain the relative transformation relationship of each first road element matching pair contained in the second sub-graph matching pair.

[0082] The first road element matching pair includes road element matching pairs that do not include lane lines. Non-lane line road elements include traffic lights, road signs, streetlights, and markings on the road surface that do not include lane lines (such as text markings). The relative transformation relationship of the first road element matching pairs includes a transformation matrix between the first road element matching pairs, which includes a translation matrix.

[0083] To improve the computational efficiency of relative transformation relationships, each road element in the first road element matching pair can be treated as a point, and the coordinate difference between the two points can be used as the relative transformation relationship. Specifically, for each first road element matching pair contained in the second subgraph matching pair, the midpoint of each road element in the first road element matching pair can be obtained, and the coordinate difference between the midpoints can be used to determine the relative transformation relationship. For traffic lights, road signs, streetlights, etc., which consist of poles and sign areas, when obtaining the midpoint of each road element in the first road element matching pair, the poles can be ignored, and the midpoint of the sign area can be directly selected as the midpoint of the road element.

[0084] A2. Determine the optimal relative transformation relationship based on the relative transformation relationship of each first road element matching pair in the second subgraph matching pair, and determine the target transformation range within the error range and including the optimal relative transformation relationship.

[0085] The specific implementation method for determining the optimal relative transformation relationship based on the relative transformation relationship of each first road element matching pair in the second subgraph matching pair includes: using algorithms such as RANSAC (Random Sample Consensus) or least squares to calculate the relative transformation relationship of each first road element matching pair in the second subgraph matching pair to obtain the optimal relative transformation relationship. The error range can be (-δ, +δ), the optimal relative transformation relationship can be the midpoint of the target transformation range, and the difference between the optimal relative transformation relationship and the start and end points of the target transformation range are δm, for example, 2m.

[0086] A3. Filter out matching pairs of first road elements whose relative transformation relationship is not within the target transformation range, and filter out matching pairs of second road elements whose relative transformation relationship is not within the target transformation range.

[0087] The second road element matching pair includes matching pairs of each dashed line segment in the lane lines where the lines are presented as dashed lines. The relative transformation relationship of the dashed line segment matching pair can be calculated as follows: first, take the midpoint of each dashed line segment in the dashed line segment matching pair, and then calculate the difference between the two midpoints as the relative transformation relationship of the dashed line segment matching pair.

[0088] When the relative transformation relationship of the first road element matching pair or the second road element matching pair is not within the target transformation range, it means that the positions of the two road elements contained in the first road element matching pair or the second road element matching pair are significantly different and they do not match. Therefore, the first road element matching pair or the second road element matching pair is an abnormal road element matching pair and can be filtered out.

[0089] A4. Select at least one target point from any road element in the third road element matching pair, and transform each target point according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point.

[0090] The third road element matching pair includes lane line matching pairs where the lines are presented as solid lines. The process of calculating the conversion point coordinates is the reverse of the aforementioned calculation of the relative conversion relationship. For example, when the method for calculating the relative conversion relationship is the midpoint coordinates of a road element belonging to target map A minus the midpoint coordinates of a road element belonging to map B, and at least one selected target point is located in target map A, the coordinates of the conversion point = coordinates of the road element point - optimal relative conversion relationship; when the method for calculating the relative conversion relationship is the midpoint coordinates of a road element belonging to target map A minus the midpoint coordinates of a road element belonging to target map B, and at least one selected target point is located in map B, the coordinates of the conversion point = coordinates of the target point + optimal relative conversion relationship.

[0091] A5. Filter out abnormal third-road element matching pairs based on the target distance and error range from each conversion point to the other road element in the third-road element matching pair.

[0092] Specifically, the target distance from each conversion point to the other road element in the third road element matching pair can be calculated. If, in at least one target point, there is a target distance greater than or equal to a preset proportion that is outside the error range, the third road element matching pair is filtered out. The preset proportion can be determined based on practical experience, for example, it can be 50%.

[0093] When at least one target point has a target distance greater than or equal to a preset proportion that is outside the error range, it means that most target points have not been mapped to another road element through the optimal relative transformation relationship. For example, most points on one lane have not been mapped to another lane through the optimal relative transformation relationship. This indicates that the two road elements in the third road element matching pair output by the map matching model are far apart and do not match, so they can be filtered out.

[0094] This method determines that the third road element matching pair is abnormal and filters it only when the target distance greater than or equal to a preset proportion is outside the error range. This avoids mistaking the third road element matching pair for anomaly due to a small number of target points having target distances outside the error range, thereby improving the accuracy of anomaly detection for the third road element matching pair.

[0095] It should be added that the above method mainly considers whether the matching result is abnormal from the range of the matching pairs in the second subgraph. Therefore, the above process can be called local consistency check.

[0096] This application embodiment first determines the optimal relative transformation relationship from the relative transformation relationship of each first road element matching pair in each second sub-map matching pair output by the map matching model, and then determines the target transformation range within the error range and including the optimal relative transformation relationship. For non-lane line road element matching pairs and each dashed line segment matching pair in dashed lane lines, anomaly filtering can be performed directly through this target transformation range. For solid line lane line matching pairs (i.e., third road element matching pairs), at least one target point can be selected from any road element in the third road element matching pair, and each target point can be transformed according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point. Then, based on the target distance from each transformation point to the other road element in the third road element matching pair and the error range, abnormal third road element matching pairs are filtered. Thus, this application embodiment can use different methods for anomaly filtering according to different road elements, and the filtering range is each second sub-map matching pair, thereby achieving local consistency verification of map matching results.

[0097] In one implementation, the post-processing module may further include: after performing anomaly filtering for road element matching pairs on each second subgraph matching pair output by the map matching model, filtering the second subgraph matching pairs corresponding to the optimal relative transformation relationship that does not satisfy the linear distribution.

[0098] Since the trajectory is continuous, all optimal relative transformation relationships between two target maps have a certain linear distribution pattern. Therefore, outliers with jumps can be eliminated, that is, the matching pairs of the second sub-maps corresponding to optimal relative transformation relationships that do not satisfy the linear distribution can be filtered out.

[0099] In this embodiment of the application, after completing the local consistency check for each second subgraph matching pair, the global consistency check of the complete map can be achieved by determining whether the optimal relative transformation relationship corresponding to each second subgraph matching pair conforms to a linear distribution, thereby improving the accuracy of the post-processing module.

[0100] In one implementation, to further improve the evaluation criteria of the map matching algorithm module and enhance the matching quality of the final obtained map matching algorithm module, multiple target map pairs can be matched, and the matching results can be evaluated. When a task includes M... total After processing by the map matching algorithm module, M target map pairs were matched. ddmatch There are M target map pairs, of which M are successfully matched. success If there are [number] instances, then the precision' and recall' for map matching in this task segment are represented as follows:

[0101] Precision′=M success / M ddmatch

[0102] Precall'=M success / M total

[0103] If the precision of the task segment is greater than or equal to the third precision threshold (e.g., 98%), and the recall of the task segment is greater than or equal to the third recall threshold (e.g., 80%), then the matching quality of the map matching algorithm module is considered to meet the requirements.

[0104] Corresponding to the above method embodiments, another embodiment of this application provides an evaluation device for a map matching algorithm module. The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number and to match the sub-maps in the two target maps, such as... Figure 3 As shown, the device includes:

[0105] The acquisition unit 210 is used to acquire the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of sub-maps that have been successfully matched;

[0106] The statistics unit 220 is used to count the target number of first subgraph matching pairs that satisfy a preset difference condition based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding true value of the relative transformation relationship.

[0107] The judgment unit 230 is used to determine whether the number of targets is the same as the number of sub-maps of the target map;

[0108] The determining unit 240 is used to determine that the two target maps are successfully matched when the number of targets is the same as the number of sub-maps of the target map.

[0109] In one possible implementation, the judgment unit 230 is used to determine whether the number of targets is the same as the number of sub-maps of the target map when the ratio of the number of targets to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold.

[0110] In one possible implementation, the acquisition unit 210 is configured to acquire the relative transformation relationship between each first sub-map matching pair and its corresponding true value, provided that the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements; or, provided that the precision and / or recall of all road element matching pairs in each first sub-map matching pair output by the map matching algorithm module meet the requirements, acquire the relative transformation relationship between each first sub-map matching pair and its corresponding true value.

[0111] In one possible implementation, the map matching algorithm module includes a map matching model and a post-processing module. The post-processing module includes: for each second sub-map matching pair output by the map matching model, obtaining the relative transformation relationship of each first road element matching pair contained in the second sub-map matching pair, wherein the first road element matching pair includes road element matching pairs that are not lane lines; determining the optimal relative transformation relationship based on the relative transformation relationship of each first road element matching pair in the second sub-map matching pair, and determining a target transformation range within the error range and containing the optimal relative transformation relationship; filtering out first road element matching pairs whose relative transformation relationship is not within the target transformation range. The process involves pairing and filtering second road element matching pairs whose relative transformation relationship is not within the target transformation range. The second road element matching pair includes matching pairs of each dashed line segment in lane lines where the line presentation is dashed. At least one target point is selected from any road element in the third road element matching pair, and each target point is transformed according to the optimal relative transformation relationship to obtain a transformation point corresponding to each target point. The third road element matching pair includes lane line matching pairs where the line presentation is solid. Abnormal third road element matching pairs are filtered based on the target distance from each transformation point to another road element in the third road element matching pair and the error range.

[0112] In one possible implementation, the post-processing module performs the step of filtering abnormal third road element matching pairs based on the target distance from each conversion point to the other road element in the third road element matching pair and the error range, including: calculating the target distance from each conversion point to the other road element in the third road element matching pair, and filtering the third road element matching pair if there is a target distance greater than or equal to a preset proportion outside the error range in the target distances corresponding to the at least one target point.

[0113] In one possible implementation, the post-processing module further includes: after performing anomaly filtering for road element matching pairs on each second subgraph matching pair output by the map matching model, filtering the second subgraph matching pairs corresponding to the optimal relative transformation relationship that does not satisfy linear distribution.

[0114] The evaluation device for the map matching algorithm module provided in this application can divide two target maps into multiple sub-maps of the same number for matching, and obtain the relative transformation relationship and its corresponding truth value between each first sub-map matching pair (i.e., a successfully matched pair of sub-maps) output by the map matching algorithm module. Based on the difference between the relative transformation relationship between each first sub-map matching pair and its corresponding truth value, the number of targets in the first sub-map matching pair that meet the preset difference condition is counted. Finally, by judging whether the number of targets is the same as the number of sub-maps of the target map, it is determined whether the two target maps are successfully matched. Therefore, determining whether the complete map is successfully matched by summarizing the sub-map matching differences not only ensures the evaluation accuracy of the map matching algorithm module's map matching effect but also improves the evaluation efficiency.

[0115] Based on the above method embodiments, another embodiment of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the above embodiments.

[0116] Based on the above method embodiments, another embodiment of this application provides an electronic device or computer device, including:

[0117] One or more processors;

[0118] The processor is coupled to a storage device for storing one or more programs;

[0119] When the one or more programs are executed by the one or more processors, the electronic device or computer device performs the method as described in any of the above embodiments.

[0120] Based on the above method embodiments, another embodiment of this application provides a vehicle that includes the apparatus as described in any of the above embodiments, or includes electronic devices as described above.

[0121] The vehicle includes a CPU (Central Processing Unit) and a T-Box (Telematics Box). The T-Box can act as a gateway to communicate with the server. The CPU can evaluate the quality of the map matching algorithm module by executing the evaluation method described above. Alternatively, the CPU can send information such as the map matching algorithm module and two target maps to the server via the T-Box, allowing the server to execute the evaluation method and evaluate the quality of the map matching algorithm module. Or, technicians can directly store the map matching algorithm module and two target maps in the server, which will then execute the evaluation method to evaluate the quality of the map matching algorithm module.

[0122] Based on the above embodiments, another embodiment of this application provides a computer program product, which includes instructions that, when executed on a computer or processor, cause the computer or processor to perform the method described in any of the above embodiments.

[0123] The above-described apparatus embodiments correspond to the method embodiments and have the same technical effects. For detailed descriptions, please refer to the method embodiments. The apparatus embodiments are derived from the method embodiments; detailed descriptions can be found in the method embodiments section, and will not be repeated here. Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application.

[0124] Those skilled in the art will understand that the modules in the apparatus of the embodiments can be distributed in the apparatus of the embodiments as described in the embodiments, or they can be located in one or more devices different from this embodiment with corresponding changes. The modules of the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for evaluating a map matching algorithm module, characterized in that, The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number, and to match the sub-maps in the two target maps. The method includes: Obtain the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of sub-maps that have been successfully matched; Based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding relative transformation relationship truth value, the target number of first subgraph matching pairs that satisfy the preset difference condition is counted. Determine whether the number of targets is the same as the number of sub-maps of the target map; If the number of targets is the same as the number of sub-maps of the target map, then the two target maps are determined to be successfully matched. The map matching algorithm module includes a map matching model and a post-processing module. The post-processing module includes: For each second sub-map matching pair output by the map matching model, obtain the relative transformation relationship of each first road element matching pair contained in the second sub-map matching pair, wherein the first road element matching pair includes road element matching pairs that are not lane lines; The optimal relative transformation relationship is determined based on the relative transformation relationship of each first road element matching pair in the second subgraph matching pair, and the target transformation range within the error range and including the optimal relative transformation relationship is determined. The first road element matching pairs that are not within the target conversion range are filtered out, and the second road element matching pairs that are not within the target conversion range are filtered out, wherein the second road element matching pairs include matching pairs of each dashed line segment in lane lines where the line presentation is dashed; At least one target point is selected from any road element in the third road element matching pair, and each target point is transformed according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point. The third road element matching pair includes lane line matching pairs with solid lines. Based on the target distance from each of the transition points to the other road element in the third road element matching pair and the error range, abnormal third road element matching pairs are filtered out.

2. The method according to claim 1, characterized in that, The step of determining whether the number of targets is the same as the number of sub-maps of the target map includes: If the ratio of the number of target maps to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold, it is determined whether the number of target maps is the same as the number of sub-map maps of the target map.

3. The method according to claim 1, characterized in that, The step of obtaining the relative transformation relationship between each first sub-graph matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value includes: If the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements, obtain the relative transformation relationship between each first sub-map matching pair and its corresponding truth value; or, If the precision and / or recall of the road element matching pairs in each of the first sub-graph matching pairs output by the map matching algorithm module meet the requirements, the relative transformation relationship between each of the first sub-graph matching pairs and the corresponding truth value of the relative transformation relationship are obtained.

4. The method according to claim 1, characterized in that, The step of filtering out abnormal third road element matching pairs based on the target distance from each conversion point to the other road element in the third road element matching pair and the error range includes: Calculate the target distance from each of the conversion points to the other road element in the third road element matching pair, and filter the third road element matching pair if there is a target distance greater than or equal to a preset proportion outside the error range in the target distances corresponding to the at least one target point.

5. The method according to claim 1, characterized in that, The post-processing module further includes: After performing anomaly filtering for road element matching pairs on each of the second subgraph matching pairs output by the map matching model, the second subgraph matching pairs corresponding to the optimal relative transformation relationship that do not satisfy the linear distribution are filtered.

6. An evaluation device for a map matching algorithm module, characterized in that, The map matching algorithm module is used to divide two target maps into multiple sub-maps of the same number, and to match the sub-maps in the two target maps. The device includes: The acquisition unit is used to acquire the relative transformation relationship between each first sub-map matching pair output by the map matching algorithm module and its corresponding relative transformation relationship truth value, wherein the first sub-map matching pair includes a pair of sub-maps that have been successfully matched; The statistical unit is used to count the target number of first subgraph matching pairs that satisfy a preset difference condition based on the difference between the relative transformation relationship between each first subgraph matching pair and its corresponding relative transformation relationship truth value. The judgment unit is used to determine whether the number of targets is the same as the number of sub-maps of the target map; A determining unit is configured to determine that the two target maps are successfully matched when the number of targets is the same as the number of sub-maps of the target map; The map matching algorithm module includes a map matching model and a post-processing module. The post-processing module includes: for each second sub-map matching pair output by the map matching model, obtaining the relative transformation relationship of each first road element matching pair contained in the second sub-map matching pair, wherein the first road element matching pair includes road element matching pairs that are not lane lines; determining the optimal relative transformation relationship based on the relative transformation relationship of each first road element matching pair in the second sub-map matching pair, and determining a target transformation range within the error range and containing the optimal relative transformation relationship; filtering out first road element matching pairs whose relative transformation relationship is not within the target transformation range, and over-filtering out... The process involves filtering out second road element matching pairs whose relative transformation relationship is not within the target transformation range, wherein the second road element matching pair includes matching pairs of each dashed line segment in lane lines where the line presentation is dashed; selecting at least one target point from any road element in the third road element matching pair, and transforming each target point according to the optimal relative transformation relationship to obtain the transformation point corresponding to each target point, wherein the third road element matching pair includes matching pairs of lane lines where the line presentation is solid; and filtering out abnormal third road element matching pairs based on the target distance from each transformation point to another road element in the third road element matching pair and the error range.

7. The apparatus according to claim 6, characterized in that, The judgment unit is used to determine whether the number of targets is the same as the number of sub-maps of the target map when the ratio of the number of targets to the number of first sub-map matching pairs output by the map matching algorithm module is greater than or equal to a preset percentage threshold.

8. The apparatus according to claim 6, characterized in that, The acquisition unit is configured to acquire the relative transformation relationship between each first sub-map matching pair and its corresponding true value, provided that the precision and / or recall of all road element matching pairs output by the map matching algorithm module meet the requirements; or, provided that the precision and / or recall of all road element matching pairs in each first sub-map matching pair output by the map matching algorithm module meet the requirements, acquire the relative transformation relationship between each first sub-map matching pair and its corresponding true value.

9. The apparatus according to claim 6, characterized in that, The post-processing module further includes: after performing anomaly filtering for road element matching pairs on each second subgraph matching pair output by the map matching model, filtering the second subgraph matching pairs corresponding to the optimal relative transformation relationship that do not satisfy linear distribution.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1-5.

11. An electronic device, characterized in that, The electronic device includes: One or more processors; The processor is coupled to a storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the electronic device performs the method as described in any one of claims 1-5.

12. A vehicle, characterized in that, The vehicle includes the device as described in any one of claims 6-9, or the electronic device as described in claim 11.

Images