Map fence acquisition method, system, map, storage medium and device

Abstract

The present application discloses a method, system, map, storage medium and device for obtaining a map fence, and belongs to the field of autonomous driving technology. The method mainly includes receiving multi-frame echo point cloud data collected in real time by a millimeter-wave radar; using the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose density of echo point cloud points is greater than a predetermined threshold and distributed in a straight line; fitting the echo point cloud into a corresponding fitting line segment, and calculating the length of the fitting line segment; if the length of the fitting line segment is greater than the preset length threshold, and the number of times the echo point cloud corresponding to the fitting line segment appears in the multi-frame echo point cloud data collected in real time is greater than the first threshold, then the fitting line segment is determined as the current metal fence line segment corresponding to the current frame. The present application can improve the accuracy of metal fence recognition, and at the same time, the autonomous driving vehicle can locate its own position on the road according to the metal fence map, thereby ensuring the safety of autonomous driving.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a method, system, map, storage medium, and device for obtaining a map barrier. Background Art

[0002] Millimeter-wave radar has become an indispensable sensor for environmental perception in autonomous vehicles. Millimeter-wave radar operates in the millimeter-wave band. Millimeter waves typically have a wavelength of 1 to 10 mm. Because their wavelength lies between centimeter waves and light waves, millimeter-wave radar combines the advantages of microwave guidance and photoelectric guidance. To reduce or even prevent traffic accidents, millimeter-wave radar is being used to provide safety alerts. The characteristics of millimeter waves ensure their adaptability to harsh weather conditions. For example, in rainy and foggy conditions with low visibility, where collision avoidance systems are crucial, laser and ultrasonic sensors may not function properly, but millimeter-wave radar is not affected. Furthermore, millimeter-wave radar antennas are immune to dust contamination, which can cause errors, making them suitable for road use.

[0003] A metal fence is a protective barrier consisting of metal posts and horizontal strips. Currently, many roads use metal fences to regulate traffic order. Metal fences clarify the right of way for all parties, reduce interference between motor vehicles, non-motor vehicles, and pedestrians, and prevent pedestrians and non-motor vehicles from crossing sideways. Therefore, when autonomous vehicles are driving on the road, they need to accurately detect metal fences between the auxiliary road and the main road to avoid collisions. Since metal fences are extremely common on roads, there is currently no effective method to identify and determine their location. Therefore, identifying and determining the location of metal fences on roads has become a pressing issue. Summary of the Invention

[0004] In response to the problem that the existing technology cannot effectively identify metal fences on the road, this application mainly provides a method, system, map, storage medium and device for obtaining map fences.

[0005] A technical solution adopted in the present application is: providing a method for obtaining a map fence, which includes: receiving multiple frames of echo point cloud data collected in real time by a millimeter-wave radar; using the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose echo point cloud density is greater than a predetermined threshold and is distributed in a straight line; fitting the echo point cloud into a corresponding fitting line segment, and calculating the length of the fitting line segment; if the length of the fitting line segment is greater than a preset length threshold, and the echo point cloud data corresponding to the fitting line segment appears more than a first threshold number of times in the multiple frames of echo point cloud data collected in real time, then the fitting line segment is determined as the current metal fence line segment corresponding to the current frame.

[0006] Another technical solution adopted in the present application is: providing a map fence acquisition system, which includes: a receiving module, which is used to receive multi-frame echo point cloud data collected in real time by millimeter wave radar; an echo point cloud determination module, which is used to use the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose echo point cloud density is greater than a predetermined threshold and is distributed in a straight line; a fitting calculation module, which is used to fit the echo point cloud into a corresponding fitting line segment and calculate the length of the fitting line segment; a metal fence line segment determination module, which is used to determine the current metal fence line segment corresponding to the current frame by the fitting line segment if the length of the fitting line segment is greater than a preset length threshold and the number of times the echo point cloud corresponding to the fitting line segment appears in the multi-frame echo point cloud data collected in real time is greater than a first threshold.

[0007] Another technical solution adopted by the present application is to provide a map, which includes: metal fence line segments obtained according to the above-mentioned map fence acquisition method.

[0008] Another technical solution adopted by the present application is: providing a computer-readable storage medium storing computer instructions, which are operated to execute any of the above-mentioned map fence acquisition methods.

[0009] Another technical solution adopted in the present application is: providing a computer device, which includes a processor and a memory, the memory storing computer instructions, wherein the processor operates the computer instructions to execute any of the above-mentioned map fence acquisition methods.

[0010] The technical solution of the present application can achieve the following beneficial effects: the present application receives multiple frames of echo point cloud data collected in real time by millimeter-wave radar; uses the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose density of echo point cloud points is greater than a predetermined threshold and can be fitted into a line segment; fits the echo point cloud that can be fitted into a line segment into a corresponding line segment, and calculates the length of the fitted line segment; if the length of the fitted line segment is greater than the preset length threshold, and the echo point cloud data corresponding to the fitted line segment appears more than the first threshold in the real-time collected echo point cloud data, then the fitted line segment is determined to be the current metal fence line segment corresponding to the current frame. The technical solution of the present application utilizes the principle that the millimeter-wave radar has a strong reflection echo to metal, and can identify metal fences on the road. Then, by using the preset length threshold and the first threshold, it can improve the accuracy of metal fence recognition. At the same time, the autonomous driving vehicle can locate its own position on the road according to the metal fence map, ensuring the safety of autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a schematic diagram of a system architecture applicable to a method for obtaining a map fence in the present application;

[0012] Figure 2This is a schematic diagram of a specific implementation of a method for obtaining a map fence in the present application;

[0013] Figure 3 This is a schematic diagram of a specific embodiment of a method for obtaining a map fence according to the present application;

[0014] Figure 4 It is a schematic diagram of a specific embodiment of a map fence acquisition system of the present application. DETAILED DESCRIPTION

[0015] The preferred embodiments of the present application are described in detail below in conjunction with the accompanying drawings so that the advantages and features of the present application can be more easily understood by those skilled in the art, thereby making a clearer and more definite definition of the protection scope of the present application.

[0016] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, the elements defined by the phrase "comprising..." do not exclude the presence of other identical elements in the process, method, article, or device comprising the elements.

[0017] Before explaining the specific embodiments of the present application in detail, the terms involved in the specific embodiments of the present application are first explained.

[0018] Fences: A protective barrier consisting of fence posts and horizontal strips, typically installed between main and secondary roads in urban areas, regulate traffic flow and reduce interference between motor vehicles, non-motor vehicles, and pedestrians. Road fences are typically metal and range in height from 0.5 to 2 meters. They can consist of one or more sections, with gaps between adjacent sections. They are generally arranged in a straight line, demonstrating a linear nature.

[0019] Echo point cloud data refers to the record of the scanned object in the form of three-dimensional data. Each echo point cloud contains corresponding three-dimensional coordinates, and some may contain color information (RGB) or reflection intensity information (Intensity). Exemplarily, the vehicle can collect echo point cloud data in the environment in which the vehicle is located while driving. As an example, the vehicle can use an on-board millimeter wave scanning system to perform real-time scanning of the environment in which it is located to obtain echo point cloud data. As another example, the vehicle can use an on-board millimeter wave radar to perform real-time scanning of the environment in which it is located to obtain echo point cloud data. The environment in which the vehicle is located generally includes metal fences, the ground, and obstacles on the ground. Therefore, the echo point cloud data includes metal fence echo point cloud data and / or obstacle echo point cloud data. The vehicle can be a collection vehicle dedicated to collecting echo point cloud data, or it can be an autonomous driving vehicle, an unmanned delivery vehicle, or other ordinary vehicles equipped with millimeter wave radars. The echo point cloud data of the present application is collected using a millimeter wave radar.

[0020] Next, the application scenarios involved in the specific implementation methods or specific examples of this application are described. The map fence acquisition method provided in this application is applied to identify and determine the position of metal fences on the road, for example, to identify and determine the position of metal fences on urban roads. As an example, in an autonomous driving scenario, when an autonomous driving vehicle is traveling on urban roads, the metal fences on the road can be identified and determined in accordance with the method provided in this application, so as to accurately identify the metal fences between the auxiliary road and the main road to avoid collisions. As another example, in an unmanned delivery scenario, when an unmanned delivery vehicle is traveling on urban roads to deliver items, the metal fences on the road can also be identified in accordance with the method provided in this application to avoid collisions. Among them, the unmanned delivery vehicle can be used to deliver express or takeout items.

[0021] Next, the implementation environment involved in the specific implementation or specific embodiment of this application is introduced. The method for obtaining a map fence provided by this application is applied to an electronic device, which can be installed in a vehicle so as to identify metal fences in the vehicle's environment while the vehicle is driving. The electronic device can be a vehicle-mounted terminal, a processor, a vehicle-mounted computer or an on-board unit (OBU), etc., and the electronic device is used to control the automatic driving of the vehicle. Figure 1 The system architecture to which the map fence acquisition method provided by this application is applicable is shown as follows: Figure 1As shown, the map fence acquisition system of the present application may include: a millimeter-wave radar and an onboard computer. The millimeter-wave radar can be connected to the onboard computer via a Controller Area Network (CAN) bus. Furthermore, the millimeter-wave radar can transmit the acquired echo point cloud data via the CAN bus to the onboard computer, so that the onboard computer can perform the map fence acquisition method provided in the present application on the echo point cloud data to identify metal fence segments.

[0022] Next, the method for obtaining the map fence of this application is described in detail. Figure 2 A specific embodiment of a method for obtaining a map fence of the present application is shown. The method uses a millimeter wave radar to collect echo point cloud data, and the vehicle-mounted computer is used as the execution body. Figure 2 In the specific embodiment shown, the method for obtaining a map barrier of the present application may include: Step S201. This step mainly includes: receiving multiple frames of echo point cloud data collected in real time by a millimeter wave radar.

[0023] In this specific embodiment, a millimeter-wave radar is installed on the vehicle. Utilizing the principle that millimeter-wave radar reflects strong echoes from metal, it can obtain parameter information such as the distance and shape of obstacles and / or metal fences in the vehicle's environment, namely, echo point cloud data of the obstacles and / or metal fences. This echo point cloud data of the obstacles and / or metal fences is then transmitted to the onboard computer. This specific embodiment does not provide a detailed explanation of the principle of collecting echo point cloud data by the millimeter-wave radar.

[0024] Since the echo point cloud data collected continuously in real time by the millimeter-wave radar may contain both metal fence echo point cloud data and obstacle echo point cloud data, in order to improve the accuracy of metal fence recognition, it is necessary to filter out the obstacle echo point cloud data as much as possible and only process the metal fence echo point cloud data.

[0025] exist Figure 2 In the specific embodiment shown, the map barrier acquisition method of the present application may include: step S202. This step mainly includes: using the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose point density is greater than a predetermined threshold and is distributed in a straight line.

[0026] In this specific embodiment, utilizing the principle that millimeter-wave radar has a stronger reflection echo from metal, the real-time collected echo point cloud data will not only include the echo point cloud data of metal fences, but may also include the echo point cloud data of obstacles such as other vehicles, roadside metal signs or indicator lights. However, the density of obstacle echo point clouds may be less than that of metal fence echo point clouds. Therefore, setting a predetermined threshold can filter out some obstacle echo point clouds, reduce false detections, and improve the accuracy of the metal fence echo point cloud. This application does not impose any specific restrictions on the value of the predetermined threshold, and those skilled in the art can set it according to actual circumstances.

[0027] At the same time, since metal fences are generally distributed in straight lines and have straight line characteristics, they appear in the form of straight lines or line segments in the echo point cloud data collected by the millimeter-wave radar. Therefore, only the echo point cloud that can be fitted into a line segment is likely to be the metal fence echo point cloud. This can also filter out some obstacle echo point clouds that cannot be fitted into a line segment, reduce false detections, and further improve the accuracy of the metal fence echo point cloud.

[0028] exist Figure 2 In the specific embodiment shown, the map barrier acquisition method of the present application may include: step S203. This step mainly includes: fitting the echo point cloud into corresponding fitting line segments, and calculating the length of the fitting line segments.

[0029] In this specific embodiment, because the metal fence appears in the echo point cloud of the millimeter wave radar as a straight line or line segment, the line segment in the echo point cloud can be found through a straight line search method, thereby ensuring that the metal fence can be accurately identified.

[0030] In one example of the present application, a line-finding method can be used to perform line segment fitting on echo point cloud data that can be fitted into line segments, wherein the method can be a Hough transform method or an LSD (Line Segment Detector) method, etc. The advantages of the Hough transform method are strong anti-interference ability, insensitivity to incomplete parts of straight lines in the image, noise, and other coexisting non-linear structures, tolerance of gaps in feature boundary descriptions, and relative immunity to image noise. However, the characteristics of the Hough transform method result in high time complexity and space complexity, and only the direction of the straight line can be determined during the detection process, and the length information of the line segment is lost. Since discretization is performed during the Hough transform process, the detection accuracy is constrained by the discrete interval of the parameters.

[0031] In another example of the present application, multiple straight line finding methods can be used to perform line segment fitting on the echo point cloud data that can be fitted into line segments. Then, based on the fitting results obtained by the multiple straight line finding methods, the fitting line segments can be comprehensively determined to ensure the accuracy of the fitting line segments.

[0032] Specifically, because millimeter-wave radar can provide information such as the distance, angle, radial velocity, and target number of the echo point cloud from the vehicle, the onboard computer uses a millimeter-wave radar data parsing program to analyze the echo point cloud data continuously collected by the millimeter-wave radar in real time to determine the position of the echo point cloud relative to the vehicle. A millimeter-wave radar data positioning program is then used to generate a two-dimensional coordinate map of these echo point clouds, with the vehicle's position as the X-axis and the vehicle's travel direction as the Y-axis. The echo point cloud data is then used to identify those echo point clouds whose point density exceeds a predetermined threshold and can be fitted into a line segment. Finally, a line-finding method is used to obtain the fitted line segment.

[0033] In a specific embodiment of the present application, metal fences may appear on the left and / or right sides of a vehicle during its travel. Since the vehicle generally travels forward normally, a millimeter-wave radar can be installed at a certain position in front of the vehicle, specifically in the middle of the vehicle's front bumper. The millimeter-wave radar plane is perpendicular to the ground. This allows the millimeter-wave radar to more promptly collect echo point cloud data in front of the vehicle, including the echo point cloud of the metal fence. Furthermore, during normal forward travel, it is generally necessary to observe the metal fences on the left and / or right front sides of the vehicle. Therefore, the fitted line segments are those located on the left and / or right front sides of the millimeter-wave radar.

[0034] exist Figure 2 In the specific embodiment shown, the method for obtaining a map fence of the present application may include: Step S204. This step mainly includes: if the length of the fitting line segment is greater than a preset length threshold, and the number of times the echo point cloud corresponding to the fitting line segment appears in the multi-frame echo point cloud data collected in real time is greater than a first threshold, then determining the fitting line segment as the current metal fence line segment corresponding to the current frame.

[0035] In this specific embodiment, during the real-time continuous acquisition process of the millimeter-wave radar, other obstacle echo point clouds may also generate short line segments, but these short line segments are short in length. Therefore, a preset length threshold can be set to filter out some short line segments generated by obstacle echo point clouds, reducing false detections and improving the accuracy of metal fence line segment recognition.

[0036] It should be noted that the onboard computer can perform frame processing on the echo point cloud data continuously acquired in real time by the millimeter-wave radar. For example, the echo point cloud data can be framed based on the acquisition time. For example, point cloud data acquired from 1-5 seconds can be used as one frame of echo point cloud data, while point cloud data acquired from 5-10 seconds can be used as another frame of echo point cloud data. This can also generate multiple frames of echo point cloud data.

[0037] During real-time, continuous acquisition by millimeter-wave radar, the line segments generated by the metal fence echo point cloud are relatively stable, though they can overlap partially across multiple frames. However, the short line segments generated by other obstacle echo point clouds are unstable. Therefore, setting a first threshold can filter out these unstable short line segments, reducing false detections and improving the accuracy of metal fence line segment recognition.

[0038] In this application, the values of the preset length threshold and the first threshold are not specifically limited. Those skilled in the art can set the preset length threshold and the first threshold based on actual conditions. For example, the preset length threshold and the first threshold can be set based on the actual conditions of the collected road metal fence. For example, if the road metal fence is continuous and dense, or the road metal fence is very long, the preset length threshold and the first threshold can be set to larger values. If the road metal fence is only one section, the preset length threshold and the first threshold can be set to smaller values. For example, the preset length threshold and the first threshold can be set based on the vehicle speed, the acquisition time of the millimeter wave radar, and the number of frames divided according to the acquisition time. Figure 3 The metal fence segment is identified with the preset length threshold set to 10m. Figure 3 Just illustrative.

[0039] In a specific embodiment of the present application, the process of determining the fitting line segment as the current metal fence line segment corresponding to the current frame includes: obtaining the proportion of the echo point cloud corresponding to the fitting line segment in all echo point clouds in the current frame; if the proportion is greater than a second threshold, the fitting line segment corresponding to the current frame is determined as the current metal fence line segment.

[0040] In this specific embodiment, during the real-time continuous acquisition process of the millimeter-wave radar, the metal fence echo point cloud may appear in multiple consecutive frames, that is, the echo point clouds in multiple frames partially overlap, so a second threshold can be set for the proportion of the overlapping part of the echo point cloud corresponding to the fitted line segment in multiple frames to filter out unstable short line segments generated by the echo point cloud of some obstacles, reduce false detection, and improve the accuracy of metal fence line segment recognition. This application does not impose a specific restriction on the value of the second threshold, and those skilled in the art can set it according to actual conditions. For example, in the five frames of echo point cloud acquired continuously in real time, the proportion of the echo point cloud corresponding to the fitted line segment in the fifth frame of some echo point cloud data needs to be greater than the second threshold.

[0041] In a specific embodiment of the present application, the larger the first threshold is set, the smaller the second threshold is set.

[0042] In this specific embodiment, the more times the metal fence echo point cloud appears in multiple frames, the smaller its proportion in the current frame. Therefore, those skilled in the art can set the first threshold and the second threshold accordingly according to actual conditions to ensure that the metal fence echo point cloud is judged to avoid missed detection.

[0043] In a specific embodiment of the present application, after determining the fitting line segment as the metal fence line segment corresponding to the current frame, the method also includes: finding the first historical frame adjacent to the current frame, wherein the first historical frame has a corresponding first historical metal fence line segment; updating the current metal fence line segment according to the first historical metal fence line segment to obtain an updated metal fence line segment corresponding to the current frame.

[0044] In a specific embodiment of the present application, the current metal fence segment is updated according to the first historical metal fence segment to obtain an updated metal fence segment corresponding to the current frame, including: obtaining the historical slope of the first historical metal fence segment relative to the driving direction and the current slope of the current metal fence segment relative to the driving direction; if the difference between the historical slope and the current slope is less than a predetermined slope difference threshold, the current metal fence segment is spliced with the first historical metal fence segment to obtain an updated metal fence segment.

[0045] In this specific embodiment, if the current frame can determine the current metal fence segment, the first historical frame adjacent to the current frame is found, where if the difference between the slope of the first historical metal fence segment corresponding to the first historical frame relative to the driving direction and the slope of the current metal fence segment relative to the driving direction is less than a predetermined slope difference threshold, the current metal fence segment is spliced with the first historical metal fence segment.

[0046] In a specific embodiment of the present application, the current metal fence segment is updated according to the first historical metal fence segment to obtain an updated metal fence segment corresponding to the current frame, and also includes: determining the overlapping length of the first historical metal fence segment and the current metal fence segment according to the positions of the first historical metal fence segment and the current metal fence segment; if the overlapping length is greater than a predetermined overlapping length threshold, the current metal fence segment is spliced with the first historical metal fence segment to obtain an updated metal fence segment.

[0047] In this embodiment, if both the current frame and the previous frame can determine the metal fence line segment, but do not meet the predetermined slope difference threshold and the predetermined overlap length threshold, it means that the current metal fence line segment and the metal fence line segment of the previous frame may be false metal fence line segments. In order to further prove the authenticity of the current metal fence line segment and the metal fence line segment of the previous frame, that is, to prove that the current metal fence line segment and the metal fence line segment of the previous frame may be false detections, it can be further determined whether the previous two frames can determine the metal fence line segment. If not, it means that the vehicle may have driven out of the fence range in the previous two frames, proving that the current metal fence line segment and the metal fence line segment of the previous frame may be false detections. If yes, the current metal fence segment and the metal fence segments of the previous two frames are checked to see if they meet the predetermined slope difference threshold and the predetermined overlap length threshold. If so, the metal fence segment of the previous frame may be a segment generated by an obstacle, and the metal fence segment of the previous frame is filtered out. This can avoid false detections and thus ensure the accuracy of metal fence segment recognition. If not, the metal fence segment of the previous frame and the current metal fence segment may be false metal fence segments, which can reduce false detections and thus ensure the accuracy of metal fence segment recognition.

[0048] In the present application, because the fence has a straight line feature and is generally parallel to the vehicle, in this specific embodiment, it is determined whether the current metal fence segment and the first historical metal fence segment meet the predetermined slope difference threshold and the predetermined overlap length threshold. This can filter out the metal fence segments generated by some obstacles, and can determine the authenticity of the current metal fence segment and the first historical metal fence segment, reduce false detection and missed detection, and ensure the accuracy of metal fence segment recognition.

[0049] In an example of this specific embodiment, if the current frame can determine the current metal fence line segment, and if the previous frame can also determine the metal fence line segment, the current metal fence line segment and the metal fence line segment of the previous frame are judged.

[0050] If the difference between the slope of the metal fence line segment of the previous frame relative to the driving direction and the slope of the current metal fence line segment relative to the driving direction is less than a predetermined slope difference threshold, and the overlap length of the metal fence line segment of the previous frame and the current metal fence line segment is greater than a predetermined overlap length threshold, then the previous frame is taken as the first historical frame, and the metal fence line segment of the previous frame is taken as the first historical metal fence line segment.

[0051] If the difference between the slope of the metal fence segment of the previous frame relative to the driving direction and the slope of the current metal fence segment relative to the driving direction is greater than or equal to the predetermined slope difference threshold, and / or the overlapping length of the metal fence segment of the previous frame and the current metal fence segment is less than or equal to the predetermined overlapping length threshold, or the metal fence segment cannot be determined in the previous frame, then determine whether the fence can be determined in the previous two frames. If the metal fence segment can be determined in the previous two frames, then determine the current metal fence segment and the metal fence segment of the previous two frames.

[0052] If the difference between the slope of the metal fence line segment of the previous two frames relative to the driving direction and the slope of the current metal fence line segment relative to the driving direction is less than the predetermined slope difference threshold, and the overlapping length of the metal fence line segment of the previous two frames and the current metal fence line segment is greater than the predetermined overlapping length threshold, then the previous two frames are taken as the first historical frames, and the metal fence line segment of the previous two frames is the first historical metal fence line segment; otherwise, the judgment of the previous three frames will continue until the first historical frame is found.

[0053] In this example, if there is no metal fence segment in the previous frame, it will be determined whether there is a metal fence segment in the previous two frames. If there is a metal fence segment in the previous two frames, it may mean that the metal fence segment was missed in the previous frame, or the fences are not continuous and dense, resulting in no metal fence segment in the previous frame. Then, the current frame is identified, which can reduce missed detections and thus ensure the accuracy of metal fence segment identification. If the slope of the current metal fence segment and the metal fence segment of the previous two frames meets the predetermined slope difference threshold and the predetermined overlap length threshold, it means that the current metal fence segment is relatively real, so the metal fence segment of the previous two frames and the current metal fence segment are retained. If it does not meet the requirements, it means that the current metal fence segment may be a false metal fence segment, which can reduce false detections and thus ensure the accuracy of metal fence segment identification.

[0054] In a specific embodiment of the present application, the current metal fence segment is spliced with the first historical metal fence segment to obtain an updated metal fence segment, including: superimposing the overlapping part of the current metal fence segment and the first historical metal fence segment to obtain a superimposed metal fence segment; splicing the non-overlapping part of the current metal fence segment, the non-overlapping part of the first historical metal fence segment, and the superimposed metal fence segment to obtain an updated metal fence segment.

[0055] Preferably, because the fence has a straight line feature, least squares fitting or other fitting methods can be used to fit the metal fence segment between the current metal fence segment and the first historical metal fence segment into a straight line segment, that is, the overlapping part between the end of the first historical metal fence segment and the front end of the current metal fence segment is superimposed.

[0056] In a specific embodiment of the present application, it also includes: determining the overlapping length of the current metal fence segment and the first historical metal fence segment based on the two-dimensional coordinates of the current metal fence segment and the two-dimensional coordinates of the first historical metal fence segment, and when the overlapping length is greater than a predetermined overlapping length threshold, the current metal fence segment and the first historical metal fence segment are spliced.

[0057] The present application receives echo point cloud data collected continuously in real time by a millimeter-wave radar; uses the echo point cloud data to determine echo point clouds whose density of echo point clouds is greater than a predetermined threshold and which can be fitted into line segments; fits the echo point clouds that can be fitted into line segments into corresponding line segments, and calculates the length of the fitted line segments; if the length of the fitted line segment is greater than a preset length threshold, and the echo point cloud data corresponding to the fitted line segment appears greater than a first threshold in the real-time continuously collected echo point cloud data, then the fitted line segment is determined to be a metal fence line segment. The technical solution of the present application utilizes the principle that millimeter-wave radar has a strong reflected echo from metal, and can identify metal fences on the road. By using the preset length threshold and the first threshold, the accuracy of metal fence recognition can be improved. At the same time, the autonomous driving vehicle can locate its own position on the road according to the metal fence map, ensuring the safety of autonomous driving.

[0058] Figure 4 A specific embodiment of a map fence acquisition system of the present application is shown. In this specific embodiment, the map fence acquisition system of the present application may include: module 401, a receiving module, for receiving multiple frames of echo point cloud data collected in real time by a millimeter-wave radar; module 402, an echo point cloud determination module, for using the echo point cloud data in the current frame to determine the echo point cloud in the current frame whose echo point cloud density is greater than a predetermined threshold and distributed in a straight line; module 403, a fitting calculation module, for fitting the echo point cloud into a corresponding fitting line segment and calculating the length of the fitting line segment; module 404, a metal fence line segment determination module, for determining the fitting line segment as the current metal fence line segment corresponding to the current frame if the length of the fitting line segment is greater than a preset length threshold and the number of times the echo point cloud data corresponding to the fitting line appears in the multiple frames of echo point cloud data collected in real time is greater than a first threshold.

[0059] In a specific embodiment of the present application, the map fence acquisition system of the present application may further include: a splicing module. This module is mainly used to:

[0060] If the current frame can determine the current metal fence segment, then the first historical frame adjacent to the current frame is searched, wherein if the difference between the slope of the first historical metal fence segment corresponding to the first historical frame relative to the driving direction and the slope of the current metal fence segment relative to the driving direction is less than a predetermined slope difference threshold, then the current metal fence segment is spliced with the first historical metal fence segment. If the overlap length between the first historical metal fence segment corresponding to the first historical frame and the current metal fence segment is greater than a predetermined overlap length threshold, then the current metal fence segment is spliced with the first historical metal fence segment. In this specific embodiment, the splicing module is further configured to: superimpose the overlapping portion of the current metal fence segment and the first historical metal fence segment to obtain a superimposed metal fence segment; and combine the metal fence segment that does not overlap with the first historical metal fence segment and the superimposed metal fence segment to form an updated metal fence segment of the current frame.

[0061] In this specific embodiment, the splicing module is further used to: determine the overlapping length of the current metal fence segment and the first historical metal fence segment based on the two-dimensional coordinates of the current metal fence segment and the two-dimensional coordinates of the first historical metal fence segment; when the overlapping length is greater than a predetermined overlapping length threshold, the current metal fence segment and the first historical metal fence segment are spliced.

[0062] The map barrier acquisition system provided in this application can be used to execute the map barrier acquisition method described in any of the above embodiments. Its implementation principles and technical effects are similar and will not be repeated here.

[0063] In a specific embodiment of the present application, the receiving module, the echo point cloud determination module, the fitting calculation module and the metal fence line segment determination module in a map fence acquisition system of the present application can be directly in the hardware, in a software module executed by a processor, or in a combination of the two.

[0064] The software modules may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from and write information to the storage medium.

[0065] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic, discrete hardware components, or any combination thereof. A general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors combined with a DSP core, or any other such configuration. In the alternative, the storage medium may be integral to the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In the alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0066] In another specific embodiment of the present application, a map includes: metal fence line segments obtained according to the above-mentioned map fence acquisition method.

[0067] In another specific embodiment of the present application, a computer-readable storage medium stores computer instructions, wherein the computer instructions are operated to execute the method for obtaining a map barrier described in any embodiment.

[0068] In another specific embodiment of the present application, a computer device includes a processor and a memory, wherein the memory stores computer instructions, wherein the processor operates the computer instructions to execute any of the above-mentioned map barrier acquisition methods.

[0069] In the several embodiments provided in this 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 merely schematic. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components 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 through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0070] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0071] The above description is merely an embodiment of the present application and does not limit the patent scope of the present application. Any equivalent structural transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for obtaining a map fence, characterized in that: include: Receiving multiple frames of echo point cloud data collected continuously in real time by a millimeter-wave radar, wherein the echo point cloud includes a metal fence echo point cloud and an obstacle echo point cloud, and the obstacle echo point cloud is filtered out of the echo point cloud; Determining, using the echo point cloud data in the current frame, an echo point cloud in the current frame whose density of echo point clouds is greater than a predetermined threshold and which is distributed in a straight line; Fitting the echo point cloud into a corresponding fitting line segment, calculating the length of the fitting line segment, and obtaining the proportion of the echo point cloud corresponding to the fitting line segment in all echo point clouds in the current frame; If the length of the fitting line segment is greater than a preset length threshold, and the number of times the echo point cloud data corresponding to the fitting line segment appears in the multi-frame echo point cloud data collected in real time is greater than a first threshold, and if the proportion is greater than a second threshold, then the fitting line segment corresponding to the current frame is determined as the current metal fence line segment, wherein the second threshold reflects the proportion of the echo point cloud corresponding to the fitting line segment of each frame in the multi-frame echo point cloud that is the overlapped part of the point cloud; After determining the fitting line segment as the metal fence line segment corresponding to the current frame, the method further includes: Finding a first historical frame that is immediately adjacent to the current frame, wherein the first historical frame has a corresponding first historical metal fence line segment; updating the current metal fence line segment according to the first historical metal fence line segment to obtain an updated metal fence line segment corresponding to the current frame; The updating of the current metal fence line segment according to the first historical metal fence line segment to obtain the updated metal fence line segment corresponding to the current frame includes: Obtaining a historical slope of the first historical metal fence line segment relative to the driving direction and a current slope of the current metal fence line segment relative to the driving direction; If the difference between the historical slope and the current slope is less than a predetermined slope difference threshold, the current metal fence line segment is spliced with the first historical metal fence line segment to obtain the updated metal fence line segment.

2. The method for obtaining a map barrier according to claim 1, characterized in that: The updating of the current metal fence line segment according to the first historical metal fence line segment to obtain the updated metal fence line segment corresponding to the current frame further includes: Determining an overlapping length of the first historical metal fence segment and the current metal fence segment according to positions of the first historical metal fence segment and the current metal fence segment; If the overlap length is greater than a predetermined overlap length threshold, the current metal fence line segment is spliced with the first historical metal fence line segment to obtain the updated metal fence line segment.

3. The method for obtaining a map fence according to claim 2, characterized in that: The step of splicing the current metal fence line segment with the first historical metal fence line segment to obtain the updated metal fence line segment includes: The overlapping part of the current metal fence line segment and the first historical metal fence line segment is superimposed to obtain a superimposed metal fence line segment; the non-overlapping part of the current metal fence line segment, the non-overlapping part of the first historical metal fence line segment and the superimposed metal fence line segment are spliced to obtain the updated metal fence line segment.

4. A map fence acquisition system, characterized in that: include: A receiving module, configured to receive multiple frames of echo point cloud data collected continuously in real time by the millimeter-wave radar, wherein the echo point cloud includes a metal fence echo point cloud and an obstacle echo point cloud, and the obstacle echo point cloud is filtered out of the echo point cloud; an echo point cloud determination module, configured to use the echo point cloud data in the current frame to determine an echo point cloud in the current frame whose density of echo point clouds is greater than a predetermined threshold and which is distributed in a straight line; a fitting calculation module, configured to fit the echo point cloud into a corresponding fitting line segment, calculate the length of the fitting line segment, and obtain the proportion of the echo point cloud corresponding to the fitting line segment in all echo point clouds in the current frame; A metal fence line segment determination module is configured to determine the fitting line segment corresponding to the current frame as the current metal fence line segment if the length of the fitting line segment is greater than a preset length threshold, and the number of occurrences of the echo point cloud data corresponding to the fitting line segment in the multi-frame echo point cloud data collected in real time is greater than a first threshold, and if the proportion is greater than a second threshold, wherein the second threshold reflects the proportion of the echo point cloud corresponding to the fitting line segment in each frame of the multi-frame echo point cloud that overlaps in the multi-frame echo point cloud. After determining the fitting line segment as the metal fence line segment corresponding to the current frame, the method further includes: Finding a first historical frame that is immediately adjacent to the current frame, wherein the first historical frame has a corresponding first historical metal fence line segment; updating the current metal fence line segment according to the first historical metal fence line segment to obtain an updated metal fence line segment corresponding to the current frame; The updating of the current metal fence line segment according to the first historical metal fence line segment to obtain the updated metal fence line segment corresponding to the current frame includes: Obtaining a historical slope of the first historical metal fence line segment relative to the driving direction and a current slope of the current metal fence line segment relative to the driving direction; If the difference between the historical slope and the current slope is less than a predetermined slope difference threshold, the current metal fence line segment is spliced with the first historical metal fence line segment to obtain the updated metal fence line segment.

5. A map, characterized in that: include: A metal fence line segment obtained according to the method for obtaining a map fence according to any one of claims 1 to 3.

6. A computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are operated to execute the method for obtaining a map barrier according to any one of claims 1 to 3.

7. A computer device comprising a processor and a memory, wherein the memory stores computer instructions, wherein: The processor operates the computer instructions to execute the method for obtaining a map barrier according to any one of claims 1 to 3.

Images