Lane mapping method for memorizing driving, electronic equipment and storage medium

Abstract

The invention provides a lane mapping method for memorizing driving, electronic equipment and a storage medium, and the method comprises the steps: obtaining the lane line information of a current frame, determining a lane line point set according to the lane line information, dividing the lane line point set into a plurality of effective region subsets according to the distance between points in the lane line point set and a vehicle, according to the method, the points in the effective area subsets are optimized through the optimization strategies corresponding to the effective area subsets, the optimized point set of the current frame is obtained, and the memory driving map point set is updated according to the optimized point set of the current frame, so that lane mapping for memory driving is realized, the problems of high mapping cost, poor effectiveness, high resource consumption and the like are solved, and the method is suitable for being popularized and applied. The method does not depend on a specific mathematical model, guarantees the isolation of data, avoids the influence of local detection errors on the global accuracy, can achieve the optimization process from coarse to fine, and improves the mapping stability and the mapping accuracy.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a lane mapping method, electronic device, and storage medium for memorizing driving lanes. Background Art

[0002] Memory Driving is a vision-based autonomous driving technology that uses onboard cameras and sensors to capture information about the vehicle's surroundings and matches this information with map data, enabling autonomous driving. Efficiently and quickly building a map using detected lane lines is a key technology in Memory Driving.

[0003] Existing lane mapping technology relies primarily on high-precision maps to provide prior knowledge, aligning lane lines detected by vision or other sensors onto the HD map. However, this approach has the following problems: 1. Mapping algorithms that rely on HD maps have high maintenance costs, poor effectiveness, and require additional resources to align maps with lane lines from other sources; 2. In complex traffic environments, lane line detection and recognition may be disrupted, affecting the accuracy and reliability of mapping; 3. Existing mapping algorithms typically require a large amount of computing resources and time, which is a challenge for autonomous driving applications with high real-time requirements. Summary of the Invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, the present application aims to provide a lane mapping method, electronic device and storage medium for memorizing driving lanes, so as to solve the problems of high mapping cost, poor effectiveness, low mapping accuracy and inability to meet real-time requirements in the prior art.

[0005] The present application provides a lane mapping method for memory driving, the method comprising:

[0006] Obtaining lane line information of the current frame, and determining a lane line point set based on the lane line information;

[0007] Dividing the lane point set into a plurality of valid area subsets according to the distance between the points in the lane point set and the vehicle;

[0008] Based on the optimization strategies corresponding to the respective effective area subsets, the points in the effective area subsets are optimized to obtain the optimized point set of the current frame;

[0009] The memory driving map point set is updated based on the optimized point set of the current frame.

[0010] Optionally, after determining the lane line point set based on the lane line information, the method further includes:

[0011] Obtaining the vehicle pose of the current frame;

[0012] Converting the vehicle posture from the local coordinate system corresponding to the current frame to the reference coordinate system to obtain the relative posture of the vehicle body;

[0013] Based on the relative position of the vehicle body, the points in the lane line point set are converted from the local coordinate system to a reference coordinate system.

[0014] Optionally, after obtaining the lane information of the current frame, the following steps are also included:

[0015] For each lane line in the lane line information, determine whether the lane line is associated with other lane lines in the previous frame; if so, determine the number of the associated other lane line as the number of the lane line; otherwise, assign a new number to the lane line;

[0016] After updating the memory driving map point set based on the optimized point set of the current frame, the method further includes:

[0017] Based on the numbers corresponding to the points in the memory driving map point set, points with the same numbers between different frames are determined to be points under the same lane line.

[0018] Optionally, before determining the lane line point set based on the lane line information, the method further includes:

[0019] For each lane line in the lane line information, determine whether the lane line intersects or branches with other adjacent lane lines, and if so, determine a branch intersection point;

[0020] At the branch intersection, each lane line is split into multiple line segments;

[0021] The proximal and distal segments of the lane where the vehicle is located are fitted in all line segments to obtain a new lane line, and the remaining line segments are used as the new lane line.

[0022] Optionally, determining whether the lane line intersects or branches with other adjacent lane lines includes:

[0023] determining a first lateral distance between a starting point of the lane line and a starting point of another adjacent lane line, and if the first lateral distance is less than a preset first threshold, determining that a branch exists between the lane line and the other adjacent lane line; and

[0024] Determine a second lateral distance between the end point of the lane line and the end point of another adjacent lane line. If the second lateral distance is less than a preset second threshold, determine that there is an intersection between the lane line and the other adjacent lane line.

[0025] Optionally, the lane line point set is divided into multiple valid area subsets, including:

[0026] Dividing the lane line point set into a first valid area subset and a second valid area subset, wherein the distance between the points in the first valid area subset and the vehicle is greater than the distance between the points in the second valid area subset and the vehicle;

[0027] Based on the optimization strategies corresponding to the respective effective area subsets, the points within the effective area subsets are optimized, including:

[0028] For the first valid area subset, determine all outliers in the first valid area subset, remove all outliers from the first valid area subset, perform curve fitting on the first valid area subset after removal to obtain a first fitting curve, and perform discrete sampling on the first fitting curve to obtain a corresponding optimized point set;

[0029] For the second valid area subset, curve fitting is performed on the second valid area subset to obtain a second fitting curve, and based on the distance between each point in the second valid area subset and the second fitting curve, some points in the second valid area subset are eliminated to obtain a corresponding optimized point set.

[0030] Optionally, determining all outliers within the first valid area subset includes:

[0031] Randomly sampling the first valid area subset to obtain a current sample subset, and performing curve fitting on the current sample subset to obtain a current sample curve;

[0032] For a remainder set other than the current sample subset within the first valid area subset, determining a distance between each point in the remainder set and the current sample curve, determining a current inlier set based on points whose distances are less than a preset distance threshold, and determining the number of inliers in the current inlier set;

[0033] Return to the step of randomly selecting the first valid area subset until the number of rounds reaches the set number, and select the sample curve with the largest number of inliers as the target curve;

[0034] The target curve is updated based on the set of interior points corresponding to the target curve, and all exterior points in the first valid area subset are determined according to the updated target curve.

[0035] Optionally, before dividing the lane point set into a plurality of valid area subsets based on the distances between points in the lane point set and the vehicle, the method further includes:

[0036] Determine whether the vehicle's data collection frequency is greater than a preset frequency threshold;

[0037] If so, determining whether the frame count value corresponding to the current frame reaches a preset optimized frame number; if so, executing the step of dividing the lane line point set into a plurality of valid area subsets based on the distance between the points in the lane line point set and the vehicle; otherwise, accumulating the frame count value;

[0038] After optimizing the points in the valid area subsets based on the optimization strategies corresponding to the valid area subsets, the method further includes:

[0039] The frame count value is cleared.

[0040] The present application also provides a lane mapping device for memorizing driving, the device comprising:

[0041] A lane line acquisition module is used to obtain lane line information of the current frame and determine a lane line point set based on the lane line information;

[0042] a point set partitioning module, configured to divide the lane line point set into a plurality of valid area subsets according to the distance between the points in the lane line point set and the vehicle;

[0043] A partition optimization module, configured to optimize the points in the valid area subset based on the optimization strategies corresponding to the valid area subsets, to obtain an optimized point set for the current frame;

[0044] The optimization output module is used to update the memory driving map point set based on the optimization point set of the current frame.

[0045] An embodiment of the present application further provides an electronic device, comprising:

[0046] processor and memory;

[0047] The processor is used to execute the steps of the lane mapping method for memorizing driving provided in any embodiment of the present application by calling the program or instructions stored in the memory.

[0048] An embodiment of the present application further provides a computer-readable storage medium storing a program or instruction, wherein the program or instruction enables a computer to execute the steps of the lane mapping method for memorizing driving provided in any embodiment of the present application.

[0049] In summary, the present application proposes a lane mapping method for memory driving. The method obtains lane line information of the current frame, determines a lane line point set based on the lane line information, and divides the lane line point set into multiple valid area subsets based on the distance between the points in the lane line point set and the vehicle. The points therein are optimized using the optimization strategy corresponding to each valid area subset to obtain the optimized point set of the current frame. The memory driving map point set is then updated based on the optimized point set of the current frame, thereby realizing lane mapping for memory driving without relying on high-precision maps. This solves the problems of high mapping cost, poor effectiveness, and high resource consumption. By extracting lane line point sets from lane line information, the method can discretize the road model into point sets, does not rely on a specific mathematical model, ensures the isolation between data, and avoids the impact of local detection errors on global accuracy. In addition, the method divides the point set into multiple subsets and optimizes different subsets separately, so that points in different areas are optimized in different ways, which can achieve a coarse-to-fine optimization process and improve mapping stability and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0051] Figure 1 This is a flow chart of a lane mapping method for memorizing driving provided in an embodiment of the present application;

[0052] Figure 2 This is a flow chart of establishing a vehicle motion trajectory provided by an embodiment of the present application;

[0053] Figure 3 This is a schematic diagram of a branching situation provided in an embodiment of the present application;

[0054] Figure 4 This is a schematic diagram of lane line reconstruction in a branching situation provided by an embodiment of the present application;

[0055] Figure 5 This is a flow chart for generating a lane line point set provided by an embodiment of the present application;

[0056] Figure 6 This is an optimized flow chart provided in an embodiment of the present application;

[0057] Figure 7 This is a schematic structural diagram of a lane mapping device for memorizing driving provided in an embodiment of the present application;

[0058] Figure 8 This is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0059] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0060] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in this application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0061] As mentioned in the background art, to address the problems in the prior art, this application proposes a lane mapping method for memory driving. Figure 1 This is a flow chart of a lane mapping method for memory driving provided by an embodiment of the present application. Figure 1 The lane mapping method for memory driving specifically includes:

[0062] S110 , obtaining lane line information of the current frame, and determining a lane line point set based on the lane line information.

[0063] The lane information may describe all lanes located near the vehicle, for example, the lanes may be described in the form of a cubic equation. The lane information may be determined based on camera detection data and / or sensor detection data of the current frame.

[0064] For example, at least one of the camera detection data and the sensor detection data can be input into a pre-trained lane detection model to obtain the lane information for the current frame output by the model. In addition to obtaining the lane information for the current frame, the vehicle pose, that is, the vehicle body motion information for the current frame, can also be obtained.

[0065] Specifically, after obtaining lane line information, each lane line can be numbered and associated with the lane line of the previous frame. For example, if the current frame is the starting frame, the lane lines of the current frame can be numbered consecutively starting from the drivable area side based on the vehicle position. If the current frame is not the starting frame, it can be determined whether the lane lines of the current frame are associated with the lane lines of the previous frame, that is, whether the lane lines of the current frame and the lane lines of the previous frame belong to the same lane line. If so, the lane lines of the current frame can inherit the number of the lane lines of the previous frame with which they are associated. If not, the lane lines of the current frame can be assigned new numbers.

[0066] In addition, each lane line in the current frame can be uniformly sampled to obtain a discrete lane line point set, which can be used as a control point set. Furthermore, all points in the lane line point set can be transformed into a reference coordinate system. The reference coordinate system can be the local coordinate system corresponding to the starting frame or the ending frame.

[0067] In a specific embodiment, after determining the lane line point set based on the lane line information, the following steps are further included:

[0068] Step 11: Get the vehicle pose of the current frame;

[0069] Step 12: Convert the vehicle pose from the local coordinate system corresponding to the current frame to the reference coordinate system to obtain the relative pose of the vehicle body;

[0070] Step 13: Based on the relative position of the vehicle body, the points in the lane line point set are converted from the local coordinate system to the reference coordinate system.

[0071] In step 11, the vehicle pose of the current frame can be obtained, and the vehicle pose includes the vehicle position and the vehicle posture. Furthermore, in step 12, the vehicle pose of the starting frame or the vehicle pose of the ending frame can be used as the reference pose, and the local coordinate system corresponding to the starting frame or the ending frame can be used as the reference coordinate system. Then, the vehicle pose of the current frame is converted from the corresponding local coordinate system to the reference coordinate system, and the offset of the vehicle pose of the current frame relative to the reference pose is obtained as the relative pose of the vehicle body in the current frame.

[0072] For example, Figure 2 This is a flow chart of establishing a vehicle motion trajectory provided by an embodiment of the present application, such as Figure 2 As shown, the vehicle position of the start frame or the end frame can be set as the reference state, and the corresponding local coordinate system is the reference coordinate system. Then, the odometer of the vehicle position in each frame is calculated, and a transformation matrix is constructed to transform the vehicle motion information of each frame into the reference coordinate system.

[0073] After converting the vehicle posture of the current frame to the reference coordinate system, further, in step 13, the points in the lane line point set can be converted from the local coordinate system corresponding to the current frame to the reference coordinate system according to the relative posture of the vehicle body to obtain the position of the points in the lane line point set of the current frame relative to the reference posture.

[0074] Through steps 11 to 13 above, the relative posture of the vehicle body in the current frame relative to the reference posture can be determined first. Then, combined with the relative posture of the vehicle body, the lane line point set of the current frame is converted from the vehicle posture to the reference posture, ensuring the position accuracy of the points in the lane line point set and thus ensuring the reliability of subsequent optimization.

[0075] In an embodiment of the present application, taking into account the possibility of branches or intersections between adjacent lane lines, in order to improve the accuracy of the lane lines and avoid partial line segments in an actual lane line being identified as other lane lines in the case of branches or intersections, it is also possible to identify branches and intersections before uniformly sampling the lane lines of the current frame, and reconstruct the lane lines of the current frame when there are branches or intersections.

[0076] In a specific embodiment, before determining the lane line point set based on the lane line information, the following steps are further included:

[0077] Step 21: For each lane line in the lane line information, determine whether the lane line intersects or branches with other adjacent lane lines. If so, determine the branch intersection point.

[0078] Step 22: At the branch intersection, split each lane line into multiple line segments;

[0079] Step 23: Fit the proximal and distal line segments of the lane where the vehicle is located in all line segments to obtain a new lane line, and use the remaining line segments as the new lane line.

[0080] In step 21, for each lane line in the current frame, a determination is made as to whether the lane line intersects or branches with any adjacent lane lines. A branch may refer to two lane lines transitioning from describing the same lane to describing different lanes, and an intersection may refer to two lane lines transitioning from describing different lanes to describing the same lane. For example, the determination of an intersection or branch may be based on the distance between the two lane lines.

[0081] In one example, determining whether a lane line intersects or branches with another adjacent lane line includes:

[0082] Determine a first lateral distance between a starting point of the lane line and a starting point of another adjacent lane line, and if the first lateral distance is less than a preset first threshold, determine that a branch exists between the lane line and the other adjacent lane line;

[0083] Also, a second lateral distance between the end point of the lane line and the end point of another adjacent lane line is determined. If the second lateral distance is less than a preset second threshold, it is determined that there is an intersection between the lane line and the other adjacent lane lines.

[0084] Specifically, whether a branch exists can be identified by the lateral distance between the starting points of two lane lines, and whether an intersection exists can be identified by the lateral distance between the end points of two lane lines.

[0085] For example, the lateral distance between the starting point of a lane line and the starting point of other adjacent lane lines can be calculated to obtain a first lateral distance, and then compared with a preset first threshold. If it is less than the preset first distance threshold, it means that the starting point of the lane line is close to the starting point of other adjacent lane lines. At this time, it can be determined that there is a branch between the two lane lines.

[0086] In addition, the lateral distance between the end point of the lane line and the end point of other adjacent lane lines can be calculated to obtain a second lateral distance, and compared with the preset second threshold. If it is less than the preset second threshold, it means that the end point of the lane line is close to the end point of other adjacent lane lines. At this time, it can be determined that there is an intersection between the two lane lines.

[0087] Through the above examples, accurate detection of branches and intersections can be achieved, which facilitates the subsequent reconstruction of lane lines and ensures the accuracy of lane lines.

[0088] After identifying a lane line's divergence or intersection with another adjacent lane line, the location where the divergence or intersection occurs can be determined as a divergence point. For example, the location where the lateral distance between two lane lines is the shortest can be determined as a divergence point.

[0089] For example, Figure 3 This is a schematic diagram of a branch situation provided by an embodiment of the present application, such as Figure 3 As shown, the lateral distance between the starting point of lane line a and the starting point of lane line b is small, so it can be determined that there is a branch. The lateral distance between the starting point of lane line a and lane line b is the shortest, which can be determined as the branch intersection.

[0090] After determining the branch intersection, in step 22, each lane line can be split into multiple segments at the branch intersection, that is, each lane line is interrupted at the branch intersection. Furthermore, in step 23, for each segment obtained by splitting, the proximal segment and the distal segment of the lane where the vehicle is located can be fitted to form a new lane line, and the remaining segments can also be used as new lane lines.

[0091] Continuing with the above example, Figure 4 This is a schematic diagram of lane line reconstruction in a branching situation provided by an embodiment of the present application, such as Figure 3-Figure 4 As shown, after each lane line is split at the branch intersection, lane line b and lane line c are both split into two upper and lower line segments; after lane line a is split, a line segment that is still lane line a is obtained.

[0092] Furthermore, the left proximal segment of the lane where the vehicle is located (i.e., the lower segment split from lane line b) is fitted with the left distal segment (i.e., lane line a) to obtain a new lane line ( Figure 4The left dashed line of the lane where the vehicle is located is in the middle); and the right proximal line segment of the lane where the vehicle is located (i.e., lane line f) is fitted with the right distal line segment (lane line d) to obtain a new lane line ( Figure 4 The remaining line segments can be used as new lane lines, namely the upper line segment split from lane line b and the upper and lower line segments split from lane line a.

[0093] Through the above steps 21 to 23, the lane lines can be reconstructed when there is an intersection or branch between the lane lines and other adjacent lane lines, so as to avoid part of the lane line of the lane where the vehicle is located at the branch intersection point from being assigned to other lane lines, so that the lane line description of the lane where the vehicle is located is more accurate, thereby facilitating more accurate calculation of information such as lane line curvature in the subsequent process, and improving the control reliability of the memory driving function.

[0094] After the lane lines are reconstructed, uniform sampling can be performed on the new lane lines to obtain a lane line point set, and each point in the point set is transformed into a reference coordinate system.

[0095] For example, Figure 5 This is a flow chart of generating a lane line point set provided by an embodiment of the present application, such as Figure 5 As shown in the figure, the lane line parameter equation can be obtained first, and then the lane line parameter equation can be used to determine whether there is a branch or intersection. The road is reconstructed for the branch intersection point, and a new lane line parameter equation is obtained after reconstruction. Key points are sampled on the lane line parameter equation as road control points to obtain a lane line point set. The lane line point set of the current frame is converted to the reference coordinate system and associated with the overlapping part with the previous frame.

[0096] S120 : Divide the lane point set into a plurality of valid area subsets based on the distance between the points in the lane point set and the vehicle.

[0097] Specifically, after transforming the points in the lane line point set to the reference coordinate system, the points in the lane line point set may be further partitioned and optimized.

[0098] Among them, the distance between the points in the lane line point set and the vehicle position in the current frame can be calculated first, and then the valid area where the points in the lane line point set are located can be determined according to the distance, and the points in the lane line point set can be placed in the subset of the corresponding valid area to obtain multiple valid area subsets.

[0099] For example, the entire area can be divided into a first valid area and a second valid area with the vehicle position in the current frame as the center, and corresponding first area distances and second area distances are set, with the first area distance being greater than the second area distance. If the distance between a point in the lane line point set and the vehicle position is less than the second area distance, the point can be placed in the second valid area subset. If the distance between a point in the lane line point set and the vehicle position is greater than the second area distance but less than the first area distance, the point can be placed in the first valid area subset.

[0100] S130 , based on the optimization strategies corresponding to the respective effective area subsets, optimizing the points in the effective area subsets to obtain an optimized point set of the current frame.

[0101] Specifically, corresponding optimization strategies can be pre-set for different effective areas. In the embodiment of the present application, considering that the accuracy of points close to the vehicle is usually more important during the mapping process, in order to ensure the control reliability of the memory driving function, fine optimization can be performed on points near the vehicle to make them as close to the actual road as possible. At the same time, in order to improve the efficiency of mapping and avoid excessive mapping resources, coarse optimization can be performed on points far away from the vehicle to filter out points with large errors.

[0102] Therefore, for the subset of valid areas that are close to the vehicle, a fine-precision optimization strategy can be adopted, and for the subset of valid areas that are far from the vehicle, a coarse-precision optimization strategy can be adopted.

[0103] In a specific embodiment, dividing the lane line point set into a plurality of valid area subsets includes: dividing the lane line point set into a first valid area subset and a second valid area subset, wherein a distance between a point in the first valid area subset and the vehicle is greater than a distance between a point in the second valid area subset and the vehicle;

[0104] Based on the optimization strategies corresponding to each valid area subset, the points in the valid area subset are optimized, including the following steps:

[0105] Step 31: for the first valid area subset, determine all outliers in the first valid area subset, remove all outliers from the first valid area subset, perform curve fitting on the first valid area subset after removal to obtain a first fitting curve, and perform discrete sampling on the first fitting curve to obtain a corresponding optimized point set;

[0106] Step 32: For the second effective area subset, curve fitting is performed on the second effective area subset to obtain a second fitting curve, and based on the distance between each point in the second effective area subset and the second fitting curve, some points in the second effective area subset are eliminated to obtain a corresponding optimized point set.

[0107] Specifically, for the first valid area subset that is far away from the vehicle, all external points can be screened out from the first valid area subset, wherein the external points may be points that do not belong to the first valid area subset and can be understood as outliers.

[0108] Regarding step 31 above, in some embodiments, determining all outliers within the first valid area subset includes the following steps:

[0109] Step 311: Randomly extract the first valid area subset to obtain a current sample subset, and perform curve fitting on the current sample subset to obtain a current sample curve;

[0110] Step 312: for the remainder of the first valid region subset excluding the current sample subset, determine the distance between each point in the remainder and the current sample curve, determine the current inlier set based on the points whose distance is less than a preset distance threshold, and determine the number of inliers in the current inlier set;

[0111] Step 313: Return to the step of randomly selecting the first valid area subset until the number of rounds reaches the set number, and select the sample curve with the largest number of inliers as the target curve;

[0112] Step 314: Update the target curve based on the set of interior points corresponding to the target curve, and determine all exterior points in the first valid region subset according to the updated target curve.

[0113] In step 311, multiple points may be randomly selected from the first valid area subset to form a current sample subset, and curve fitting may be performed on the points in the current sample subset to obtain a current sample curve. Specifically, during the random selection, a certain number of points may be selected for each lane line number, and curve fitting may be performed on the points selected with the same number to obtain the current sample curve corresponding to each lane line number.

[0114] Furthermore, in step 312, a complement can be formed based on the points in the first valid area subset except the current sample subset, and then the distance between each point in the complement and the current sample curve can be determined, and the points with a distance less than a preset distance threshold are taken as inliers to obtain the current inlier set corresponding to the current sample curve, and the number of inliers in the current inlier set can be determined.

[0115] Furthermore, in step 313, the process can return to step 311 to re-extract the current sample subset and perform curve fitting to determine the current inlier set until the number of rounds returned to execution reaches the set number of rounds. At this time, the inlier set with the largest number of inliers can be selected from the inlier sets obtained in all rounds, and its corresponding sample curve can be used as the target curve.

[0116] Furthermore, in step 314, after determining the target curve, curve fitting can be performed again based on the set of inner points corresponding to the target curve, and the fitting result can be used as the new target curve to update the target curve, and then the distance between each point in the first valid area subset and the updated target curve is calculated, and the points with a distance greater than the set threshold are determined as outliers.

[0117] In the above steps 311 to 314, each sample curve is constructed and the corresponding inlier set is determined by random sampling and multiple rounds of fitting. This can increase the accuracy of the sample curve as the number of iterations increases, thereby ensuring the accuracy of the outliers. Moreover, after selecting the sample curve with the largest number of inliers, the curve is re-estimated based on the inlier set, which can further improve the accuracy of the curve and thus improve the accuracy of the outliers.

[0118] It should be noted that, in addition to determining the outliers based on the above method, all outliers can also be screened out from the first valid area subset through clustering, Euclidean distance and other methods, which is not limited in this embodiment of the present application.

[0119] After determining all external points in the first valid area subset, further, all external points can be removed from the first valid area subset, and curve fitting can be performed based on the remaining points in the first valid area subset to obtain a first fitting curve, and then discrete sampling is performed on the first fitting curve, and the points obtained by discrete sampling are placed in the optimized point set.

[0120] In addition, for the second effective area subset that is close to the vehicle, curve fitting can be performed on the points in the second effective area subset to obtain a second fitting curve, and then the distance between each point in the second effective area subset and the second fitting curve is calculated, and the points with a distance greater than the set distance threshold are removed from the second effective area subset, and the remaining points are placed in the optimized point set.

[0121] Through steps 31 and 32 above, for the first valid area subset far from the vehicle, points that differ significantly from the actual road can be filtered out by eliminating outliers and refitting the samples, achieving the goal of coarse optimization. For the second valid area subset close to the vehicle, all points that do not conform to the actual road can be filtered out based on the distance between the points and the fitted curve, achieving the goal of fine optimization. This ensures that points near the vehicle are more accurate than those farther away, ensuring that lane control points within a limited distance near the vehicle are closer to the actual road. This improves mapping efficiency while ensuring mapping accuracy, thereby ensuring the reliability of the memory driving function.

[0122] For example, Figure 6 This is an optimized flow chart provided by the embodiment of the present application, such as Figure 6As shown, based on the lane line point set entering the queue, the lane line point set is divided into a first valid area subset and a second valid area subset, wherein the external points in the first valid area subset are eliminated, and the curve fitting and sampling are re-performed to obtain the corresponding optimized point set, and the points in the second valid area subset whose distance to the fitted curve exceeds the threshold are eliminated to obtain the corresponding optimized point set. Finally, the optimized point set of the current frame is output.

[0123] In an embodiment of the present application, the above steps S110-S130 can be repeated for each frame, that is, the lane line information of the current frame is obtained and converted, and then divided into multiple valid area subsets for partition optimization.

[0124] Considering that different vehicles have different data collection frequencies, when high-frequency data collection is used, the time difference between two adjacent frames is extremely small, making the lane line information collected in the two adjacent frames very similar. Therefore, to further improve mapping efficiency, when high-frequency data collection is used, there is no need to optimize the lane line point set for each frame. Instead, optimization can be performed once every fixed frame to improve mapping efficiency while ensuring mapping accuracy.

[0125] In one example, before dividing the lane point set into a plurality of valid area subsets based on the distances between points in the lane point set and the vehicle, the method further includes:

[0126] Determine whether the vehicle's data collection frequency is greater than a preset frequency threshold; if so, determine whether the frame count value corresponding to the current frame reaches a preset optimized frame number. If so, divide the lane point set into multiple valid area subsets based on the distance between the points in the lane point set and the vehicle; otherwise, accumulate the frame count values;

[0127] After optimizing the points in the valid area subsets based on the optimization strategies corresponding to the valid area subsets, the method further includes: clearing the frame count value.

[0128] Specifically, after obtaining the lane line information of the current frame, it is possible to first determine whether the vehicle's data acquisition frequency is greater than a preset frequency threshold, where the data acquisition frequency can be the acquisition frequency of the vehicle's camera or sensor. If the data acquisition frequency is greater than the preset frequency threshold, it can be determined that the vehicle is in a high-frequency data acquisition state. At this time, before optimizing the lane line information of the current frame, it is possible to first determine whether the frame count value corresponding to the current frame reaches the preset optimization frame number.

[0129] The preset number of optimized frames can be determined based on the data collection frequency of the vehicle. The higher the data collection frequency, the larger the preset number of optimized frames. The frame count value corresponding to the current frame can be the number of frames that have not been optimized.

[0130] For example, if the frame count value corresponding to the current frame does not reach the preset optimized frame number, it means that the optimized frame number is not met at this time. The frame count value can be accumulated to continue to obtain the lane line information of the next frame and repeat the above process.

[0131] If the frame count value corresponding to the current frame reaches the preset number of optimized frames, it means that optimization can be performed at this time. The steps of dividing the lane line point set of the current frame into multiple valid area subsets and performing partition optimization are executed. Moreover, after the optimization process is completed, the frame count value is cleared to re-accumulate the number of frames that have not been optimized.

[0132] Through the above example, when the data acquisition frequency is high, considering that the acquired data between two adjacent frames are very similar, optimization can be performed once every certain number of frames to further improve the efficiency of mapping.

[0133] S140: Update the memory driving map point set based on the optimized point set of the current frame.

[0134] After obtaining the optimized point set of the current frame, the memory driving map point set can be updated using the optimized point set of the current frame. For example, the optimized point set of the current frame is written to the memory driving map point set to complete the map update. The memory driving map point set is used to describe the memory driving map. Writing the optimized point set of the current frame to the memory driving map point set can be writing the entire optimized point set of the current frame to the memory driving map point set, or writing a portion of the optimized point set of the current frame (the portion that differs from the memory driving map point set) to the memory driving map point set.

[0135] In a specific embodiment, after obtaining lane line information of the current frame, the method further includes: for each lane line in the lane line information, determining whether the lane line is associated with other lane lines in the previous frame; if so, determining the number of the associated other lane line as the number of the lane line; otherwise, assigning a new number to the lane line;

[0136] After updating the memory driving map point set based on the optimized point set of the current frame, the method further includes: determining points with the same number between different frames as points under the same lane line based on the numbers corresponding to the points in the memory driving map point set.

[0137] Specifically, after obtaining the lane lines of the current frame, they can be associated with the lane lines of the previous frame. For the associated lane lines, they inherit the numbering of the previous frame. For the unassociated lane lines, they can be assigned new numbers to achieve global numbering of lane lines, avoiding local numbering of each frame, which will cause confusion in subsequent numbering and affect map construction.

[0138] After updating the memory driving map point set based on the optimized point set of the current frame, points with the same number in the memory driving map point set can be determined as points under the same lane line, that is, points with the same number describe the same lane line.

[0139] The above implementation makes it easier to distinguish between the same lane line and different lane lines when updating the map, thereby ensuring the accuracy of mapping, avoiding the same lane line being identified as multiple different lane lines on the actual road, ensuring the accuracy of steps such as lane line curvature calculation, and thus improving the reliability of the memory driving function.

[0140] The lane mapping method for memory driving provided in an embodiment of the present application obtains lane line information for the current frame, determines a lane line point set based on the lane line information, and divides the lane line point set into multiple valid area subsets based on the distance between the points in the lane line point set and the vehicle. The points in the lane line point set are optimized using the optimization strategy corresponding to each valid area subset to obtain the optimized point set for the current frame. The memory driving map point set is then updated based on the optimized point set for the current frame. This method implements lane mapping for memory driving without relying on high-precision maps, solving problems such as high mapping costs, poor effectiveness, and high resource consumption. By extracting lane line point sets from lane line information, the method can discretize the road model into point sets. This method does not rely on a specific mathematical model, ensures data isolation, and avoids the impact of local detection errors on global accuracy. Furthermore, the method divides the point set into multiple subsets and optimizes different subsets separately, so that points in different areas are optimized differently. This method can implement a coarse-to-fine optimization process and improve mapping stability and accuracy.

[0141] Compared with other mapping algorithms, for example, 1. Fitting-based methods: This method models lane lines as curves and forms a map by splicing these curves. However, simply relying on modeled curves to fit lane lines and splicing them into the final map will make the final map not close to the real road surface; 2. By establishing a topological map based on the association relationship of lane lines, the lane lines are associated using a matching algorithm and a designed loss function to form the final map. However, the matching algorithm is prone to falling into a local optimal solution and is usually accompanied by a one-to-one matching, and the matching algorithm will change the matching result due to changes in convergence or the number of regressions; 3. Deep learning-based methods: By using deep learning algorithms, such as convolutional neural networks and recurrent neural networks, the features of lane lines are automatically learned and matched with map data. However, the end-to-end algorithm that uses neural networks and map data for matching mainly requires a large amount of training data and has a high investment cost. The mapping method provided in the embodiment of the present application can avoid the impact of local detection errors on global accuracy, ensure the accuracy and reliability of mapping, achieve the optimization goal from coarse to fine through partition optimization, and improve the stability of mapping.

[0142] Figure 7 7 is a schematic structural diagram of a lane mapping device for memory driving provided in an embodiment of the present application. The device includes a lane line acquisition module 710, a point set partitioning module 720, a partition optimization module 730, and an optimization output module 740, wherein:

[0143] A lane line acquisition module 710 is configured to acquire lane line information of a current frame and determine a lane line point set based on the lane line information;

[0144] a point set partitioning module 720 for partitioning the lane point set into a plurality of valid area subsets based on the distance between the points in the lane point set and the vehicle;

[0145] A partition optimization module 730 is configured to optimize the points in the valid area subset based on the optimization strategies corresponding to the valid area subsets to obtain an optimized point set for the current frame;

[0146] The optimization output module 740 is used to update the memory driving map point set based on the optimization point set of the current frame.

[0147] Based on the above embodiments, the lane line acquisition module 710 optionally includes a conversion unit, which is used to obtain the vehicle posture of the current frame; convert the vehicle posture from the local coordinate system corresponding to the current frame to the reference coordinate system to obtain the relative posture of the vehicle body; and convert the points in the lane line point set from the local coordinate system to the reference coordinate system based on the relative posture of the vehicle body.

[0148] On the basis of the above-mentioned embodiments, optionally, the lane line acquisition module 710 is further used to determine, for each lane line in the lane line information, whether the lane line is associated with other lane lines in the previous frame; if so, the numbers of the other associated lane lines are determined as the number of the lane line; otherwise, a new number is assigned to the lane line; the optimization output module 740 is further used to determine points with the same number between different frames as points under the same lane line based on the numbers corresponding to each point in the memory driving map point set.

[0149] On the basis of the above-mentioned embodiments, the lane line acquisition module 710 optionally includes a lane line reconstruction unit, which is used to determine, for each lane line in the lane line information, whether there is an intersection or branch between the lane line and other adjacent lane lines, and if so, determine the branch intersection point; at the branch intersection point, split each lane line into multiple line segments; fit the proximal line segment and the distal line segment of the lane where the vehicle is located in all line segments to obtain a new lane line, and use the remaining line segments as the new lane line.

[0150] Based on the above embodiments, optionally, the lane line reconstruction unit is further used to determine a first lateral distance between the starting point of the lane line and the starting point of other adjacent lane lines. If the first lateral distance is less than a preset first threshold, it is determined that a branch exists between the lane line and other adjacent lane lines; and to determine a second lateral distance between the end point of the lane line and the end point of other adjacent lane lines. If the second lateral distance is less than a preset second threshold, it is determined that an intersection exists between the lane line and other adjacent lane lines.

[0151] Based on the above embodiments, optionally, the point set partitioning module 720 is specifically configured to divide the lane line point set into a first valid area subset and a second valid area subset, wherein the distance between the points in the first valid area subset and the vehicle is greater than the distance between the points in the second valid area subset and the vehicle;

[0152] Partition optimization module 730 is specifically used to determine all external points in the first valid area subset for the first valid area subset, remove all external points from the first valid area subset, perform curve fitting on the first valid area subset after removal to obtain a first fitting curve, perform discrete sampling on the first fitting curve to obtain a corresponding optimized point set; for the second valid area subset, perform curve fitting on the second valid area subset to obtain a second fitting curve, and remove some points in the second valid area subset based on the distance between each point in the second valid area subset and the second fitting curve to obtain a corresponding optimized point set.

[0153] On the basis of the above-mentioned embodiments, optionally, the partition optimization module 730 is further used to randomly extract the first valid area subset to obtain the current sample subset, and perform curve fitting on the current sample subset to obtain the current sample curve; for the remainder of the first valid area subset except the current sample subset, determine the distance between each point in the remainder and the current sample curve, determine the current inlier set based on the points whose distance is less than a preset distance threshold, and determine the number of inliers in the current inlier set; return to execute the step of randomly extracting the first valid area subset until the number of rounds executed reaches the set number of rounds, and select the sample curve with the largest number of inliers as the target curve; update the target curve based on the inlier set corresponding to the target curve, and determine all the outliers in the first valid area subset according to the updated target curve.

[0154] Based on the above embodiments, the device may further include an optimization judgment unit, configured to judge whether the data acquisition frequency of the vehicle is greater than a preset frequency threshold; if so, to judge whether the frame count value corresponding to the current frame reaches a preset optimization frame number; if so, to divide the lane line point set into a plurality of valid area subsets according to the distance between the points in the lane line point set and the vehicle; otherwise, to accumulate the frame count values;

[0155] The partition optimization module 730 is further configured to clear the frame count value after optimizing the points in the valid area subset based on the optimization strategies corresponding to the valid area subsets.

[0156] The lane mapping device for driving memory provided in the embodiment of the present application can execute the steps of the lane mapping method for driving memory provided in the method embodiment of the present application, and the execution steps and beneficial effects are not repeated here.

[0157] Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 8 As shown, the electronic device 400 includes one or more processors 401 and a memory 402 .

[0158] The processor 401 may be a central processing unit (CPU) or other forms of processing units having data processing capabilities and / or instruction execution capabilities, and may control other components in the electronic device 400 to perform desired functions.

[0159] The memory 402 may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc. One or more computer program instructions may be stored on the computer-readable storage medium, and the processor 401 may execute the program instructions to implement the lane mapping method for memory driving and / or other desired functions of any embodiment of the present application described above. Various contents such as initial external parameters, thresholds, etc. may also be stored in the computer-readable storage medium.

[0160] In one example, electronic device 400 may further include an input device 403 and an output device 404, which are interconnected via a bus system and / or other connection mechanisms (not shown). Input device 403 may include, for example, a keyboard, a mouse, etc. Output device 404 may output various information to the outside, including warning information, braking force, etc. Output device 404 may include, for example, a display, a speaker, a printer, a communication network, and remote output devices connected thereto.

[0161] Of course, to simplify, Figure 8 Only some of the components related to the present application in the electronic device 400 are shown, and components such as buses, input / output interfaces, etc. are omitted. In addition, the electronic device 400 may further include any other appropriate components according to specific application scenarios.

[0162] In addition to the above-mentioned methods and devices, an embodiment of the present application may also be a computer program product, which includes computer program instructions, which, when executed by a processor, enable the processor to execute the steps of the lane mapping method for memorizing driving provided in any embodiment of the present application.

[0163] The computer program product may be written in any combination of one or more programming languages to implement the program code for performing the operations of the embodiments of the present application, including object-oriented programming languages such as Java, C++, and conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server.

[0164] In addition, an embodiment of the present application may also be a computer-readable storage medium having computer program instructions stored thereon. When the computer program instructions are executed by a processor, the processor executes the steps of the lane mapping method for memorizing driving provided in any embodiment of the present application.

[0165] The computer-readable storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, include but is not limited to a system, device or component of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0166] It should be noted that the terms used in this application are only for describing specific embodiments and are not intended to limit the scope of this application. As shown in the specification and claims of this application, unless the context clearly indicates an exception, the words "one", "an", "a kind of" and / or "the" do not specifically refer to the singular and may also include the plural. The terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method or device. In the absence of further restrictions, the elements defined by the sentence "comprise a..." do not exclude the presence of other identical elements in the process, method or device comprising the elements.

[0167] It should also be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on this application. Unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be a communication between the internal parts of two elements. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0168] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. The above is only the preferred implementation method of this application. It should be pointed out that due to the limitations of textual expression, there are objectively infinite specific structures. For ordinary technicians in this technical field, without departing from the principles of this application, they can also make several improvements, modifications or changes, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the scope of protection of this application.

Claims

1. A lane mapping method for memory driving, characterized in that: include: Obtaining lane line information of the current frame, and determining a lane line point set based on the lane line information; Dividing the lane point set into a plurality of valid area subsets according to the distance between the points in the lane point set and the vehicle; Based on the optimization strategies corresponding to the respective effective area subsets, the points in the effective area subsets are optimized to obtain the optimized point set of the current frame; The memory driving map point set is updated based on the optimized point set of the current frame.

2. The method according to claim 1, characterized in that After determining the lane line point set based on the lane line information, the method further includes: Obtaining the vehicle pose of the current frame; Converting the vehicle posture from the local coordinate system corresponding to the current frame to the reference coordinate system to obtain the relative posture of the vehicle body; Based on the relative position of the vehicle body, the points in the lane line point set are converted from the local coordinate system to a reference coordinate system.

3. The method according to claim 1, characterized in that After obtaining the lane line information of the current frame, it also includes: For each lane line in the lane line information, determine whether the lane line is associated with other lane lines in the previous frame; if so, determine the number of the associated other lane line as the number of the lane line; otherwise, assign a new number to the lane line; After updating the memory driving map point set based on the optimized point set of the current frame, the method further includes: Based on the numbers corresponding to the points in the memory driving map point set, points with the same numbers between different frames are determined to be points under the same lane line.

4. The method according to claim 1, wherein Before determining the lane line point set based on the lane line information, the method further includes: For each lane line in the lane line information, determine whether the lane line intersects or branches with other adjacent lane lines, and if so, determine a branch intersection point; At the branch intersection, each lane line is split into multiple line segments; The proximal and distal segments of the lane where the vehicle is located are fitted in all line segments to obtain a new lane line, and the remaining line segments are used as the new lane line.

5. The method according to claim 4, characterized in that Determining whether the lane line intersects or branches with other adjacent lane lines includes: determining a first lateral distance between a starting point of the lane line and a starting point of another adjacent lane line, and if the first lateral distance is less than a preset first threshold, determining that a branch exists between the lane line and the other adjacent lane line; and Determine a second lateral distance between the end point of the lane line and the end point of another adjacent lane line. If the second lateral distance is less than a preset second threshold, determine that there is an intersection between the lane line and the other adjacent lane line.

6. The method according to claim 1, characterized in that The lane line point set is divided into multiple valid area subsets, including: Dividing the lane line point set into a first valid area subset and a second valid area subset, wherein the distance between the points in the first valid area subset and the vehicle is greater than the distance between the points in the second valid area subset and the vehicle; Based on the optimization strategies corresponding to the respective effective area subsets, the points within the effective area subsets are optimized, including: For the first valid area subset, determine all outliers in the first valid area subset, remove all outliers from the first valid area subset, perform curve fitting on the first valid area subset after removal to obtain a first fitting curve, and perform discrete sampling on the first fitting curve to obtain a corresponding optimized point set; For the second valid area subset, curve fitting is performed on the second valid area subset to obtain a second fitting curve, and based on the distance between each point in the second valid area subset and the second fitting curve, some points in the second valid area subset are eliminated to obtain a corresponding optimized point set.

7. The method according to claim 6, characterized in that Determining all outliers within the first valid region subset includes: Randomly sampling the first valid area subset to obtain a current sample subset, and performing curve fitting on the current sample subset to obtain a current sample curve; For a remainder set other than the current sample subset within the first valid area subset, determining a distance between each point in the remainder set and the current sample curve, determining a current inlier set based on points whose distances are less than a preset distance threshold, and determining the number of inliers in the current inlier set; Return to the step of randomly selecting the first valid area subset until the number of rounds reaches the set number, and select the sample curve with the largest number of inliers as the target curve; The target curve is updated based on the set of interior points corresponding to the target curve, and all exterior points in the first valid area subset are determined according to the updated target curve.

8. The method according to claim 1, characterized in that Before dividing the lane point set into a plurality of valid area subsets based on the distances between points in the lane point set and the vehicle, the method further includes: Determine whether the vehicle's data collection frequency is greater than a preset frequency threshold; If so, determining whether the frame count value corresponding to the current frame reaches a preset optimized frame number; if so, executing the step of dividing the lane line point set into a plurality of valid area subsets based on the distance between the points in the lane line point set and the vehicle; otherwise, accumulating the frame count value; After optimizing the points in the valid area subsets based on the optimization strategies corresponding to the valid area subsets, the method further includes: The frame count value is cleared.

9. An electronic device, characterized in that: The electronic device comprises: processor and memory; The processor is configured to execute the steps of the lane mapping method for memory driving according to any one of claims 1 to 8 by calling the program or instructions stored in the memory.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program or instruction, and the program or instruction enables a computer to execute the steps of the lane mapping method for memory driving according to any one of claims 1 to 8.

Images