Road data processing method and device, storage medium and program product

By using sheep horn spiral fitting technology to process the coordinate and slope information of road three-dimensional data in intelligent vehicle driving, the problem of low map accuracy in polynomial fitting in long-distance scenarios is solved, and higher map drawing accuracy and adaptability are achieved.

CN118941728APending Publication Date: 2024-11-12NAVINFO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410985505.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-22
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the map accuracy obtained based on polynomial fitting is poor in long-distance scenarios such as ramps, bends, straight turns, etc., resulting in a decrease in map accuracy.

Method used

By extracting the three-dimensional road data in each frame of scene image, the coordinate information and slope information of each three-dimensional point in each road three-dimensional data are calculated, and the fitting process is performed using the sheep horn spiral line to generate the fitting line of the target road element. , to improve map drawing accuracy.

Benefits of technology

The map drawing accuracy is improved, especially in scenes with large curvature changes, the fitting line is smoother, more adaptable, and can describe road characteristics more accurately.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118941728A_ABST
    Figure CN118941728A_ABST
Patent Text Reader

Abstract

The embodiment of the invention provides a road data processing method and device, a storage medium and a program product. The method comprises the following steps: acquiring road three-dimensional data of each scene image in at least one scene image; carrying out sampling calculation processing on the road three-dimensional data of each scene image to obtain a sampling set; according to the coordinate information and the slope information of each control point, performing cavel spiral fitting processing on each control point, and generating a fitting line corresponding to the target road element; wherein the fitting line is used for generating a map. According to the method, the fitting line with higher precision can be obtained, and the map drawing precision is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of map technology, and in particular to a road data processing method, device, storage medium and program product. Background Art

[0002] High-precision maps provide rich road information during the vehicle's intelligent driving process. They can identify target road data based on images collected by on-board equipment for use in drawing high-precision maps.

[0003] In the prior art, all pixel points of the target road are obtained from the captured image, and all pixel points of the target road are fitted into a cubic polynomial curve, that is, a fitting line corresponding to the target road is obtained for generating a map.

[0004] However, in the above process, in long-distance scenarios such as ramps, curves, and straight turns, the accuracy of the fitting line obtained based on polynomial fitting is poor, resulting in a decrease in the accuracy of the drawn map. Summary of the invention

[0005] The embodiments of the present application provide a road data processing method, device, storage medium and program product to solve the technical problem of poor map accuracy obtained based on polynomial fitting.

[0006] In a first aspect, an embodiment of the present application provides a method for processing road data, the method comprising:

[0007] Acquire three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which the vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image;

[0008] Sampling and calculating the road three-dimensional data of each of the scene images to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents a three-dimensional point of a target road element; the control point has coordinate information and slope information of the three-dimensional point in a preset global coordinate system;

[0009] According to the coordinate information and slope information of each control point, a clevis fitting process is performed on each control point to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0010] In a possible implementation manner, the scene images have a collection time; and sampling and calculating the road three-dimensional data of each of the scene images to obtain a sampling set includes:

[0011] The three-dimensional road data are grouped and processed to obtain a head-end data set, an intermediate data set, and a tail-end data set; wherein the acquisition time of the scene image to which the three-dimensional road data in the head-end data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs; and the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the tail-end data set belongs;

[0012] The sampling set is determined according to the head-end data set, the middle data set and the tail-end data set.

[0013] In a possible implementation, determining the sampling set according to the head-end data set, the intermediate data set, and the tail-end data set includes:

[0014] Performing fitting processing on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set; wherein the first control point and the last control point both have coordinate information;

[0015] Perform multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain each first intermediate point in the sampling set; wherein the first intermediate point is the control point remaining except the first control point and the last control point in the sampling set; the first intermediate point has the coordinate information of the control point in the preset global coordinate system;

[0016] The slope information of the control point in the sampling set is determined according to the coordinate information of the control point in the sampling set.

[0017] In a possible implementation manner, fitting processing is performed on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set, including:

[0018] Performing polynomial fitting processing on the head-end data set and the tail-end data set to obtain fitting data; wherein the fitting data includes at least two fitting points; the fitting points represent three-dimensional points of target road elements; the fitting points have coordinate information of the three-dimensional points in the preset global coordinate system;

[0019] The first fitting point in the fitting data is determined to be the first control point in the sampling set, and the last fitting point in the fitting data is determined to be the last control point in the sampling set.

[0020] In a possible implementation, performing multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain the first intermediate point in the sampling set includes:

[0021] Extracting at least one target data set from the intermediate data set according to a preset interval time; wherein the target data set includes three-dimensional road data of a preset number of scene images; and the acquisition moments of the scene images corresponding to the target data set are continuous;

[0022] Determine a first coordinate set corresponding to the target data set; wherein the first coordinate set includes first coordinate values ​​corresponding to each road three-dimensional data in the target data set; the first coordinate value is a coordinate value of each three-dimensional point in the road three-dimensional data located on a right coordinate axis in a preset three-dimensional coordinate system; the right coordinate axis points to the right side of the vehicle's forward direction;

[0023] determining a difference between the first coordinate value and an average value corresponding to each of the first coordinate values;

[0024] The three-dimensional point to which the first coordinate value with the smallest difference in the first coordinate set belongs is determined as the first intermediate point in the sampling set; wherein the first intermediate point has coordinate information of the three-dimensional point in the global coordinate system.

[0025] In a possible implementation manner, the control points in the sampling set are sorted from small to large according to the forward distance, where the forward distance represents the horizontal distance between the position of the control point and the position of the vehicle;

[0026] Determining the slope information of the control point in the sampling set according to the coordinate information of the control point in the sampling set includes:

[0027] Determine the first endpoint, the last endpoint, and each second intermediate point in the sampling set; wherein the first endpoint is the first control point in the sampling set, the last endpoint is the last control point in the sampling set, and the second intermediate point is the remaining control point in the sampling set except the first endpoint and the last endpoint;

[0028] Performing polynomial derivation processing on the first endpoint and the first end point to obtain slope information of the first endpoint and slope information of the first end point;

[0029] An interpolation calculation process is performed on the second middle point to obtain slope information of the second middle point.

[0030] In a possible implementation manner, performing interpolation calculation processing on the second intermediate point to obtain the slope information of the second intermediate point includes:

[0031] Determine a front neighboring point of the second midpoint and a rear neighboring point of the second midpoint from the sampling set; wherein the front neighboring point is a control point that is arranged before the second midpoint in the sampling set and is adjacent to the second midpoint; and the rear neighboring point is a control point that is arranged after the second midpoint in the sampling set and is adjacent to the second midpoint;

[0032] The slope information of the second intermediate point is determined according to the second intermediate point, a previous adjacent point of the second intermediate point, and a subsequent adjacent point of the second intermediate point.

[0033] In a possible implementation manner, determining the slope information of the second intermediate point according to the second intermediate point, a previous adjacent point of the second intermediate point, and a subsequent adjacent point of the second intermediate point includes:

[0034] Determine the corresponding angle information between the second middle point, the front adjacent point, and the rear adjacent point; wherein the angle information represents the angle between the first vector and the second vector of the second middle point in the preset coordinate system; the first vector is a vector formed by the second middle point and the front adjacent point of the second middle point, and the second vector is a vector formed by the second middle point and the rear adjacent point of the second middle point;

[0035] If it is determined that the angle represented by the angle information is less than or equal to the preset angle, determining the slope information of the second midpoint according to the second vector of the second midpoint;

[0036] If it is determined that the angle represented by the angle information is greater than the preset angle, a fitting calculation process is performed on the second middle point, the front adjacent point, and the rear adjacent point to obtain the slope information of the second middle point.

[0037] In a possible implementation manner, performing fitting calculation processing on the second middle point, the front adjacent point, and the rear adjacent point to obtain slope information of the second middle point includes:

[0038] Fitting the coordinate information of the second middle point, the coordinate information of the front adjacent point, and the coordinate information of the rear adjacent point to obtain circle information of the second middle point; wherein the circle information is information of a fitted circle corresponding to the second middle point, the front adjacent point of the second middle point, and the rear adjacent point of the second middle point;

[0039] The slope information of the second middle point is determined according to the coordinate information of the second middle point and the circle information.

[0040] In a possible implementation manner, the method further includes:

[0041] Determine at least one second coordinate set corresponding to the three-dimensional road data of each of the scene images; wherein the second coordinate set includes at least one second coordinate value; the second coordinate value is the coordinate value of each three-dimensional point in the three-dimensional road data located on the right coordinate axis in a preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction;

[0042] Calculating and processing each second coordinate value in the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set;

[0043] According to the coordinate mean corresponding to the second coordinate set, the three-dimensional data of each road corresponding to the second coordinate set are filtered out to obtain a processed second coordinate set; wherein the processed second coordinate set includes three-dimensional points whose second coordinate values ​​in the three-dimensional data of each road are in a reserved interval corresponding to the second coordinate set; the reserved interval is a preset numerical interval in which the coordinate mean corresponding to the second coordinate set is located.

[0044] In a possible implementation, calculating and processing each second coordinate value in the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set includes:

[0045] Determine at least one preset interval corresponding to the second coordinate set, and determine a first number of points in the preset interval; wherein the preset interval is a numerical interval in which a preset second coordinate value is located, and the first number of points represents the total number of three-dimensional points in the second coordinate set that belong to the second coordinate value in the preset interval;

[0046] Determine a target interval of a preset number corresponding to the second coordinate set, and determine a second number of points corresponding to the second coordinate set; wherein the target interval is a preset interval within a preset number range of the largest number of first points among all preset intervals; and the second number of points is the total number of all three-dimensional points corresponding to the second coordinate set;

[0047] For each target interval corresponding to the second coordinate set, determine a weight of the target interval according to the first number of points in the target interval and the second number of points corresponding to the second coordinate set; wherein the weight represents the importance of the target interval in all preset intervals;

[0048] According to the weights of each target interval corresponding to the second coordinate set, the second coordinate values ​​in each target interval corresponding to the second coordinate set are weighted averaged to obtain the coordinate mean corresponding to the second coordinate set.

[0049] In a second aspect, an embodiment of the present application provides a road data processing device, the device comprising:

[0050] An acquisition module, used to acquire three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which the vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image;

[0051] A sampling module, used for performing sampling calculation processing on the road three-dimensional data of each of the scene images to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents a three-dimensional point of a target road element; the control point has coordinate information and slope information of the three-dimensional point in a preset global coordinate system;

[0052] The fitting module is used to perform a clevis fitting process on each of the control points according to the coordinate information and slope information of each of the control points to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0053] In a possible implementation manner, the scene image has a collection time; and the sampling module is specifically used for:

[0054] The three-dimensional road data are grouped and processed to obtain a head-end data set, an intermediate data set, and a tail-end data set; wherein the acquisition time of the scene image to which the three-dimensional road data in the head-end data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs; and the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the tail-end data set belongs;

[0055] The sampling set is determined according to the head-end data set, the middle data set and the tail-end data set.

[0056] In a possible implementation manner, the sampling module is specifically used for:

[0057] Performing fitting processing on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set; wherein the first control point and the last control point both have coordinate information;

[0058] Perform multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain each first intermediate point in the sampling set; wherein the first intermediate point is the control point remaining except the first control point and the last control point in the sampling set; the first intermediate point has the coordinate information of the control point in the preset global coordinate system;

[0059] The slope information of the control point in the sampling set is determined according to the coordinate information of the control point in the sampling set.

[0060] In a possible implementation manner, the sampling module is specifically used for:

[0061] Performing polynomial fitting processing on the head-end data set and the tail-end data set to obtain fitting data; wherein the fitting data includes at least two fitting points; the fitting points represent three-dimensional points of target road elements; the fitting points have coordinate information of the three-dimensional points in the preset global coordinate system;

[0062] The first fitting point in the fitting data is determined to be the first control point in the sampling set, and the last fitting point in the fitting data is determined to be the last control point in the sampling set.

[0063] In a possible implementation manner, the sampling module is specifically used for:

[0064] Extracting at least one target data set from the intermediate data set according to a preset interval time; wherein the target data set includes three-dimensional road data of a preset number of scene images; and the acquisition moments of the scene images corresponding to the target data set are continuous;

[0065] Determine a first coordinate set corresponding to the target data set; wherein the first coordinate set includes first coordinate values ​​corresponding to each road three-dimensional data in the target data set; the first coordinate value is a coordinate value of each three-dimensional point in the road three-dimensional data located on a right coordinate axis in a preset three-dimensional coordinate system; the right coordinate axis points to the right side of the vehicle's forward direction;

[0066] determining a difference between the first coordinate value and an average value corresponding to each of the first coordinate values;

[0067] The three-dimensional point to which the first coordinate value with the smallest difference in the first coordinate set belongs is determined as the first intermediate point in the sampling set; wherein the first intermediate point has coordinate information of the three-dimensional point in the global coordinate system.

[0068] In a possible implementation manner, the control points in the sampling set are sorted in ascending order according to the forward distance, where the forward distance represents the horizontal distance between the position of the control point and the position of the vehicle; and the sampling module is specifically configured to:

[0069] Determine the first endpoint, the last endpoint, and each second intermediate point in the sampling set; wherein the first endpoint is the first control point in the sampling set, the last endpoint is the last control point in the sampling set, and the second intermediate point is the remaining control point in the sampling set except the first endpoint and the last endpoint;

[0070] Performing polynomial derivation processing on the first endpoint and the first end point to obtain slope information of the first endpoint and slope information of the first end point;

[0071] An interpolation calculation process is performed on the second middle point to obtain slope information of the second middle point.

[0072] In a possible implementation manner, the sampling module is further specifically used for:

[0073] Determine a front neighboring point of the second midpoint and a rear neighboring point of the second midpoint from the sampling set; wherein the front neighboring point is a control point that is arranged before the second midpoint in the sampling set and is adjacent to the second midpoint; and the rear neighboring point is a control point that is arranged after the second midpoint in the sampling set and is adjacent to the second midpoint;

[0074] The slope information of the second intermediate point is determined according to the second intermediate point, a previous adjacent point of the second intermediate point, and a subsequent adjacent point of the second intermediate point.

[0075] In a possible implementation manner, the sampling module is further specifically used for:

[0076] Determine the corresponding angle information between the second middle point, the front adjacent point, and the rear adjacent point; wherein the angle information represents the angle between the first vector and the second vector of the second middle point in the preset coordinate system; the first vector is a vector formed by the second middle point and the front adjacent point of the second middle point, and the second vector is a vector formed by the second middle point and the rear adjacent point of the second middle point;

[0077] If it is determined that the angle represented by the angle information is less than or equal to the preset angle, determining the slope information of the second midpoint according to the second vector of the second midpoint;

[0078] If it is determined that the angle represented by the angle information is greater than the preset angle, a fitting calculation process is performed on the second middle point, the front adjacent point, and the rear adjacent point to obtain the slope information of the second middle point.

[0079] In a possible implementation manner, the sampling module is further specifically used for:

[0080] Fitting the coordinate information of the second middle point, the coordinate information of the front adjacent point, and the coordinate information of the rear adjacent point to obtain circle information of the second middle point; wherein the circle information is information of a fitted circle corresponding to the second middle point, the front adjacent point of the second middle point, and the rear adjacent point of the second middle point;

[0081] The slope information of the second middle point is determined according to the coordinate information of the second middle point and the circle information.

[0082] In a possible implementation manner, the device is further used for:

[0083] Determine at least one second coordinate set corresponding to the three-dimensional road data of each of the scene images; wherein the second coordinate set includes at least one second coordinate value; the second coordinate value is the coordinate value of each three-dimensional point in the three-dimensional road data located on the right coordinate axis in a preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction;

[0084] Calculating and processing each second coordinate value in the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set;

[0085] According to the coordinate mean corresponding to the second coordinate set, the three-dimensional data of each road corresponding to the second coordinate set are filtered out to obtain a processed second coordinate set; wherein the processed second coordinate set includes three-dimensional points whose second coordinate values ​​in the three-dimensional data of each road are in a reserved interval corresponding to the second coordinate set; the reserved interval is a preset numerical interval in which the coordinate mean corresponding to the second coordinate set is located.

[0086] In a possible implementation manner, the device is further specifically used for:

[0087] Determine at least one preset interval corresponding to the second coordinate set, and determine a first number of points in the preset interval; wherein the preset interval is a numerical interval in which a preset second coordinate value is located, and the first number of points represents the total number of three-dimensional points in the second coordinate set that belong to the second coordinate value in the preset interval;

[0088] Determine a target interval of a preset number corresponding to the second coordinate set, and determine a second number of points corresponding to the second coordinate set; wherein the target interval is a preset interval within a preset number range of the largest number of first points among all preset intervals; and the second number of points is the total number of all three-dimensional points corresponding to the second coordinate set;

[0089] For each target interval corresponding to the second coordinate set, determine a weight of the target interval according to the first number of points in the target interval and the second number of points corresponding to the second coordinate set; wherein the weight represents the importance of the target interval in all preset intervals;

[0090] According to the weights of the target intervals corresponding to the second coordinate set, weighted averaging processing is performed on the second coordinate values ​​in the target intervals corresponding to the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set.

[0091] In a third aspect, an embodiment of the present application provides a computing device, including: a memory, a processor;

[0092] The memory stores computer-executable instructions;

[0093] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0094] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementations of the first aspect.

[0095] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0096] The road data processing method, device, storage medium and program product provided in the embodiments of the present application extract the three-dimensional road data in each frame of the scene image, calculate the coordinate information and slope information of each three-dimensional point in each three-dimensional road data, and use the ram's horn spiral to implement the fitting processing of the road data, so as to obtain the fitting line corresponding to the target road element for generating a map; further, by using the ram's horn spiral fitting to implement the fitting processing of the road three-dimensional data, a more accurate fitting line is obtained, which can improve the accuracy of map drawing. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0098] Figure 1 A schematic diagram of an application scenario provided for this application;

[0099] Figure 2 A schematic diagram of a process flow of a road data processing method provided in an embodiment of the present application;

[0100] Figure 3 A schematic diagram of a road data fitting error provided in an embodiment of the present application;

[0101] Figure 4 A road data fitting comparison effect diagram for a lane line straight road scenario provided in an embodiment of the present application;

[0102] Figure 5A road data fitting comparison effect diagram for a paved edge curve scenario provided in an embodiment of the present application;

[0103] Figure 6 A flowchart of another method for processing road data provided in an embodiment of the present application;

[0104] Figure 7 The effect diagram of the three-dimensional road data of a guardrail provided in an embodiment of the present application before and after denoising;

[0105] Figure 8 A flowchart of another method for processing road data provided in an embodiment of the present application;

[0106] Fig. 9 The effect diagram of the 3D data of a paved road before and after denoising provided in an embodiment of the present application;

[0107] Fig.10 A schematic diagram of the structure of a road data processing device provided in an embodiment of the present application;

[0108] Fig.11 A schematic diagram of the structure of a computing device provided in an embodiment of the present application.

[0109] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0110] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application.

[0111] During the intelligent driving process of the vehicle, it is necessary to identify the target road data based on the images collected by the on-board equipment, such as lane lines, guardrails, paved edges, curbs and other road linear element point cloud data, for use in drawing maps and vehicle intelligent driving.

[0112] In one example, all pixel points of the target road are obtained from the captured image, and all pixel points of the target road are fitted into a cubic polynomial curve, that is, a fitting line corresponding to the target road is obtained.

[0113] However, in the above process, for scenes with curvature changes and long distances, such as ramps, curves, straight turns, etc., the accuracy of the fitting line obtained based on polynomial fitting is poor, resulting in a decrease in the accuracy of the drawn map.

[0114] The road data processing method provided in the present application extracts the three-dimensional road data in each frame of the scene image, calculates the coordinate information and slope information of each three-dimensional point in each three-dimensional road data, and uses the ram's horn spiral to realize the fitting processing of the road data, thereby obtaining the fitting line corresponding to the target road element for use in generating a map, thereby solving the technical problem that the fitting line obtained based on polynomial fitting has poor accuracy, resulting in reduced accuracy of the drawn map.

[0115] Figure 1 A schematic diagram of an application scenario provided for this application, such as Figure 1 As shown, the specific application scenario of the present application includes a computing device 101 and an image acquisition device 102. The image acquisition device 102 acquires scene images of driving scenes in which multiple vehicles are located. The computing device 101 obtains these scene images from the image acquisition device 102, performs image processing on these scene images, and obtains three-dimensional road data of each scene image. Then, these three-dimensional road data are sampled and calculated, and subjected to ram's horn spiral fitting processing to obtain fitting lines corresponding to target road elements in the scene image for use in generating maps.

[0116] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0117] Figure 2 A flowchart of a method for processing road data provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the method includes:

[0118] 201. Obtain three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which a vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image.

[0119] Exemplarily, the execution subject of this embodiment may be a computing device, hereinafter referred to as the device, which may be a physical device or a virtual device for processing road data. The image acquisition device acquires a scene image representing the driving scene in which the vehicle is located, and the device obtains one or more scene images from the image acquisition device, and performs image processing, such as target detection and three-dimensional reconstruction processing, on each acquired scene image to obtain three-dimensional road data for each scene image, such as point cloud data of target road elements such as lane lines, guardrails, paved edges, and curbs, that is, three-dimensional road data. The three-dimensional road data of each target road element includes the three-dimensional points of the target road element in the corresponding scene image; taking the lane line as an example, for each frame of image acquired by the camera, the lane line on the image is first detected by the lane line detection algorithm, and then the lane line point cloud of the lane line in the frame is generated by the three-dimensional reconstruction method.

[0120] 202. Sampling and calculating the three-dimensional road data of each scene image to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents a three-dimensional point of a target road element; the control point has coordinate information and slope information of the corresponding three-dimensional point in a preset global coordinate system.

[0121] Exemplarily, through preset sampling rules, the device performs sampling and calculation processing on the road three-dimensional data of each scene image collected to obtain a sampling set corresponding to all road three-dimensional data; wherein the sampling set includes one or more control points, wherein each control point represents a three-dimensional point corresponding to the target road element, that is, each control point belongs to any three-dimensional point sampled from the multiple three-dimensional points acquired; each control point has coordinate information of the three-dimensional point in a preset global coordinate system, such as coordinate information in a Gaussian coordinate system (G system), and by calculating the coordinate information of each control point, the slope information corresponding to each control point in the preset global coordinate system is obtained.

[0122] 203. Perform a clevis fitting process on each control point according to the coordinate information and slope information of each control point to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0123] Exemplarily, the device calculates the coordinate information and slope information of each control point based on the ram's horn fitting algorithm according to the coordinate information and slope information of each control point, and then performs ram's horn fitting processing on each control point to obtain a fitting line, that is, a fitting line corresponding to the target road element corresponding to each control point, wherein the fitting line can be used to generate a map, such as a high-precision map, for use in the fields of intelligent driving of vehicles. For example, the error between the cubic polynomial fitting and the true value of the highly automated driving map HAD (High Automated Driving, abbreviated) is calculated, and the error between the ram's horn fitting and the true value of HAD is calculated, and the two errors are compared, Figure 3 A schematic diagram of a road data fitting error provided in an embodiment of the present application is shown in FIG. Figure 3 As shown, it can be seen that the fitting line accuracy after using the horn spiral fitting can be improved by about 0.2m compared with the fitting after using the cubic polynomial fitting; Figure 4 A road data fitting comparison effect diagram for a lane line straight road scenario provided in an embodiment of the present application, such as Figure 4 As shown in the figure, it can be seen that the lane lines fitted by the cubic polynomial are on both sides of the guardrail, which is obviously inconsistent with the actual scene, while the lane lines fitted by the ram's horn spiral are on the right side of the guardrail, close to the trajectory point, which is obviously more consistent with the actual scene; Figure 5 A road data fitting comparison effect diagram for a paved edge curve scenario provided in an embodiment of the present application, such as Figure 5 As shown, it can be seen that the cubic polynomial fitting has breaks, while the ram's horn spiral fitting has no breaks, and the fitting is smoother, and the transition from a straight line to a curve is smooth (the curvature change will be relatively large). The ram's horn spiral fitting algorithm can adapt well to such scenarios. Furthermore, for vehicles that perform assisted driving based on the lane lines obtained by the ram's horn spiral fitting algorithm, the vehicle does not need to decelerate sharply when entering a curve from a straight line, thereby improving the driving experience of the vehicle users.

[0124] In this embodiment, a method for processing road data is provided, by extracting three-dimensional road data in each frame of scene image, calculating the coordinate information and slope information of each three-dimensional point in each three-dimensional road data, and using the sheep's horn spiral to implement the fitting processing of the road data, and obtain the fitting line corresponding to the target road element for generating a map; further, by using the sheep's horn spiral fitting to implement the fitting processing of the road three-dimensional data, a more accurate fitting line is obtained, which can improve the accuracy of map drawing.

[0125] Figure 6 A flowchart of another method for processing road data provided in an embodiment of the present application is shown as follows: Figure 6 As shown, the method includes:

[0126] 301. Obtain three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which a vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image.

[0127] Exemplarily, this step may refer to step 201 and will not be described in detail here.

[0128] 302. Determine at least one second coordinate set corresponding to the three-dimensional road data of each scene image; wherein the second coordinate set includes at least one second coordinate value; the second coordinate value is the coordinate value of each three-dimensional point in the three-dimensional road data located on the right coordinate axis in a preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction.

[0129] For example, due to detection errors, reconstruction errors, etc., the road 3D data of the generated single-frame scene image contains noise, which will affect the fitting accuracy. Therefore, it is necessary to use a multi-frame data statistics method to implement noise filtering processing on the road 3D data of each scene image. The device processes the road 3D data of all scene images multiple times, such as taking multiple frames for processing each time, and extracts the coordinate value of each 3D point on the right coordinate axis in the preset 3D coordinate system for each 3D point in each road 3D data taken each time, that is, the second coordinate value. The preset 3D coordinate system is the vehicle coordinate system, which refers to the 3D coordinate system established with the vehicle itself as the reference point; the direction of the right coordinate axis in the coordinate system is the lateral right direction of the vehicle's forward direction; and then, the second coordinate set corresponding to each road 3D data taken each time is obtained; wherein the forward coordinate axis in the preset 3D coordinate system is the coordinate axis whose direction is consistent with the vehicle's forward direction in the preset 3D coordinate system.

[0130] For example, in the vehicle coordinate system (B system), the XYZ axes correspond to the front, right and downward directions of the vehicle respectively, that is, the x-axis in the vehicle coordinate system points to the vehicle's forward direction, and the y-axis points to the lateral right direction of the vehicle's forward direction; the z-axis points to the ground perpendicular to the ground and points downward. Each time, 8 frames of road 3D data of the scene image are taken, and the right-hand coordinates of each 3D point in the 8 frames of road 3D data of the scene image are sorted from small to large in the 3D coordinates of each 3D point in the 8 frames of road data of the scene image in the vehicle coordinate system (B system), and recorded as the second coordinate set {Xi, i=1, 2, 3...N} corresponding to the 8 frames of road 3D data of the scene image, where N represents the number of points.

[0131] 303. Calculate and process each second coordinate value in the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set.

[0132] Exemplarily, for each second coordinate set, the device calculates and processes each second coordinate value in each second coordinate set based on a preset mean calculation algorithm to obtain a coordinate mean corresponding to the second coordinate set.

[0133] In one example, step 303 includes the following steps:

[0134] The first step of step 303 is to determine at least one preset interval corresponding to the second coordinate set, and determine the first number of points in the preset interval; wherein the preset interval is a numerical interval in which the preset second coordinate value is located, and the first number of points represents the total number of three-dimensional points belonging to the second coordinate value in the second coordinate set that is within the preset interval.

[0135] The second step of step 303 is to determine the target interval of the preset number corresponding to the second coordinate set, and determine the second number of points corresponding to the second coordinate set; wherein the target interval is a preset interval within the preset number range of the largest number of first points in all preset intervals; and the second number of points is the total number of all three-dimensional points corresponding to the second coordinate set.

[0136] The third step of step 303 is to determine the weight of the target interval for each target interval corresponding to the second coordinate set according to the first number of points in the target interval and the second number of points corresponding to the second coordinate set; wherein the weight represents the importance of the target interval in all preset intervals.

[0137] The fourth step of step 303 is to perform weighted averaging processing on the second coordinate values ​​in each target interval corresponding to the second coordinate set according to the weights of each target interval corresponding to the second coordinate set, so as to obtain the coordinate mean corresponding to the second coordinate set.

[0138] Exemplarily, when the device calculates the coordinate mean corresponding to each second coordinate set, for each second coordinate set, according to the numerical range of each second coordinate value in the second coordinate set, it can be pre-divided into multiple preset intervals, and the total number of three-dimensional points belonging to the second coordinate values ​​in the second coordinate set in each preset interval is determined, which is the number of first points in each preset interval in the second coordinate set; for each second coordinate set, the preset interval within the preset number range with the largest number of first points in all preset intervals is determined as the preset number target interval corresponding to the second coordinate set, and the total number of all three-dimensional points corresponding to all second coordinate values ​​in the second coordinate set is determined, which is the number of first points for the second coordinate set. the device calculates the first point number of each target interval and the second point number corresponding to the second coordinate set for each target interval corresponding to each second coordinate set based on a preset calculation algorithm, and obtains the weight of each target interval in the second coordinate set to characterize the importance of the target interval in all preset intervals; the device performs weighted averaging processing on the second coordinate values ​​in each target interval corresponding to each second coordinate set based on a preset weighted averaging algorithm and according to the weights of each target interval corresponding to each second coordinate set, and obtains the coordinate mean corresponding to each second coordinate set; thereby, a more accurate coordinate mean can be obtained, so as to improve the accuracy of the results obtained in subsequent calculation processes.

[0139] For example, in the interval X0-XN, it is divided into several preset intervals at intervals of 1m. The number of points falling in each preset interval is counted. Assuming Hj represents the jth interval, the number of points falling in this interval is Nj, then the weight of this interval is wj=Nj / N. All intervals are sorted from large to small according to the weight, and the first two intervals H0 and H1 are taken; the right-hand coordinates of all points in the B system falling in the intervals H0 and H1 are weighted averaged, and the calculation formula is as follows: Calculate the coordinate mean X mean ; Among them, X0 and X1 represent the point sets falling on H0 and H1 respectively, w0 and w1 represent the weights of H0 and H1 respectively, and N0 and N1 represent the number of points falling on the intervals of H0 and H1 respectively.

[0140] 304. According to the coordinate mean corresponding to the second coordinate set, the three-dimensional data of each road corresponding to the second coordinate set are filtered out to obtain a processed second coordinate set; wherein the processed second coordinate set includes three-dimensional points whose second coordinate values ​​in the three-dimensional data of each road are in a reserved interval corresponding to the second coordinate set; the reserved interval is a preset numerical interval in which the coordinate mean corresponding to the second coordinate set is located.

[0141] Exemplarily, the device determines the preset value interval of the coordinate mean corresponding to each second coordinate set as the reserved interval corresponding to each second coordinate set according to the obtained coordinate mean corresponding to each second coordinate set, determines the three-dimensional points whose second coordinate values ​​in each road three-dimensional data are not in the reserved interval corresponding to the second coordinate set as noise points, and performs filtering processing to retain the three-dimensional points whose second coordinate values ​​in each road three-dimensional data are in the reserved interval corresponding to the second coordinate set, that is, to obtain each processed second coordinate set; for example, Figure 7 The effect diagram of the 3D road data of a guardrail provided in the embodiment of the present application before and after denoising is shown as follows: Figure 7 As shown, the absolute value abs(Xi-Xmean) of the difference between the coordinate mean Xmean and each second coordinate value Xi can be calculated. If abs(Xi-Xmean)>0.6, the point corresponding to Xi is noise and needs to be filtered out, that is, the point with the right coordinate within the interval of ±0.6m of the average value is retained; then, by using the method of multi-frame data statistics to filter the noise of the road three-dimensional data, the denoised road three-dimensional data is obtained for fitting processing, which can reduce detection errors and reconstruction errors and further improve the fitting accuracy.

[0142] 305. Sampling and calculating the three-dimensional road data of each scene image to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents the three-dimensional point of the target road element; the control point has coordinate information and slope information of the three-dimensional point in a preset global coordinate system.

[0143] Exemplarily, this step may refer to step 202 and will not be described in detail here.

[0144] In one example, the scene image has a collection time; step 305 includes:

[0145] Step 1: Group the three-dimensional data of each road to obtain a head-end data set, an intermediate data set and a tail-end data set; wherein the acquisition time of the scene image to which the road three-dimensional data in the head-end data set belongs is earlier than the acquisition time of the scene image to which the road three-dimensional data in the intermediate data set belongs; the acquisition time of the scene image to which the road three-dimensional data in the intermediate data set belongs is earlier than the acquisition time of the scene image to which the road three-dimensional data in the tail-end data set belongs.

[0146] Step 2: Determine the sampling set based on the head-end data set, the middle data set and the tail-end data set.

[0147] Specifically, each scene image acquired by the device has a collection time. According to the collection time of each scene image acquired, the road three-dimensional data of each scene image are sorted and grouped in time order to obtain a head-end data set, an intermediate data set and a tail-end data set, wherein the collection time of the scene image to which the road three-dimensional data in the head-end data set belongs is earlier than the collection time of the scene image to which the road three-dimensional data in the intermediate data set belongs, and the collection time of the scene image to which the road three-dimensional data in the intermediate data set belongs is earlier than the collection time of the scene image to which the road three-dimensional data in the tail-end data set belongs; for example, for the road three-dimensional data of all scene images, the head-end data set contains the road three-dimensional data of the 10 frames of scene images with the earliest collection time, the head-end data set contains the road three-dimensional data of the 10 frames of scene images with the earliest collection time, the tail-end data set contains the road three-dimensional data of the 10 frames of scene images with the latest collection time, and the remaining road three-dimensional data of each scene image constitute the intermediate data set. For the road three-dimensional data in the obtained head-end data set, intermediate data set and tail-end data set, sampling and calculation processing are performed on each three-dimensional point in these road three-dimensional data based on a preset sampling calculation and processing algorithm, that is, each control point in the sampling set is obtained; for example, the road three-dimensional data in the head-end data set is first sampled and calculated to obtain a part of the control points in the sampling set, and the road three-dimensional data in the intermediate data set and the tail-end data set are sampled and calculated to obtain the control points of the remaining part of the sampling set; then the control points in the sampling set are fitted and calculated to obtain the coordinate information and slope information of each control point in the sampling set in the preset global coordinate system.

[0148] 306. Perform a clevis fitting process on each control point according to the coordinate information and slope information of each control point to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0149] Exemplarily, this step may refer to step 203 and will not be described in detail here.

[0150] In this embodiment, based on the above embodiment, a multi-frame data statistics method is used to filter the noise of the road three-dimensional data, and the denoised road three-dimensional data is obtained for fitting processing, which can reduce detection errors and reconstruction errors and further improve the fitting accuracy.

[0151] Figure 8 A flowchart of another method for processing road data provided in an embodiment of the present application is shown as follows: Figure 8 As shown, the method includes:

[0152] 401. Obtain three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which a vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image.

[0153] Exemplarily, this step may refer to step 201 and will not be described in detail here.

[0154] 402. Group the three-dimensional data of each road to obtain a head-end data set, an intermediate data set, and a tail-end data set; wherein the acquisition time of the scene image to which the three-dimensional data of the road in the head-end data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional data of the road in the intermediate data set belongs; and the acquisition time of the scene image to which the three-dimensional data of the road in the intermediate data set is earlier than the acquisition time of the scene image to which the three-dimensional data of the road in the tail-end data set belongs.

[0155] Among them, the scene image has a collection time.

[0156] Exemplarily, each scene image acquired by the device has an acquisition time. According to the acquisition time of each scene image acquired, the road three-dimensional data of each scene image are sorted and grouped in chronological order to obtain a head-end data set, an intermediate data set and a tail-end data set, wherein the acquisition time of the scene image to which the road three-dimensional data in the head-end data set belongs is earlier than the acquisition time of the scene image to which the road three-dimensional data in the intermediate data set belongs, and the acquisition time of the scene image to which the road three-dimensional data in the intermediate data set belongs is earlier than the acquisition time of the scene image to which the road three-dimensional data in the tail-end data set belongs. For example, for the road three-dimensional data of all scene images, the head-end data set contains the road three-dimensional data of the 10 frames of scene images with the earliest acquisition time, the head-end data set contains the road three-dimensional data of the 10 frames of scene images with the earliest acquisition time, the tail-end data set contains the road three-dimensional data of the 10 frames of scene images with the latest acquisition time, and the remaining road three-dimensional data of each scene image constitute the intermediate data set.

[0157] 403. Perform fitting processing on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set; wherein the first control point and the last control point both have coordinate information.

[0158] Exemplarily, for each 3D point in the road 3D data in the head-end data set and the tail-end data set, the device first determines the coordinate information of each 3D point in the preset global coordinate system, and based on the preset fitting algorithm, performs fitting processing on these 3D points according to the coordinate information of these 3D points to obtain the fitted head-end data set and the fitted tail-end data set. Two points can be determined from the fitted head-end data set and the fitted tail-end data set as the first control point and the last control point in the sampling set. For example, a point is determined from the fitted head-end data set as the first control point in the sampling set, and a point is determined from the fitted tail-end data set as the last control point in the sampling set. It is worth noting that the order of the control points in the sampling set can be the processing order of the fitting process in the subsequent fitting algorithm.

[0159] In one example, step 403 includes the following steps:

[0160] The first step of step 403 is to perform polynomial fitting processing on the head-end data set and the tail-end data set to obtain fitting data; wherein the fitting data includes at least two fitting points; the fitting points represent the three-dimensional points of the target road elements; the fitting points have the coordinate information of the three-dimensional points in the preset global coordinate system.

[0161] The second step of step 403 is to determine the first fitting point in the fitting data as the first control point in the sampling set, and to determine the last fitting point in the fitting data as the last control point in the sampling set.

[0162] Exemplarily, in order to make the selected control points more accurate, for each three-dimensional point in the three-dimensional data of the road in the head-end data set and the tail-end data set, the device performs polynomial fitting processing on the head-end data set and the tail-end data set based on a preset polynomial fitting algorithm and according to the coordinate information of each three-dimensional point in a preset global coordinate system, to obtain fitting data such as the fitting data being a fitting equation; wherein the fitting data includes at least two fitting points; each fitting point represents a three-dimensional point of a target road element; each fitting point has the coordinate information of the three-dimensional point in a preset global coordinate system, and the first fitting point in the fitting data is determined as the first control point in the sampling set, and the last fitting point in the fitting data is determined as the last control point in the sampling set; thereby, a more accurate first control point and last control point in the sampling set can be obtained.

[0163] For example, for each 3D point in the road 3D data in the head-end data set and the tail-end data set, the device determines the local coordinate system (I system) through principal component analysis (PCA) and then performs a cubic polynomial fitting, such as according to formula 1 and 2 formula We can get formula 3 yi=a1*xi3+b1*xi2+c1*xi+d1 and formula 4 zi=a2*xi3+b2*xi2+c2*xi+d2; among them, a1, b1, c1 and d1 are the coefficients of cubic polynomial fitting in the y direction, and similarly, a2, b2, c2 and d2 are the coefficients of cubic polynomial fitting in the z direction, and the fitting equation is obtained. The first endpoint is the first fitting point in the fitting equation, and the last endpoint is the last fitting point in the fitting equation.

[0164] 404. Perform multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain each first intermediate point in the sampling set; wherein the first intermediate point is the control point remaining except the first control point and the last control point in the sampling set; the first intermediate point has the coordinate information of the control point in the preset global coordinate system.

[0165] Exemplarily, for each three-dimensional point in the road three-dimensional data in the intermediate data set, the device determines the coordinate information of the three-dimensional point in a preset global coordinate system, and based on a preset multi-frame fusion calculation algorithm, performs multi-frame fusion calculation processing on the three-dimensional points in the road three-dimensional data in the intermediate data set according to the coordinate information of each three-dimensional point, and selects the first intermediate points in the sampling set from the road three-dimensional data in the intermediate data set, that is, the remaining control points except the first control point and the last control point in the sampling set, wherein the coordinate information of each intermediate point is the coordinate information of the point in the preset global coordinate system.

[0166] In one example, step 404 includes the following steps:

[0167] The first step of step 404 is to extract at least one target data set from the intermediate data set according to a preset interval time; wherein the target data set includes road three-dimensional data of a preset number of scene images; and the acquisition moments of each scene image corresponding to the target data set are continuous.

[0168] The second step of step 404 is to determine the first coordinate set corresponding to the target data set; wherein the first coordinate set includes each first coordinate value corresponding to each road three-dimensional data in the target data set; the first coordinate value is the coordinate value of each three-dimensional point in the road three-dimensional data located on the right coordinate axis in the preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction.

[0169] The third step of step 404 is to determine the difference between the first coordinate value and the average value corresponding to each first coordinate value.

[0170] The fourth step of step 404 is to determine the three-dimensional point to which the first coordinate value with the smallest difference in the first coordinate set belongs, as the first intermediate point in the sampling set; wherein the first intermediate point has the coordinate information of the three-dimensional point in the global coordinate system.

[0171] Exemplarily, in order to make the selected control points more accurate, for the intermediate data set, a target data set is extracted from the intermediate data set each time according to a preset interval time, and the target data set includes road three-dimensional data of a preset number of scene images, and the acquisition moments of each scene image corresponding to the target data set are continuous, thereby obtaining one or more target data sets; for example, at every acquisition moment of 8 frames of scene images, 10 frames of road three-dimensional data of scene images are taken from the intermediate data set. For each target data set extracted, the device determines the first coordinate value corresponding to each three-dimensional point in each road three-dimensional data in the target data set, and obtains the first coordinate set corresponding to the target data set, wherein the first coordinate value of each three-dimensional point is the coordinate value of the three-dimensional point on the right coordinate axis in the preset three-dimensional coordinate system, and the forward coordinate axis is a coordinate axis whose direction is consistent with the vehicle's forward direction in the preset three-dimensional coordinate system; such as the vehicle coordinate system (B system), XYZ corresponds to the forward, right and downward directions of the vehicle respectively, that is, the direction of the x-axis in the vehicle coordinate system is the vehicle's forward direction, and the direction of the y-axis is the lateral right direction of the vehicle's forward direction; the direction of the z-axis is perpendicular to the ground, pointing downward, and the first coordinate value is the coordinate value on the right Y-axis. For each first coordinate set, the device calculates the corresponding average value between each first coordinate value in the first coordinate set, and calculates the difference between each first coordinate value in the first coordinate set and the average value, and determines the three-dimensional point to which the first coordinate value with the smallest difference in the first coordinate set belongs as a first intermediate point in the sampling set, and then all the first intermediate points in the sampling set are obtained, wherein the coordinate information of each intermediate point is the coordinate information of the corresponding three-dimensional point in the global coordinate system; thereby, more accurate first intermediate points in the sampling set are obtained to improve the fitting accuracy.

[0172] 405. Determine the slope information of the control point in the sampling set according to the coordinate information of the control point in the sampling set.

[0173] Exemplarily, for the coordinate information of each control point in the sampling set, based on a preset calculation method, the device calculates the coordinate information of each control point in the sampling set to obtain the slope information of each control point in the sampling set; thereby, a more accurate sampling set is obtained. Fig. 9 The denoising effect diagram of the three-dimensional data of a paved road provided in an embodiment of the present application is shown in FIG. Fig. 9 As shown in the figure, accurate selection of control points can ensure that the points are at the center of the road 3D data, thereby ensuring the fitting accuracy.

[0174] In one example, each control point in the sampling set is sorted from small to large according to the forward distance, and the forward distance represents the horizontal distance between the position of the control point and the position of the vehicle; step 405 includes the following steps:

[0175] The first step of step 405 is to determine the first endpoint, the last endpoint and each second intermediate point in the sampling set; wherein the first endpoint is the first control point in the sampling set, the last endpoint is the last control point in the sampling set, and the second intermediate point is the remaining control point in the sampling set except the first endpoint and the last endpoint.

[0176] The second step of step 405 is to perform polynomial derivation processing on the first endpoint and the first endpoint to obtain the slope information of the first endpoint and the slope information of the first endpoint.

[0177] The third step of step 405 is to perform interpolation calculation processing on the second middle point to obtain the slope information of the second middle point.

[0178] Exemplarily, after obtaining the sampling set, the device sorts each control point in the sampling set from small to large according to the forward distance, and the forward distance represents the horizontal distance between the position of each control point and the position of the vehicle, that is, to ensure that all control points are time-sequenced. According to the sorting of each control point in the sampling set, the device determines the first control point in the sampling set as the head end point, determines the last control point in the sampling set as the tail end point, and determines the remaining control points in the sampling set except the head end point and the tail end point as each second intermediate point. Due to the different strategies for selecting the control points at the head, tail and middle, the method for calculating the slope is also different. For the head end point and the tail end point, the device adopts a cubic polynomial fitting calculation method, that is, according to the coordinate information of the head end point and the coordinate information of the head end point, the polynomial derivative processing is performed on the head end point and the tail end point to obtain the slope information of the head end point and the slope information of the head end point. Based on the preset interpolation algorithm, the device performs interpolation calculation processing on each second intermediate point to obtain the slope information of each second intermediate point; then, through different methods for calculating the slope, accurate slope information of each control point is obtained to improve the fitting accuracy.

[0179] For example, the specific fitting method is: let P be the eigenvector matrix solved by PCA, {xg, yg, zg} be the 3D coordinates in the G system, {xi, yi, zi} be the I system calculated by PCA, then: According to formula 1 and 2 formula We can get formula 5 a*xg+b*yg+c*zg=xi, formula 6 d*xg+e*yg+f*zg=yi, and formula 7 g*xg+h*yg+i*zg=zi; taking the derivatives of formula 5, formula 6, and formula 7 with respect to xg respectively, we can get formula 8 a+b*dyg / dxg+c*dzg / dxg=dxi / dxg, formula 9 d+e*dyg / dxg+f*dzg / dxg=dyi / dxg, and formula 10 g+h*dyg / dxg+i*dzg / dxg=dzi / dxg. Let dyg / dxg=k1, dzg / dxg=k2, compare formula 9 with formula 10, and compare formula 10 with formula 8, and we can get formula 11 12 formulas According to formula 3 yi=a1*xi3+b1*xi2+c1*xi+d1, formula 4 zi=a2*xi3+b2*xi2+c2*xi+d2, take the derivative of xi on the left and right sides, record them as k3 and k4 respectively, and we can get formula 13: 14 formulas Among them, xi can be calculated from formula 5; by combining formulas 11 to 14, dyg / dxg can be solved, that is, the slope information of the first and last endpoints can be obtained. At the same time, the slope information k2 of the first endpoint can also be obtained.

[0180] In one example, the third step of step 405 includes:

[0181] Step 1: Determine the front neighboring point of the second midpoint and the rear neighboring point of the second midpoint from the sampling set; wherein the front neighboring point is a control point that is before the second midpoint in the sampling set and adjacent to the second midpoint; and the rear neighboring point is a control point that is after the second midpoint in the sampling set and adjacent to the second midpoint.

[0182] Step 2: Determine the slope information of the second midpoint according to the second midpoint, the front adjacent point of the second midpoint, and the rear adjacent point of the second midpoint.

[0183] Exemplarily, accurately solving the slope of the control point has a very important influence on the fitting accuracy of the horn spiral. Therefore, for each second midpoint in the sampling set, the device determines the control point in the sampling set that is before the second midpoint and adjacent to the second midpoint as the front neighboring point of the second midpoint, and determines the control point in the sampling set that is after the second midpoint and adjacent to the second midpoint as the rear neighboring point of the second midpoint; based on the coordinate information of the second midpoint, the front neighboring point of the second midpoint, and the rear neighboring point of the second midpoint, the three points can be interpolated to obtain the slope information of the second midpoint; the slope information of the solved control point can be further improved, and the fitting accuracy can be improved.

[0184] In one example, step 2 includes:

[0185] The first step is to determine the corresponding angle information between the second midpoint, the front adjacent point, and the rear adjacent point; wherein the angle information represents the angle between the first vector and the second vector of the second midpoint in the preset coordinate system; the first vector is a vector composed of the second midpoint and the front adjacent point of the second midpoint, and the second vector is a vector composed of the second midpoint and the rear adjacent point of the second midpoint.

[0186] The second step is to determine the slope information of the second midpoint according to the second vector of the second midpoint if it is determined that the angle represented by the angle information is less than or equal to the preset angle.

[0187] The third step is to perform fitting calculation processing on the second middle point, the front adjacent point, and the rear adjacent point to obtain the slope information of the second middle point if it is determined that the angle represented by the angle information is greater than the preset angle.

[0188] Specifically, for each second midpoint, the device first determines the vector formed by the second midpoint and the front adjacent point of the second midpoint, that is, the first vector of the second midpoint, and determines the vector formed by the second midpoint and the rear adjacent point of the second midpoint, that is, the second vector of the second midpoint, and then can calculate the angle between the first vector and the second vector of the second midpoint in the preset coordinate system, that is, the angle information corresponding to the second midpoint; compare the calculated angle corresponding to the second midpoint with the preset angle, if it is determined that the angle corresponding to the second midpoint is less than or equal to the preset angle, then based on the preset algorithm, calculate the second vector of the second midpoint to obtain the slope information of the second midpoint; if it is determined that the angle corresponding to the second midpoint is greater than the preset angle, then based on the preset fitting algorithm, perform fitting calculation processing on the second midpoint, the front adjacent point of the second midpoint, and the rear adjacent point of the second midpoint to obtain the slope information of the second midpoint; thereby, the slope information of the solved control point can be further improved, and the fitting accuracy can be improved.

[0189] For example, the intermediate control point is solved by interpolation. The vector v1 formed by the current point Pi and the previous point P(i-1) is calculated, and the vector v2 formed by the next point P(i+1) is calculated. The angle between the vectors v1 and v2 is calculated. According to the size of the angle θ, there are two scenarios: if the angle is within 12 degrees, this section is considered to be a straight line scene, and the current slope is calculated by v2, that is, k = v2y / v2x; if the angle is greater than 12 degrees, it is necessary to further fit the current point Pi, the previous point P(i-1), and the next point P(i+1) to obtain the slope information of the current point Pi.

[0190] In one example, the third step includes:

[0191] The first step is to fit the coordinate information of the second middle point, the coordinate information of the front adjacent point, and the coordinate information of the rear adjacent point to obtain the circle information of the second middle point; wherein the circle information is the information of the fitted circle corresponding to the second middle point, the front adjacent point of the second middle point, and the rear adjacent point of the second middle point.

[0192] The second step is to determine the slope information of the second middle point according to the coordinate information of the second middle point and the circle information.

[0193] Specifically, for each second midpoint, according to a preset fitting algorithm, the coordinate information of the second midpoint, the coordinate information of the front adjacent point of the second midpoint, and the coordinate information of the rear adjacent point of the second midpoint are fitted and calculated to obtain the information of the fitted circle corresponding to the three points, that is, the circle information, such as the circle equation. The coordinate information of the second midpoint and the corresponding circle information are calculated to obtain the slope information of the second midpoint.

[0194] For example, if we fit a circle based on the coordinates of these three points, and let the radius of the circle be r and the center be (x0, y0), we have 16 formulas (xgi-x0) 2 +(ygi-y0) 2 =r 2 By solving the least squares problem, x0, y0 and r can be solved. By differentiating both sides of formula 16 with respect to xgi, the slope of point Pi can be obtained as k = (x0-xgi) / (ygi-y0). This can further improve the slope information of the solved control point and improve the fitting accuracy.

[0195] 406. Perform a clevis fitting process on each control point according to the coordinate information and slope information of each control point to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0196] Exemplarily, this step may refer to step 203 and will not be described in detail here.

[0197] In this embodiment, based on the above embodiments, the first endpoint, the last endpoint, and the middle control point for fitting are selected respectively through different strategies, and the strategy is determined based on different slope information to obtain the slope information and coordinate information of each control point in the sampling set, thereby obtaining a more accurate fitting line to improve the accuracy of the drawn map.

[0198] Fig.10 A schematic diagram of a road data processing device provided in an embodiment of the present application is shown in FIG. Fig.10 As shown, the device comprises:

[0199] The acquisition module 501 is used to acquire the road three-dimensional data of each scene image in at least one scene image; wherein the scene image represents the driving scene in which the vehicle is located; and the road three-dimensional data includes the three-dimensional points of the target road elements in the scene image.

[0200] The sampling module 502 is used to perform sampling and calculation processing on the road three-dimensional data of each scene image to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents the three-dimensional point of the target road element; the control point has the coordinate information and slope information of the three-dimensional point in a preset global coordinate system.

[0201] The fitting module 503 is used to perform a clevis fitting process on each control point according to the coordinate information and slope information of each control point, and generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

[0202] In one possible implementation, the scene image has a collection time; the sampling module 502 is specifically used to: group the three-dimensional data of each road to obtain a head-end data set, an intermediate data set, and a tail-end data set; wherein the collection time of the scene image to which the road three-dimensional data in the head-end data set belongs is earlier than the collection time of the scene image to which the road three-dimensional data in the intermediate data set belongs; the collection time of the scene image to which the road three-dimensional data in the intermediate data set belongs is earlier than the collection time of the scene image to which the road three-dimensional data in the tail-end data set belongs; and determine the sampling set based on the head-end data set, the intermediate data set, and the tail-end data set.

[0203] In a possible implementation, the sampling module 502 is specifically used to: perform fitting processing on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set; wherein the first control point and the last control point both have coordinate information; perform multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain each first intermediate point in the sampling set; wherein the first intermediate point is the remaining control point except the first control point and the last control point in the sampling set; the first intermediate point has the coordinate information of the control point in the preset global coordinate system; and determine the slope information of the control point in the sampling set based on the coordinate information of the control point in the sampling set.

[0204] In one possible implementation, the sampling module 502 is specifically used to: perform polynomial fitting processing on the head-end data set and the tail-end data set to obtain fitting data; wherein the fitting data includes at least two fitting points; the fitting points represent three-dimensional points of the target road elements; the fitting points have coordinate information of the three-dimensional points in a preset global coordinate system; determine the first fitting point in the fitting data as the first control point in the sampling set, and determine the last fitting point in the fitting data as the last control point in the sampling set.

[0205] In a possible implementation, the sampling module 502 is specifically used to: extract at least one target data set from the intermediate data set according to a preset interval duration; wherein the target data set includes road three-dimensional data of a preset number of scene images; the acquisition moments of each scene image corresponding to the target data set are continuous; determine a first coordinate set corresponding to the target data set; wherein the first coordinate set includes each first coordinate value corresponding to each road three-dimensional data in the target data set; the first coordinate value is the coordinate value of each three-dimensional point in the road three-dimensional data located on the right coordinate axis in the preset three-dimensional coordinate system; the pointing direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction; determine the difference between the first coordinate value and the corresponding average value between each first coordinate value; determine the three-dimensional point to which the first coordinate value with the smallest difference in the first coordinate set belongs, as the first intermediate point in the sampling set; wherein the first intermediate point has the coordinate information of the three-dimensional point in the global coordinate system.

[0206] In a possible implementation, each control point in the sampling set is sorted from small to large according to the forward distance, and the forward distance represents the horizontal distance between the position of the control point and the position of the vehicle; the sampling module 502 is specifically used to: determine the first endpoint, the tail endpoint and each second intermediate point in the sampling set; wherein the first endpoint is the first control point in the sampling set, the tail endpoint is the last control point in the sampling set, and the second intermediate point is the remaining control point in the sampling set except the first endpoint and the tail endpoint; perform polynomial derivative processing on the first endpoint and the first endpoint to obtain the slope information of the first endpoint and the slope information of the first endpoint; perform interpolation calculation processing on the second intermediate point to obtain the slope information of the second intermediate point.

[0207] In a possible implementation, the sampling module 502 is further specifically used to: determine a front neighboring point of the second midpoint and a rear neighboring point of the second midpoint from the sampling set; wherein the front neighboring point is a control point that is arranged before the second midpoint in the sampling set and is adjacent to the second midpoint; the rear neighboring point is a control point that is arranged after the second midpoint in the sampling set and is adjacent to the second midpoint; and determine the slope information of the second midpoint based on the second midpoint, the front neighboring point of the second midpoint, and the rear neighboring point of the second midpoint.

[0208] In a possible implementation, the sampling module 502 is further specifically used to: determine the corresponding angle information between the second midpoint, the front adjacent point, and the rear adjacent point; wherein the angle information represents the angle between the first vector and the second vector of the second midpoint in a preset coordinate system; the first vector is a vector composed of the second midpoint and the front adjacent point of the second midpoint, and the second vector is a vector composed of the second midpoint and the rear adjacent point of the second midpoint; if it is determined that the angle represented by the angle information is less than or equal to the preset angle, then the slope information of the second midpoint is determined according to the second vector of the second midpoint; if it is determined that the angle represented by the angle information is greater than the preset angle, then the second midpoint, the front adjacent point, and the rear adjacent point are fitted and calculated to obtain the slope information of the second midpoint.

[0209] In a possible implementation, the sampling module 502 is further specifically used to: perform fitting processing on the coordinate information of the second midpoint, the coordinate information of the front adjacent point, and the coordinate information of the rear adjacent point to obtain the circle information of the second midpoint; wherein the circle information is the information of the fitted circle corresponding to the second midpoint, the front adjacent point of the second midpoint, and the rear adjacent point of the second midpoint; and determine the slope information of the second midpoint based on the coordinate information and the circle information of the second midpoint.

[0210] In a possible implementation, the device is also used to: determine at least one second coordinate set corresponding to the road three-dimensional data of each scene image; wherein the second coordinate set includes at least one second coordinate value; the second coordinate value is the coordinate value of each three-dimensional point in the road three-dimensional data located on the right coordinate axis in a preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction; each second coordinate value in the second coordinate set is calculated and processed to obtain a coordinate mean corresponding to the second coordinate set; according to the coordinate mean corresponding to the second coordinate set, each road three-dimensional data corresponding to the second coordinate set is filtered out to obtain a processed second coordinate set; wherein the processed second coordinate set includes three-dimensional points in each road three-dimensional data whose second coordinate values ​​are in a reserved interval corresponding to the second coordinate set; the reserved interval is a preset numerical interval in which the coordinate mean corresponding to the second coordinate set is located.

[0211] In a possible implementation, the device is further specifically used to: determine at least one preset interval corresponding to the second coordinate set, and determine a first number of points in the preset interval; wherein the preset interval is a numerical interval in which a preset second coordinate value is located, and the first number of points represents the total number of three-dimensional points to which the second coordinate value in the second coordinate set is within the preset interval; determine a preset number of target intervals corresponding to the second coordinate set, and determine a second number of points corresponding to the second coordinate set; wherein the target interval is a preset interval within a preset number range in which the first number of points is the largest among all preset intervals; the second number of points is the total number of all three-dimensional points corresponding to the second coordinate set; for each target interval corresponding to the second coordinate set, determine a weight of the target interval according to the first number of points in the target interval and the second number of points corresponding to the second coordinate set; wherein the weight represents the importance of the target interval in all preset intervals; according to the weights of each target interval corresponding to the second coordinate set, perform weighted averaging on the second coordinate values ​​in each target interval corresponding to the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set.

[0212] The device of this embodiment can execute the technical solution in the above method. Its specific implementation process and technical principles are the same and will not be repeated here.

[0213] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. Each module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The functions of each module. In addition, all or part of these modules can be integrated together, or they can be implemented independently. The processing element here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0214] Fig.11 A schematic diagram of a computing device provided in an embodiment of the present application is shown in FIG. Fig.11 As shown, the computing device includes: a memory 601 and a processor 602; the memory 601 is a memory for storing instructions executable by the processor 602.

[0215] The processor 602 is configured to execute the method provided in the above embodiment.

[0216] The computer also includes a receiver 603 and a transmitter 604. The receiver 603 is used to receive instructions and data sent by other devices, and the transmitter 604 is used to send instructions and data to external devices.

[0217] An embodiment of the present application also provides a chip for executing instructions, which is used to execute the technical solution of the processing method in the above embodiment.

[0218] The embodiment of the present application also provides a computer-readable storage medium, in which computer instructions are stored. When the computer instructions are executed on a computer, the computer executes the technical solution of the processing method of the above embodiment.

[0219] An embodiment of the present application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium. When at least one processor executes the computer program, the technical solution of the processing method in the above embodiment can be implemented.

[0220] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.

[0221] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to implement the solution of this embodiment.

[0222] In addition, each functional module in each embodiment of the present application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The above-mentioned module-composed unit can be implemented in the form of hardware or in the form of hardware plus software functional units.

[0223] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0224] An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic control unit or a main control device.

[0225] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0226] It should be further noted that, although the various steps in the flowchart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0227] It should be understood that the above-mentioned device embodiments are only illustrative, and the device of the present application can also be implemented in other ways. For example, the division of units / modules in the above-mentioned embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units, modules or components can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0228] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0229] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for processing road data, characterized in that: include: Acquire three-dimensional road data of each scene image in at least one scene image; wherein the scene image represents a driving scene in which the vehicle is located; and the three-dimensional road data includes three-dimensional points of target road elements in the scene image; Sampling and calculating the road three-dimensional data of each of the scene images to obtain a sampling set; wherein the sampling set includes at least one control point, and the control point represents a three-dimensional point of a target road element; the control point has coordinate information and slope information of the three-dimensional point in a preset global coordinate system; According to the coordinate information and slope information of each control point, a clevis fitting process is performed on each control point to generate a fitting line corresponding to the target road element; wherein the fitting line is used to generate a map.

2. The method according to claim 1, characterized in that: The scene images have a collection time; sampling and calculating are performed on the three-dimensional road data of each of the scene images to obtain a sampling set, including: The three-dimensional road data are grouped and processed to obtain a head-end data set, an intermediate data set, and a tail-end data set; wherein the acquisition time of the scene image to which the three-dimensional road data in the head-end data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs; and the acquisition time of the scene image to which the three-dimensional road data in the intermediate data set belongs is earlier than the acquisition time of the scene image to which the three-dimensional road data in the tail-end data set belongs; The sampling set is determined according to the head-end data set, the middle data set and the tail-end data set.

3. The method according to claim 2, characterized in that Determining the sampling set according to the head-end data set, the intermediate data set, and the tail-end data set includes: Performing fitting processing on the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set; wherein the first control point and the last control point both have coordinate information; Perform multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain each first intermediate point in the sampling set; wherein the first intermediate point is the control point remaining except the first control point and the last control point in the sampling set; the first intermediate point has the coordinate information of the control point in the preset global coordinate system; The slope information of the control point in the sampling set is determined according to the coordinate information of the control point in the sampling set.

4. The method according to claim 3, characterized in that Fitting the head-end data set and the tail-end data set to obtain the first control point and the last control point in the sampling set includes: Performing polynomial fitting processing on the head-end data set and the tail-end data set to obtain fitting data; wherein the fitting data includes at least two fitting points; the fitting points represent three-dimensional points of target road elements; the fitting points have coordinate information of the three-dimensional points in the preset global coordinate system; The first fitting point in the fitting data is determined to be the first control point in the sampling set, and the last fitting point in the fitting data is determined to be the last control point in the sampling set.

5. The method according to claim 3, characterized in that: Performing multi-frame fusion calculation processing on the three-dimensional road data in the intermediate data set to obtain the first intermediate point in the sampling set includes: Extracting at least one target data set from the intermediate data set according to a preset interval time; wherein the target data set includes three-dimensional road data of a preset number of scene images; and the acquisition moments of the scene images corresponding to the target data set are continuous; Determine a first coordinate set corresponding to the target data set; wherein the first coordinate set includes first coordinate values ​​corresponding to each road three-dimensional data in the target data set; the first coordinate value is a coordinate value of each three-dimensional point in the road three-dimensional data located on a right coordinate axis in a preset three-dimensional coordinate system; the right coordinate axis points to the right side of the vehicle's forward direction; A first middle point in the sampling set is determined according to the first coordinate value and each of the first coordinate values; wherein the first middle point has coordinate information of a three-dimensional point in the global coordinate system.

6. The method according to claim 3, characterized in that The control points in the sampling set are sorted in ascending order according to the forward distance, where the forward distance represents the horizontal distance between the position of the control point and the position of the vehicle; Determining the slope information of the control point in the sampling set according to the coordinate information of the control point in the sampling set includes: Determine the first endpoint, the last endpoint, and each second intermediate point in the sampling set; wherein the first endpoint is the first control point in the sampling set, the last endpoint is the last control point in the sampling set, and the second intermediate point is the remaining control point in the sampling set except the first endpoint and the last endpoint; Performing polynomial derivation processing on the first endpoint and the first end point to obtain slope information of the first endpoint and slope information of the first end point; An interpolation calculation process is performed on the second middle point to obtain slope information of the second middle point.

7. The method according to claim 6, characterized in that Performing interpolation calculation processing on the second intermediate point to obtain slope information of the second intermediate point includes: Determine a front neighboring point of the second midpoint and a rear neighboring point of the second midpoint from the sampling set; wherein the front neighboring point is a control point that is arranged before the second midpoint in the sampling set and is adjacent to the second midpoint; and the rear neighboring point is a control point that is arranged after the second midpoint in the sampling set and is adjacent to the second midpoint; The slope information of the second intermediate point is determined according to the second intermediate point, a previous adjacent point of the second intermediate point, and a subsequent adjacent point of the second intermediate point.

8. The method according to any one of claims 1 to 7, characterized in that The method further comprises: Determine at least one second coordinate set corresponding to the three-dimensional road data of each of the scene images; wherein the second coordinate set includes at least one second coordinate value; the second coordinate value is the coordinate value of each three-dimensional point in the three-dimensional road data located on the right coordinate axis in a preset three-dimensional coordinate system; the direction of the right coordinate axis is the lateral right direction of the vehicle's forward direction; Calculating and processing each second coordinate value in the second coordinate set to obtain a coordinate mean corresponding to the second coordinate set; According to the coordinate mean corresponding to the second coordinate set, the three-dimensional data of each road corresponding to the second coordinate set are filtered out to obtain a processed second coordinate set; wherein the processed second coordinate set includes three-dimensional points whose second coordinate values ​​in the three-dimensional data of each road are in a reserved interval corresponding to the second coordinate set; the reserved interval is a preset numerical interval in which the coordinate mean corresponding to the second coordinate set is located.

9. A computing device, characterized in that include: Memory, processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor performs the method according to any one of claims 1 to 8.

10. A computer-readable storage medium / computer program product, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-8; and / or the computer program product includes a computer program, which, when executed by a processor, implements the method as described in any one of claims 1-8.