Positioning accuracy determination method and apparatus, electronic device, and storage medium

Abstract

The embodiment of the present disclosure discloses a positioning accuracy determination method, by acquiring a current frame point cloud and a current frame pose, transforming the current frame point cloud to a world coordinate system according to the current frame pose, and then dividing the second layer point cloud on the ground in the current frame point cloud into a plurality of second layer block point clouds on the ground, for each second layer block point cloud, determining its grid occupancy probability average and hit rate on the map, screening out the second layer matching block point cloud, determining the positioning judgment result of the current frame pose based on the number of the second layer matching block point cloud, and determining the confidence of the current frame pose according to the positioning judgment result, so as to calculate the confidence of the positioning through the grid occupancy probability average and the hit rate of the map area which is not easy to change, solve the problem that false positive and false negative of the positioning confidence easily appear in the prior art, improve the robustness of the positioning confidence, and make the confidence more distinguishable in expressing whether the pose is accurate.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent driving technology, and in particular to a method, apparatus, electronic device, and storage medium for determining the accuracy of positioning. Background Technology

[0002] Positional confidence represents the accuracy of the current localization pose. Existing methods, when locating on a newly created map, often yield a high confidence score because the environmental point cloud scanned at the current location matches the map well; this is a true positive for pose accuracy. Even when the current localization pose is correct, slow, cumulative changes in the environment and the intrusion of dynamic objects can lead to a poorer match between the environmental point cloud scanned at the current location and the map. This results in a low average score for the current frame's point cloud matching with the map, leading to a low confidence score; this is a false negative for pose accuracy. Conversely, when the current localization pose is incorrect, mismatches occur between the current frame's point cloud and the map, resulting in a higher confidence score; this is a false positive for pose accuracy.

[0003] Therefore, the current method has poor robustness in location reliability and is prone to false positives and false negatives in location reliability. Summary of the Invention

[0004] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, electronic device, and storage medium for determining the accuracy of positioning, which solves the problem of false positives and false negatives in positioning information that are prone to occur in the prior art, and realizes a reliable judgment on whether the current frame pose positioning is accurate.

[0005] In a first aspect, embodiments of this disclosure provide a method for determining the accuracy of positioning, the method comprising:

[0006] Obtain the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0007] The ground-level second-layer point cloud in the current frame point cloud is divided into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, the average grid occupancy probability and hit rate of the ground-level second-layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the corresponding grid of the ground-level second-layer block point cloud that are occupied by objects to the total number of laser points in the ground-level second-layer block point cloud.

[0008] Based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds, a second-layer matching block point cloud is determined in each of the above-ground second-layer block point clouds;

[0009] The localization judgment result of the current frame pose is determined based on the number of point clouds in the second-layer matching block, and the confidence level of the current frame pose is determined based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0010] Secondly, embodiments of this disclosure also provide a positioning accuracy determination device, the device comprising:

[0011] The transformation module is used to acquire the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0012] The statistics module is used to divide the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. For each ground second layer block point cloud, the average grid occupancy probability and hit rate of the ground second layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the ground second layer block point cloud whose corresponding grid is occupied by an object to the total number of laser points in the ground second layer block point cloud.

[0013] The matching module is used to determine the second-layer matching block point cloud in each of the above-ground second-layer block point clouds based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds.

[0014] The determination module is used to determine the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching block, and to determine the confidence level of the current frame pose based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0015] Thirdly, embodiments of this disclosure also provide an electronic device, the electronic device comprising: one or more processors; a storage device for storing one or more programs; and when the one or more programs are executed by the one or more processors, causing the one or more processors to implement the positioning accuracy determination method as described above.

[0016] Fourthly, embodiments of this disclosure also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the positioning accuracy determination method as described above.

[0017] This disclosure provides a method for determining the accuracy of localization. It acquires the current frame point cloud and the current frame pose, transforms the current frame point cloud to a world coordinate system based on the current frame pose, and then divides the ground-level second-layer point cloud in the current frame point cloud into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, it determines the average grid occupancy probability and hit rate on the map. Second-layer matching block point clouds are selected based on the average grid occupancy probability and hit rate. The localization judgment result of the current frame pose is determined based on the number of second-layer matching block point clouds, thereby determining the localization accuracy. The result determines the confidence level of the current frame pose to filter out potential, unchanging map regions such as the second layer of point cloud on the ground. The confidence level of localization is calculated using the average grid occupancy probability and hit rate of the unchanging map regions. This solves the problem of false positives and false negatives in localization confidence in existing technologies, improves the robustness of localization confidence, and enables a reliable judgment on whether the current frame pose localization is accurate. Furthermore, this method outputs the confidence level through two stages: determining the localization judgment result and determining the confidence level based on the localization judgment result. This makes the confidence level more discriminative in expressing whether the pose is accurate. Attached Figure Description

[0018] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale.

[0019] Figure 1 This is a flowchart of a method for determining the accuracy of positioning according to an embodiment of this disclosure;

[0020] Figure 2 A schematic diagram illustrating a positioning accuracy determination process provided in an embodiment of this disclosure;

[0021] Figure 3 This is a schematic diagram of a positioning accuracy determination device according to an embodiment of the present disclosure;

[0022] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this disclosure. Detailed Implementation

[0023] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0024] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0025] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0026] Before detailing the method for determining the accuracy of positioning provided in the embodiments of this disclosure, the technical problem solved by this method will be described by way of example. In laser-based positioning, the positioning confidence is related to the degree of matching between the spatial location of the point cloud in the current frame and the map. In grid-based positioning, existing positioning confidence typically first determines the average grid occupancy probability of all points in the current frame's point cloud on the map, and then outputs whether the positioning is successful through an average value threshold. If the average value is greater than the average value threshold, the average grid occupancy probability is used as the positioning confidence, indicating that the accuracy of the current positioning meets the standard and the positioning is successful; if the average value is less than the average value threshold, the average grid occupancy probability is used as the positioning confidence, indicating that the accuracy of the current positioning does not meet the standard and the positioning fails.

[0027] The advantage of the existing method is that it is simple to calculate and can comprehensively reflect the matching degree between the current frame and the map. The disadvantage is that it only uses the average value of the occupation probability of a single grid as the basis for judging whether the localization is successful, the reliability of the localization is not robust, and it is easy to produce false positives and false negatives in the localization reliability.

[0028] Therefore, in order to solve the problem of false positives and false negatives in positioning confidence in the prior art, this disclosure provides a method for determining the accuracy of positioning, which is applicable to determining the positioning accuracy of vehicle pose. The method determines the confidence level by using potentially unchanging map areas, minimizing the impact of changes in the environment and dynamic objects on the confidence level, and improving the accuracy of the final output confidence level.

[0029] Figure 1 This is a flowchart illustrating a method for determining positioning accuracy according to an embodiment of this disclosure. The method can be executed by a positioning accuracy determination device, which can be implemented in software and / or hardware and can be configured in an electronic device. Figure 1 As shown, the method may specifically include the following steps:

[0030] S110. Obtain the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose.

[0031] The current frame point cloud can be the point cloud data collected by the vehicle-mounted LiDAR in the current cycle; the current frame pose can be the vehicle's pose in the current cycle. For example, the current frame pose can be represented by six degrees of freedom parameters, such as the coordinates in the X direction (i.e., translation along the X-axis), the coordinates in the Y direction (i.e., translation along the Y-axis), the coordinates in the Z direction (i.e., translation along the Z-axis), and the roll angle, pitch angle, and yaw angle; where the roll angle, pitch angle, and yaw angle can be understood as the angles of rotation of the vehicle body around each coordinate axis, the Y direction is the direction of the vehicle's head, the X direction is perpendicular to the Y direction and the plane formed by the X and Y directions is parallel to the ground, and the Z direction is perpendicular to the plane formed by the X and Y directions.

[0032] In this embodiment, the current frame point cloud is specifically the point cloud data in the current frame coordinate system, and the current frame pose is specifically the pose data in the map coordinate system. Therefore, in order to further determine the positioning accuracy of the current frame pose, the current frame point cloud can be transformed from the current frame coordinate system to the map coordinate system first through the current frame pose.

[0033] S120. Divide the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. For each ground second layer block point cloud, determine the average grid occupancy probability and hit rate of the ground second layer block point cloud on the map. The hit rate is the ratio of the number of laser points in the corresponding grid of the ground second layer block point cloud that are occupied by objects to the total number of laser points in the ground second layer block point cloud.

[0034] Specifically, after transforming the current frame point cloud from the current frame coordinate system to the map coordinate system, the current frame point cloud can be divided into a ground point cloud, a first-layer ground point cloud, and a second-layer ground point cloud according to the height of each laser point. The height of the laser points in the second-layer ground point cloud is greater than the height of the laser points in the first-layer ground point cloud, and the height of the laser points in the first-layer ground point cloud is greater than the height of the laser points in the ground point cloud.

[0035] Furthermore, the second-layer point cloud on the ground can be divided into blocks according to the XY plane to obtain multiple second-layer block point clouds. For example, the XY plane can be divided into multiple square regions according to a set size or a set number, and the second-layer block point clouds on the ground can be constructed based on the laser points in the second-layer point cloud that fall within each square region; or, the XY plane can be divided into multiple fan-shaped regions with the front of the vehicle as the vertex and according to a set angle, and the second-layer block point clouds on the ground can be constructed based on the laser points in the second-layer point cloud that fall within each fan-shaped region.

[0036] It should be noted that the purpose of dividing the second-layer point cloud above ground into multiple second-layer block point clouds in this embodiment is as follows: considering that the regions in the second-layer point cloud above ground usually represent stable building areas, the average grid occupancy probability and hit rate of these regions are statistically analyzed to determine whether the localization is successful. This can avoid the influence of moving objects and changes in the map environment in the first-layer point cloud above ground on the accuracy of pose determination. Compared with the method of statistically analyzing all points, this can improve the robustness of localization reliability and reduce the occurrence of false positives and false negatives in localization reliability.

[0037] Specifically, after dividing the second-layer point cloud on the ground into multiple second-layer block point clouds, the average grid occupancy probability and hit rate of each second-layer block point cloud on the map can be statistically analyzed.

[0038] For example, we can first determine the grid corresponding to each laser point in the second-level block point cloud on the map, that is, the map grid where each laser point is located. Further, we can obtain the occupancy probability of the grid corresponding to each laser point. The occupancy probability is used to describe the possibility that the grid is occupied by an object. The higher the occupancy probability, the more likely the grid is to be occupied by an object. Then, we average the occupancy probabilities of all laser points to obtain the average grid occupancy probability.

[0039] Furthermore, based on the occupancy probability of the grid corresponding to each laser point, the laser points whose corresponding grids are occupied by objects can be determined among each laser point. For example, grids with an occupancy probability greater than a preset value (such as 90%) can be determined as grids occupied by objects, and then the laser points located in grids occupied by objects can be determined as laser points whose corresponding grids are occupied by objects, i.e., hit laser points. Further, the ratio of the number of hit laser points to the total number of laser points in the second-layer block point cloud on the ground can be used as the hit rate of the second-layer block point cloud on the ground.

[0040] In this embodiment, the purpose of determining the average grid occupancy probability and hit rate of the two-layer block point cloud in each region is to facilitate the determination of the positioning judgment result of the current frame pose, that is, to determine whether the current frame pose is successfully positioned.

[0041] In one specific implementation, before dividing the ground second-layer point cloud in the current frame point cloud into multiple ground second-layer block point clouds, the method further includes: determining whether the height of the ground point cloud in the current frame point cloud matches the height of the ground in the map; if so, then performing the operation of dividing the ground second-layer point cloud in the current frame point cloud into multiple ground second-layer block point clouds; otherwise, determining the positioning judgment result of the current frame pose as positioning failure, and determining the preset confidence level as the confidence level of the current frame pose.

[0042] That is, we can first determine whether the pose localization of the current frame has failed by checking whether the height of the ground point cloud matches the height of the ground in the map. If the heights match, we can further divide the second layer of the ground point cloud and determine the localization result based on the average grid occupancy probability and hit rate of the point cloud blocks in each area. If the heights do not match, we can directly determine the localization result as a failure and use the preset confidence level as the confidence level of the pose of the current frame. The preset confidence level can be 0.

[0043] By combining the above methods with the height of the ground point cloud, we can further determine whether the pose of the current frame has been successfully localized, which further ensures the accuracy of the localization judgment result and thus ensures the reliability of the localization confidence.

[0044] S130. Based on the average grid occupancy probability and hit rate of the two-layer block point cloud in each region, determine the matching block point cloud in the two-layer block point cloud in each region.

[0045] The second-layer matching block point cloud can be any second-layer point cloud on the ground that matches the map. Specifically, second-layer point clouds on the ground with an average grid occupancy probability greater than a preset average probability can be identified as second-layer matching block point clouds; or, second-layer point clouds on the ground with a hit rate greater than a preset hit rate can be identified as second-layer matching block point clouds.

[0046] Alternatively, ground-level second-layer block point clouds with an average grid occupancy probability greater than a preset average probability and a hit rate greater than a preset hit rate can be identified as second-layer matching block point clouds.

[0047] S140. Determine the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching block, and determine the confidence level of the current frame pose based on the localization judgment result. The confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0048] The positioning result can be either successful or unsuccessful. Successful positioning indicates a high degree of matching between the current frame pose and the map, meaning the vehicle is accurately positioned. Unsuccessful positioning indicates a low degree of matching between the current frame pose and the map, meaning the vehicle is not accurately positioned.

[0049] For example, if the number of second-layer matching block point clouds does not exceed the preset number, the positioning judgment result can be determined as positioning failure; if the number of second-layer matching block point clouds exceeds the preset number, the positioning judgment result can be determined as positioning success.

[0050] In this embodiment, in order to further improve the reliability of the positioning confidence, the positioning can also be determined by combining the pose of the previous frame and the point cloud of the previous frame.

[0051] In one specific implementation, the localization judgment result of the current frame pose is determined based on the number of point clouds in the two-layer matching blocks, including the following steps:

[0052] Step 1: Obtain the point cloud and pose of the previous frame. Based on the pose of the current frame and the pose of the previous frame, transform the point cloud of the previous frame from the previous frame coordinate system to the current frame coordinate system of the point cloud of the current frame.

[0053] Step 2: Divide the current frame point cloud and the previous frame point cloud in the XY plane to obtain the current frame block point cloud and the previous frame block point cloud. The XY plane is parallel to the ground, the Y direction in the XY plane is the direction of the vehicle's front, and the X direction in the XY plane is perpendicular to the Y direction.

[0054] Step 3: Determine the environmental distribution similarity between the current frame point cloud and the previous frame point cloud based on the current frame point cloud and the previous frame point cloud. If the environmental distribution similarity is greater than the preset similarity, the positioning judgment result is determined to be successful if the number of second-layer matching point clouds is greater than the preset number. If the environmental distribution similarity does not exceed the preset similarity, the positioning judgment result is determined to be unsuccessful.

[0055] Specifically, in step 1 above, the pose increment can be determined based on the pose of the previous frame and the pose of the current frame. Then, the pose increment is used to transform the point cloud of the previous frame from the previous frame coordinate system to the current frame coordinate system of the point cloud of the current frame. The pose increment can be understood as a position transformation matrix. By multiplying the position transformation matrix with the point cloud of the previous frame, the point cloud of the previous frame can be transformed to the current frame coordinate system.

[0056] It should be noted that the purpose of performing coordinate system transformation on the previous frame point cloud is: since it is necessary to determine the environmental distribution similarity between the current frame point cloud and the previous frame point cloud, in order to ensure the accuracy of the determined environmental distribution similarity, the two frame point clouds can be aligned in the same coordinate system first, and then the environmental distribution similarity can be calculated.

[0057] Furthermore, multiple regions can be divided in the XY plane. Each current frame block point cloud is constructed based on the laser points located in each region in the current frame point cloud, and the previous frame block point cloud is constructed based on the laser points located in each region in the previous frame point cloud, thus realizing the division of the current frame point cloud and the previous frame point cloud.

[0058] In step 3 above, the environmental distribution similarity between two frame point clouds can be further determined based on the current frame point cloud and the previous frame point cloud. The environmental distribution similarity describes the degree of similarity between the environmental distribution of the current frame point cloud and the environmental distribution of the previous frame point cloud. For example, the number of laser points with the same height in the current frame point cloud and the previous frame point cloud can be counted, and the environmental distribution similarity can be determined based on this number.

[0059] Optionally, the environmental distribution similarity between the current frame point cloud and the previous frame point cloud is determined based on the current frame block point cloud and the previous frame block point cloud, including: determining the first height feature matrix corresponding to the current frame point cloud based on the maximum height of the laser points in each current frame block point cloud, and determining the second height feature matrix corresponding to the previous frame point cloud based on the maximum height of the laser points in each previous frame block point cloud; performing matrix correlation calculation on the first height feature matrix and the second height feature matrix to obtain the environmental distribution similarity between the current frame point cloud and the previous frame point cloud.

[0060] Specifically, for each current frame point cloud block, the maximum height of the laser points can be selected from all laser point heights to obtain the maximum height of the laser points in the current frame point cloud block. Further, a first height feature matrix is ​​constructed based on the maximum height of the laser points in each current frame point cloud block. Similarly, the same method can be used to process each previous frame point cloud block to construct a second height feature matrix.

[0061] Furthermore, by constructing a first height feature matrix and a second height feature matrix, the point cloud of the current frame and the point cloud of the previous frame can be reduced to a two-dimensional matrix. Then, matrix correlation calculation can be performed on the two-dimensional matrix, and the calculation result can be used as the environmental distribution similarity, which ensures the accuracy of the environmental distribution similarity. Moreover, reducing the three-dimensional point cloud data to a two-dimensional matrix can accelerate the calculation efficiency of inter-frame comparison and improve the speed of determining the location confidence.

[0062] After obtaining the environmental distribution similarity between the current frame point cloud and the previous frame point cloud, the environmental distribution similarity can be compared with a preset similarity. If the environmental distribution similarity is greater than the preset similarity, it means that the environmental distribution similarity between the current frame point cloud and the previous frame point cloud is high. In this case, the localization judgment result can be determined as successful if the number of second-layer matching block point clouds is greater than the preset number. If the environmental distribution similarity does not exceed the preset similarity, it means that the environmental distribution similarity between the current frame point cloud and the previous frame point cloud is low. In this case, the localization judgment result can be directly determined as failed.

[0063] In steps 1-3 above, the historical prior of the previous frame is introduced to assist in determining the localization result of the pose of the current frame. Compared with using only the single-frame information of the current frame, the accuracy is higher, which further ensures the accuracy of the localization confidence.

[0064] Furthermore, after obtaining the positioning judgment result, the confidence level of the pose of the current frame can be determined by combining the positioning judgment result.

[0065] In one specific implementation, determining the confidence level of the current frame pose based on the positioning judgment result includes: if the positioning judgment result of the current frame pose is successful, then determining the confidence level of the current frame pose based on the average grid occupancy probability and hit rate of each second-layer matching block point cloud; if the positioning judgment result of the current frame pose is unsuccessful, then determining the preset confidence level as the confidence level of the current frame pose.

[0066] Specifically, if the localization judgment result is localization failure, the pre-set confidence level can be taken as the confidence level of the current frame pose. If the localization judgment result is localization success, the confidence level of the current frame pose can be calculated by combining the average grid occupancy probability and hit rate of each second-layer matching block point cloud.

[0067] In one example, the confidence level of the current frame pose is determined based on the average grid occupancy probability and hit rate of each two-layer matching block point cloud, including: using the hit rate of each two-layer matching block point cloud as a weight, and performing a weighted sum with the average grid occupancy probability corresponding to each two-layer matching block point cloud to obtain the confidence level of the current frame pose.

[0068] That is, the hit rate can be used as a weight, multiplied by the average occupancy probability of the corresponding grid, and then the results of multiplying all the point clouds of the two-layer matching blocks are summed to obtain the confidence of the current frame pose. In this way, the confidence of the current frame pose can be calculated by combining the hit rate, further ensuring the accuracy of the positional confidence.

[0069] In another example, the confidence of the current frame pose is determined based on the average grid occupancy probability and hit rate of each second-layer matching block point cloud. This includes: dividing the ground first-layer point cloud in the current frame point cloud into multiple ground first-layer block point clouds; for each ground first-layer block point cloud, determining the average grid occupancy probability and hit rate of the ground first-layer block point cloud on the map; determining a first-layer matching block point cloud in each ground first-layer block point cloud based on the average grid occupancy probability and hit rate of each ground first-layer block point cloud; and using the hit rate of each first-layer matching block point cloud and each second-layer matching block point cloud as weights, performing a weighted sum with the average grid occupancy probability of each first-layer matching block point cloud and each second-layer matching block point cloud to obtain the confidence of the current frame pose.

[0070] That is, the first layer of point cloud on the ground can be divided into multiple first-layer block point clouds on the XY plane. The division method can refer to the division method of the second layer of point cloud on the ground. Furthermore, the average grid occupancy probability and hit rate of each first-layer block point cloud can be calculated to determine one of the matching block point clouds.

[0071] Furthermore, the hit rates of the first-layer and second-layer matching block point clouds can be used as weights and multiplied by the average occupancy probability of the corresponding grid. Then, the results of all multiplications can be summed to obtain the confidence level of the current frame pose.

[0072] In the example above, the confidence level is further guaranteed by determining a layer of matching block point cloud and combining the average grid occupancy probability and hit rate of the layer of matching block point cloud.

[0073] In this embodiment, the positioning accuracy of the current frame pose can be obtained by determining the confidence level of the current frame pose. The higher the confidence level, the stronger the positioning accuracy of the current frame pose.

[0074] The positioning accuracy determination method provided in this embodiment acquires the current frame point cloud and the current frame pose, transforms the current frame point cloud to the world coordinate system based on the current frame pose, and then divides the ground-level second-layer point cloud in the current frame point cloud into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, it determines the average grid occupancy probability and hit rate on the map. Second-layer matching block point clouds are selected based on the average grid occupancy probability and hit rate. The positioning judgment result of the current frame pose is determined based on the number of second-layer matching block point clouds. Therefore, the positioning judgment result is used to determine the positioning accuracy. The confidence level of the current frame pose is determined to filter out potentially unchanging map regions such as the second layer of point cloud on the ground. The confidence level of localization is calculated using the average grid occupancy probability and hit rate of the unchanging map regions. This solves the problem of false positives and false negatives in localization confidence in existing technologies, improves the robustness of localization confidence, and enables a reliable judgment on whether the localization of the current frame pose is accurate. Furthermore, the method outputs the confidence level through two stages: determining the localization judgment result and determining the confidence level based on the localization judgment result. This makes the confidence level more discriminative in expressing whether the pose is accurate.

[0075] For example, Figure 2 This is a schematic diagram illustrating a positioning accuracy determination process provided in an embodiment of the present disclosure, such as... Figure 2As shown, the environmental distribution similarity can be calculated first. If the environmental distribution similarity is greater than a preset similarity, the current frame point cloud is further divided according to the height coordinates of the laser points. It is then determined whether the height of the ground point cloud matches the ground height in the map. If they match, the second-layer point cloud on the ground is further divided into multiple second-layer block point clouds. The average grid occupancy probability and hit rate of each second-layer block point cloud are calculated, and second-layer matching block point clouds are selected. If the number of second-layer matching block point clouds is greater than a preset number, the confidence level is calculated based on the average grid occupancy probability and hit rate of the second-layer block point clouds. If the environmental distribution similarity does not exceed the preset similarity, the height of the ground point cloud in the current frame does not match the ground height in the map, and the number of second-layer matching block point clouds does not exceed the preset number, 0 is directly output as the confidence level of the current frame pose.

[0076] Compared to methods that rely solely on the average occupancy probability to determine the accuracy of the localization pose, this method employs multi-layered judgment logic to achieve more accurate localization reliability.

[0077] Furthermore, when the positioning judgment result is positioning failure, 0 is output directly. When the positioning judgment result is positioning success, the confidence score is calculated by selectively using the hit rate weighted by a portion of the point cloud. Compared with the method of using the average occupancy probability as the positioning confidence score, the confidence score can be more discriminative in expressing the pose accuracy.

[0078] Furthermore, by using height-level layering and dividing the point cloud into blocks according to the XY plane, some potentially unchanging map areas can be screened out. Compared with the background technique that uses the average occupancy probability calculated from all points as the positioning reliability without discrimination, this method can avoid the influence of moving objects on the first layer of the ground and changes in the map environment on the accuracy of the pose judgment, improve the robustness of positioning reliability, and reduce false positives and false negatives in positioning reliability.

[0079] Figure 3 This is a schematic diagram of a positioning accuracy determination device according to an embodiment of this disclosure. Figure 3 As shown: The device includes: a transformation module 310, a statistics module 320, a matching module 330, and a determination module 340.

[0080] Transformation module 310 is used to acquire the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0081] The statistics module 320 is used to divide the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. For each ground second layer block point cloud, the average grid occupancy probability and hit rate of the ground second layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the ground second layer block point cloud whose corresponding grid is occupied by an object to the total number of laser points in the ground second layer block point cloud.

[0082] Matching module 330 is used to determine a second-layer matching block point cloud in each of the above-ground second-layer block point clouds based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds.

[0083] The determination module 340 is used to determine the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching block, and to determine the confidence level of the current frame pose based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0084] Optionally, the determining module 340 is further configured to determine the confidence level of the current frame pose based on the average grid occupancy probability and hit rate corresponding to each of the two-layer matching block point clouds if the positioning judgment result of the current frame pose is successful, and to determine the preset confidence level as the confidence level of the current frame pose if the positioning judgment result of the current frame pose is unsuccessful.

[0085] Optionally, the matching module 330 is further configured to use the hit rate of each of the two-layer matching block point clouds as a weight, and perform a weighted summation with the average grid occupancy probability corresponding to each of the two-layer matching block point clouds to obtain the confidence level of the current frame pose.

[0086] Optionally, the matching module 330 is further configured to divide the ground first layer point cloud in the current frame point cloud into multiple ground first layer block point clouds; for each ground first layer block point cloud, determine the average grid occupancy probability and hit rate of the ground first layer block point cloud on the map; based on the average grid occupancy probability and hit rate of each ground first layer block point cloud, determine a matching block point cloud in each ground first layer block point cloud; and use the hit rate of each first layer matching block point cloud and each second layer matching block point cloud as weights to perform a weighted sum with the average grid occupancy probability of each first matching block point cloud and each second layer matching block point cloud to obtain the confidence level of the current frame pose.

[0087] Optionally, the device further includes a ground matching module, which is used to determine whether the height of the ground point cloud in the current frame point cloud matches the height of the ground in the map. If so, the module performs the operation of dividing the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. Otherwise, the module determines that the positioning judgment result of the current frame pose is a positioning failure and determines the preset confidence level as the confidence level of the current frame pose.

[0088] Optionally, the determining module 340 includes a positioning judgment submodule, which includes a conversion unit, a division unit, and a distribution judgment unit, wherein:

[0089] The conversion unit is used to acquire the point cloud of the previous frame and the pose of the previous frame, and based on the pose of the current frame and the pose of the previous frame, convert the point cloud of the previous frame from the coordinate system of the previous frame to the coordinate system of the current frame point cloud.

[0090] The partitioning unit is used to partition the current frame point cloud and the previous frame point cloud in the XY plane to obtain each current frame block point cloud in the current frame point cloud and each previous frame block point cloud in the previous frame point cloud. The XY plane is parallel to the ground, the Y direction in the XY plane is the direction of the vehicle's front, and the X direction in the XY plane is perpendicular to the Y direction.

[0091] The distribution judgment unit is used to determine the environmental distribution similarity between the current frame point cloud and the previous frame point cloud based on the current frame block point cloud and the previous frame block point cloud. If the environmental distribution similarity is greater than a preset similarity, the positioning judgment result is determined to be successful if the number of second-layer matching block point clouds is greater than a preset number. If the environmental distribution similarity does not exceed the preset similarity, the positioning judgment result is determined to be unsuccessful.

[0092] Optionally, the distribution determination unit is specifically used for:

[0093] The first height feature matrix corresponding to the current frame point cloud is determined based on the maximum height of the laser points in each current frame block point cloud, and the second height feature matrix corresponding to the previous frame point cloud is determined based on the maximum height of the laser points in each previous frame block point cloud. Matrix correlation calculation is performed on the first height feature matrix and the second height feature matrix to obtain the environmental distribution similarity between the current frame point cloud and the previous frame point cloud.

[0094] The positioning accuracy determination device provided in this embodiment can execute the steps in the positioning accuracy determination method provided in this embodiment, and has the execution steps and beneficial effects, which will not be repeated here.

[0095] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this disclosure. See below for details. Figure 4 It shows a schematic diagram of a structure suitable for implementing the electronic device 500 in the embodiments of this disclosure. Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0096] like Figure 4 As shown, the electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 501, which can perform various appropriate actions and processes to implement the methods of the embodiments described herein, based on a program stored in a read-only memory (ROM) 502 or a program loaded from a storage device 508 into a random access memory (RAM) 503. The RAM 503 also stores various programs and data required for the operation of the electronic device 500. The processing device 501, ROM 502, and RAM 503 are interconnected via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0097] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts, thereby implementing the positioning method as described above. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 509, or installed from a storage device 508, or installed from a ROM 502. When the computer program is executed by the processing device 501, it performs the functions defined in the methods of embodiments of this disclosure.

[0098] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0099] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device. The aforementioned computer-readable medium carries one or more programs that, when executed by the electronic device, cause the electronic device to:

[0100] Obtain the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0101] The ground-level second-layer point cloud in the current frame point cloud is divided into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, the average grid occupancy probability and hit rate of the ground-level second-layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the corresponding grid of the ground-level second-layer block point cloud that are occupied by objects to the total number of laser points in the ground-level second-layer block point cloud.

[0102] Based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds, a second-layer matching block point cloud is determined in each of the above-ground second-layer block point clouds;

[0103] The localization judgment result of the current frame pose is determined based on the number of point clouds in the second-layer matching block, and the confidence level of the current frame pose is determined based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0104] Optionally, when one or more of the above-described procedures are executed by the electronic device, the electronic device may also execute other steps described in the above embodiments.

[0105] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0106] Option 1: A method for determining the accuracy of positioning, the method comprising:

[0107] Obtain the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0108] The ground-level second-layer point cloud in the current frame point cloud is divided into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, the average grid occupancy probability and hit rate of the ground-level second-layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the corresponding grid of the ground-level second-layer block point cloud that are occupied by objects to the total number of laser points in the ground-level second-layer block point cloud.

[0109] Based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds, a second-layer matching block point cloud is determined in each of the above-ground second-layer block point clouds;

[0110] The localization judgment result of the current frame pose is determined based on the number of point clouds in the second-layer matching block, and the confidence level of the current frame pose is determined based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0111] Option 2: According to the method described in Option 1, determining the confidence level of the current frame pose based on the positioning judgment result includes:

[0112] If the localization judgment result of the current frame pose is successful, the confidence of the current frame pose is determined based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds. If the localization judgment result of the current frame pose is unsuccessful, the preset confidence is determined as the confidence of the current frame pose.

[0113] Solution 3: According to the method described in Solution 2, the step of determining the confidence level of the current frame pose based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds includes:

[0114] The hit rate of each of the two-layer matching block point clouds is used as a weight, and the weighted sum is performed with the average grid occupancy probability corresponding to each of the two-layer matching block point clouds to obtain the confidence level of the current frame pose.

[0115] Option 4: According to the method described in Option 2, the step of determining the confidence level of the current frame pose based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds includes:

[0116] The first ground layer point cloud in the current frame point cloud is divided into multiple ground layer block point clouds. For each ground layer block point cloud, the average grid occupancy probability and hit rate of the ground layer block point cloud on the map are determined. Based on the average grid occupancy probability and hit rate of each ground layer block point cloud, a matching block point cloud is determined in each ground layer block point cloud.

[0117] The hit rate of each of the first-layer matching block point clouds and each of the second-layer matching block point clouds is used as a weight, and the average grid occupancy probability of each of the first-layer matching block point clouds and each of the second-layer matching block point clouds is weighted and summed to obtain the confidence level of the current frame pose.

[0118] Option 5: According to the method described in Option 1, before dividing the ground-level second-layer point cloud in the current frame point cloud into multiple ground-level second-layer block point clouds, the method further includes:

[0119] Determine whether the height of the ground point cloud in the current frame point cloud matches the height of the ground in the map. If yes, perform the operation of dividing the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. Otherwise, determine that the positioning judgment result of the current frame pose is positioning failure, and determine the preset confidence level as the confidence level of the current frame pose.

[0120] Solution 6: According to the method described in Solution 1, the step of determining the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching blocks includes:

[0121] Obtain the point cloud and pose of the previous frame. Based on the pose of the current frame and the pose of the previous frame, transform the point cloud of the previous frame from the previous frame coordinate system to the current frame coordinate system of the point cloud of the current frame.

[0122] The current frame point cloud and the previous frame point cloud are divided in the XY plane to obtain each current frame block point cloud in the current frame point cloud and each previous frame block point cloud in the previous frame point cloud. The XY plane is parallel to the ground, the Y direction in the XY plane is the direction of the vehicle's front, and the X direction in the XY plane is perpendicular to the Y direction.

[0123] The environmental distribution similarity between the current frame point cloud and the previous frame point cloud is determined based on the current frame point cloud and the previous frame point cloud. If the environmental distribution similarity is greater than a preset similarity, the positioning judgment result is determined to be successful if the number of second-layer matching point clouds is greater than a preset number. If the environmental distribution similarity does not exceed the preset similarity, the positioning judgment result is determined to be unsuccessful.

[0124] Solution 7: According to the method described in Solution 6, determining the environmental distribution similarity between the current frame point cloud and the previous frame point cloud based on each current frame block point cloud and each previous frame block point cloud includes:

[0125] The first height feature matrix corresponding to the current frame point cloud is determined based on the maximum height of the laser points in each current frame block point cloud, and the second height feature matrix corresponding to the previous frame point cloud is determined based on the maximum height of the laser points in each previous frame block point cloud.

[0126] Matrix correlation calculation is performed on the first height feature matrix and the second height feature matrix to obtain the environmental distribution similarity between the current frame point cloud and the previous frame point cloud.

[0127] Option 8: A device for determining the accuracy of positioning, comprising:

[0128] The transformation module is used to acquire the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose;

[0129] The statistics module is used to divide the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. For each ground second layer block point cloud, the average grid occupancy probability and hit rate of the ground second layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the ground second layer block point cloud whose corresponding grid is occupied by an object to the total number of laser points in the ground second layer block point cloud.

[0130] The matching module is used to determine the second-layer matching block point cloud in each of the above-ground second-layer block point clouds based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds.

[0131] The determination module is used to determine the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching block, and to determine the confidence level of the current frame pose based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

[0132] Option 9: An electronic device, the electronic device comprising:

[0133] One or more processors;

[0134] Storage device for storing one or more programs;

[0135] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of schemes 1-7.

[0136] Option 10: A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of Options 1-7.

[0137] The above description is merely a preferred embodiment of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features disclosed in this disclosure that have similar functions.

Claims

1. A method for determining the accuracy of positioning, characterized in that, The method includes: Obtain the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose; The ground-level second-layer point cloud in the current frame point cloud is divided into multiple ground-level second-layer block point clouds. For each ground-level second-layer block point cloud, the average grid occupancy probability and hit rate of the ground-level second-layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the corresponding grid of the ground-level second-layer block point cloud that are occupied by objects to the total number of laser points in the ground-level second-layer block point cloud. Based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds, a second-layer matching block point cloud is determined in each of the above-ground second-layer block point clouds; The localization result of the current frame pose is determined based on the number of point clouds in the second-layer matching block. If the localization judgment result of the current frame pose is successful, the confidence level of the current frame pose is determined based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds. If the localization judgment result of the current frame pose is unsuccessful, the preset confidence level is determined as the confidence level of the current frame pose. The confidence level of the current frame pose reflects the localization accuracy of the current frame pose.

2. The method according to claim 1, characterized in that, The determination of the confidence level of the current frame pose based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds includes: The hit rate of each of the two-layer matching block point clouds is used as a weight, and the weighted sum is performed with the average grid occupancy probability corresponding to each of the two-layer matching block point clouds to obtain the confidence level of the current frame pose.

3. The method according to claim 1, characterized in that, The determination of the confidence level of the current frame pose based on the average grid occupancy probability and hit rate of each of the two-layer matching block point clouds includes: The first ground layer point cloud in the current frame point cloud is divided into multiple ground layer block point clouds. For each ground layer block point cloud, the average grid occupancy probability and hit rate of the ground layer block point cloud on the map are determined. Based on the average grid occupancy probability and hit rate of each ground layer block point cloud, a matching block point cloud is determined in each ground layer block point cloud. The hit rate of each of the first-layer matching block point clouds and each of the second-layer matching block point clouds is used as a weight, and the average grid occupancy probability of each of the first-layer matching block point clouds and each of the second-layer matching block point clouds is weighted and summed to obtain the confidence level of the current frame pose.

4. The method according to claim 1, characterized in that, Before dividing the ground-level second-layer point cloud in the current frame point cloud into multiple ground-level second-layer block point clouds, the method further includes: Determine whether the height of the ground point cloud in the current frame point cloud matches the height of the ground in the map. If yes, perform the operation of dividing the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. Otherwise, determine that the positioning judgment result of the current frame pose is positioning failure, and determine the preset confidence level as the confidence level of the current frame pose.

5. The method according to claim 1, characterized in that, The step of determining the localization result of the current frame pose based on the number of point clouds in the second-layer matching block includes: Obtain the point cloud and pose of the previous frame. Based on the pose of the current frame and the pose of the previous frame, transform the point cloud of the previous frame from the previous frame coordinate system to the current frame coordinate system of the point cloud of the current frame. The current frame point cloud and the previous frame point cloud are divided in the XY plane to obtain each current frame block point cloud in the current frame point cloud and each previous frame block point cloud in the previous frame point cloud. The XY plane is parallel to the ground, the Y direction in the XY plane is the direction of the vehicle's front, and the X direction in the XY plane is perpendicular to the Y direction. The environmental distribution similarity between the current frame point cloud and the previous frame point cloud is determined based on the current frame point cloud and the previous frame point cloud. If the environmental distribution similarity is greater than a preset similarity, the positioning judgment result is determined to be successful if the number of second-layer matching point clouds is greater than a preset number. If the environmental distribution similarity does not exceed the preset similarity, the positioning judgment result is determined to be unsuccessful.

6. The method according to claim 5, characterized in that, The step of determining the environmental distribution similarity between the current frame point cloud and the previous frame point cloud based on each current frame block point cloud and each previous frame block point cloud includes: The first height feature matrix corresponding to the current frame point cloud is determined based on the maximum height of the laser points in each current frame block point cloud, and the second height feature matrix corresponding to the previous frame point cloud is determined based on the maximum height of the laser points in each previous frame block point cloud. Matrix correlation calculation is performed on the first height feature matrix and the second height feature matrix to obtain the environmental distribution similarity between the current frame point cloud and the previous frame point cloud.

7. A device for determining the accuracy of positioning, characterized in that, include: The transformation module is used to acquire the current frame point cloud and the current frame pose, and transform the current frame point cloud to the world coordinate system based on the current frame pose; The statistics module is used to divide the ground second layer point cloud in the current frame point cloud into multiple ground second layer block point clouds. For each ground second layer block point cloud, the average grid occupancy probability and hit rate of the ground second layer block point cloud on the map are determined. The hit rate is the ratio of the number of laser points in the ground second layer block point cloud whose corresponding grid is occupied by an object to the total number of laser points in the ground second layer block point cloud. The matching module is used to determine the second-layer matching block point cloud in each of the above-ground second-layer block point clouds based on the average grid occupancy probability and hit rate of each of the above-ground second-layer block point clouds. The determination module is used to determine the localization judgment result of the current frame pose based on the number of point clouds in the second-layer matching block, and to determine the confidence level of the current frame pose based on the localization judgment result, wherein the confidence level of the current frame pose reflects the localization accuracy of the current frame pose. The determining module is further configured to, if the positioning judgment result of the current frame pose is successful, determine the confidence level of the current frame pose based on the average grid occupancy probability and hit rate corresponding to each of the two-layer matching block point clouds; if the positioning judgment result of the current frame pose is unsuccessful, determine the preset confidence level as the confidence level of the current frame pose.

8. An electronic device, characterized in that, The electronic device includes: One or more processors; Storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-6.

9. 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 1-6.

Images