A method and device for matching traffic lights in an autonomous driving scenario

Abstract

The present invention discloses a method and device for matching traffic lights in an autonomous driving scenario, and relates to the field of autonomous driving technology. The method comprises: calculating the offset vectors of the projection points relative to each detection frame based on the detection frames of multiple traffic lights in an image and the projection points of multiple traffic lights in a high-precision map; generating multiple matching schemes based on the offset vectors of each projection point; determining the connection lines between the projection points and the detection frames in the same matching combination; calculating the image topological relationship index based on the number of intersections between the connection lines of different matching combinations; calculating the outlier index based on the length difference between the connection lines of different matching combinations; counting the number of traffic lights detected; calculating the matching number index based on the number of traffic lights detected; and screening multiple matching schemes using the evaluation indexes in sequence according to the priority of the evaluation indexes to obtain a target matching scheme. This embodiment can accurately determine the detection frames corresponding to the traffic lights in the high-precision map.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method and device for matching traffic lights in an autonomous driving scenario. Background Art

[0002] In the field of autonomous driving, traffic light detection is crucial for vehicle operation. Currently, high-precision maps are typically used to determine the three-dimensional locations of traffic lights in the lane of the current driving direction. After coordinate conversion, the projection points of these traffic lights in the two-dimensional image captured by the camera are obtained. Based on these projection points, traffic lights are detected and classified to obtain traffic light status information.

[0003] However, in actual application scenarios, the detection process may be affected by factors such as lighting changes, obstructed traffic lights, traffic light failures, and temporary traffic light displacement. This can lead to missed detections and false detections, resulting in discrepancies between the detection information obtained from the 2D image and the traffic light information in the HD map, such as differences in number or location. Therefore, it is necessary to match the traffic lights in the HD map with the detection box to determine the status of the traffic lights. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a method and device for matching traffic lights in an autonomous driving scenario, which can accurately determine the detection frame corresponding to the traffic lights in a high-precision map.

[0005] In a first aspect, an embodiment of the present invention provides a method for matching traffic lights in an autonomous driving scenario, comprising:

[0006] Calculate, based on the detection frames of multiple traffic lights in the image and the projection points of multiple traffic lights in the high-precision map in the image, the offset vectors of the projection points relative to each detection frame;

[0007] Generating a plurality of matching schemes according to the offset vectors of the projection points; wherein any of the matching schemes includes: a plurality of matching combinations consisting of projection points and detection frames;

[0008] For each matching scheme: determining the lines connecting the projection points and the detection frame in the same matching combination; calculating the image topological relationship index based on the number of intersections between the lines of different matching combinations; calculating the outlier index based on the length difference between the lines of different matching combinations; counting the number of traffic lights detected; and calculating the matching number index based on the number of traffic lights detected;

[0009] According to the preset evaluation index priority, multiple matching solutions are screened using the evaluation indexes in turn to obtain the target matching solution;

[0010] Among them, the evaluation indicators include: the image topology relationship indicator, the outlier indicator and the matching quantity indicator; the image topology relationship indicator is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier indicator is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity indicator is used to evaluate the matching scheme from the quantity dimension of the traffic lights.

[0011] In a second aspect, an embodiment of the present invention provides a traffic light matching device in an autonomous driving scenario, comprising:

[0012] A generation module is configured to calculate, based on detection frames of multiple traffic lights in an image and projection points of multiple traffic lights in the image in a high-precision map, offset vectors of the projection points relative to each detection frame; and generate multiple matching schemes based on the offset vectors of each projection point; wherein any of the matching schemes includes: multiple matching combinations consisting of projection points and detection frames;

[0013] a calculation module configured to, for each of the matching schemes, determine the lines connecting the projection points and the detection frame in the same matching combination; calculate an image topological relationship index based on the number of intersections between the lines of different matching combinations; calculate an outlier index based on the length differences between the lines of different matching combinations; count the number of traffic lights detected; and calculate a matching number index based on the number of traffic lights detected;

[0014] A screening module is configured to screen multiple matching solutions using the evaluation indicators in sequence according to the preset evaluation indicator priorities to obtain a target matching solution;

[0015] Among them, the evaluation indicators include: the image topology relationship indicator, the outlier indicator and the matching quantity indicator; the image topology relationship indicator is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier indicator is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity indicator is used to evaluate the matching scheme from the quantity dimension of the traffic lights.

[0016] One embodiment of the above invention has the following advantages or beneficial effects: by calculating the offset vector, the projection point error caused by the conversion from the world coordinate system to the two-dimensional coordinate system of the camera can be corrected, so that the matching distance is more accurate. The image topology relationship index can solve the matching difficulties caused by the proximity of the detection frame and the projection point. The outlier index takes into account the influence of the detection frame obtained by misdetection. The matching quantity index takes into account the degree of matching between the detection frame and the projection point. The above three indicators are used to screen the matching solutions according to the set evaluation indicator priority. By evaluating the matching solutions from three dimensions, the final target matching solution can be quickly and accurately screened.

[0017] The further effects of the above-mentioned non-conventional optional manner will be described below in conjunction with specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are provided for a better understanding of the present invention and are not intended to limit the present invention.

[0019] Figure 1 This is a flowchart of a method for matching traffic lights in an autonomous driving scenario provided by an embodiment of the present invention;

[0020] Figure 2 An embodiment of the present invention provides a combination of two matches that do not have an intersection.

[0021] Figure 3 An embodiment of the present invention provides a combination of two matches that do not have an intersection.

[0022] Figure 4 An embodiment of the present invention provides a height group;

[0023] Figure 5 is another height group provided by one embodiment of the present invention;

[0024] Figure 6 Two plane groups are provided in one embodiment of the present invention;

[0025] Figure 7 This is a schematic diagram of a traffic light projection point provided by an embodiment of the present invention;

[0026] Figure 8 This is a schematic diagram of a traffic light matching device in an autonomous driving scenario provided by an embodiment of the present invention;

[0027] Figure 9 It is a schematic diagram of the structure of a computer system of a terminal device or a server suitable for implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following description of exemplary embodiments of the present invention is made in conjunction with the accompanying drawings, in which various details of the embodiments of the present invention are included to facilitate understanding. These details should be considered as merely exemplary. Therefore, it should be appreciated by those skilled in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0029] like Figure 1 As shown, an embodiment of the present invention provides a method for matching traffic lights in an autonomous driving scenario, including:

[0030] Step 101: Based on the detection frames of multiple traffic lights in the image and the projection points of multiple traffic lights in the high-precision map in the image, calculate the offset vector of the projection point relative to each detection frame.

[0031] Although false detection or missed detection may occur in actual application scenarios, considering that the number of detection boxes and projection points should be consistent when the detection is accurate, in order to ensure the accuracy of the matching results, the number of detection boxes and projection points in the embodiment of the present invention is the same.

[0032] Step 102: Generate multiple matching schemes based on the offset vectors of each projection point; wherein any matching scheme includes: multiple matching combinations consisting of projection points and detection boxes.

[0033] For the same projection point, the offset vectors relative to different detection boxes correspond to different matching schemes.

[0034] Step 103: For each matching scheme: determine the connection line between the projection point and the detection frame in the same matching combination; calculate the image topology relationship index based on the number of intersections between the connection lines of different matching combinations; calculate the outlier index based on the length difference between the connection lines of different matching combinations; count the number of traffic lights detected; and calculate the matching quantity index based on the number of traffic lights detected.

[0035] Step 104: According to the preset evaluation indicator priority, the evaluation indicators are used in sequence to screen multiple matching solutions to obtain a target matching solution.

[0036] Among them, the evaluation indicators include: image topology relationship index, outlier index and matching quantity index; the image topology relationship index is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier index is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity index is used to evaluate the matching scheme from the number dimension of the traffic lights.

[0037] For example, there are 16 matching solutions. 10 are selected based on the image topology relationship index, 6 are selected based on the outlier index, and 1 is selected from 6 based on the number of matches index as the target matching solution. The priority of the evaluation indicators can be adjusted according to the needs of the actual scenario.

[0038] By calculating the offset vector, we can correct the projection point errors caused by converting the world coordinate system to the camera's two-dimensional coordinate system, making the matching distance more accurate. The image topology relationship index can address the matching difficulties caused by the close proximity of the detection frame and the projection point. The outlier index considers the impact of misdetected detection frames, and the matching quantity index considers the degree of matching between the detection frame and the projection points. These three indicators are used to screen matching solutions according to the set evaluation priority. By evaluating matching solutions from three dimensions, we can quickly and accurately select the final target matching solution.

[0039] In one embodiment of the present invention, based on the detection frames of multiple traffic lights in an image and the projection points of multiple traffic lights in the image in a high-precision map, calculating the offset vectors of the projection points relative to each detection frame includes:

[0040] For the projection points of each traffic light in the HD map, the following operations are performed: Based on the coordinates of the projection point and the coordinates of the center of each detection frame, the offset vector of the projection point relative to each detection frame is calculated.

[0041] For example, the detection boxes are A, B and C, and the projection points are F, G and H. For the projection point F, the offset vectors of F relative to A, B and C are calculated respectively. The coordinates of F are (f1, f2), the coordinates of A are (a1, a2), and the offset vector is

[0042] The embodiment of the present invention can accurately measure the position deviation between the detection frame and the projection point through the offset vector, thereby improving the accuracy of the matching result.

[0043] In one embodiment of the present invention, multiple matching schemes are generated based on the offset vectors of each projection point, including:

[0044] Calculate the offset distance of the projection point relative to each detection frame based on the offset vector of the projection point relative to each detection frame;

[0045] Bipartite graph matching is performed according to the offset distance of each projection point to obtain multiple matching solutions.

[0046] Using the above example, The offset distance is

[0047] Preferably, the bipartite graph matching may adopt the Hungarian algorithm, etc. The embodiment of the present invention adopts bipartite graph matching to match the projection point with the detection frame, thereby improving the accuracy of the matching result.

[0048] In one embodiment of the present invention, the image topology relationship index is calculated based on the number of intersections between lines of different matching combinations, including:

[0049] Count the number of intersections between the lines of each matching combination in the matching solution;

[0050] Based on the z-value of the traffic light to which the projection point belongs in the HD map, the matching combinations in the matching scheme are divided into multiple height groups, and the number of intersections between the connecting lines of different matching combinations in each height group is counted.

[0051] Based on the x and y values ​​of the traffic light to which the projection point belongs in the HD map, each height group is divided into multiple plane groups, and the number of intersections between the lines of different matching combinations in each plane group is counted.

[0052] The image topological relationship index is calculated based on the number of intersections between the lines of each matching combination in the matching scheme, the number of intersections between the lines of different matching combinations in each height group, and the number of intersections between the lines of different matching combinations in each plane group.

[0053] The image topology relationship index can be the sum of the number of intersections between the lines of each matching combination in the matching scheme, the number of intersections between the lines of different matching combinations in each height group, and the number of intersections between the lines of different matching combinations in each plane group. The matching scheme with the smaller number of intersections is preferred.

[0054] Traffic lights are installed within a specified height range, and the vertical distribution does not vary much, while the horizontal distribution of traffic lights is related to the size of the intersection. Traffic lights in a high-precision map are arranged from left to right, so the order of the detection frames to be matched should also be from left to right. It is unreasonable to match the traffic light on the left side of the map to the detection frame on the right side. Based on this, the embodiment of the present invention counts the number of intersections between the lines of each matching combination in the matching solution, such as Figure 2 As shown, the detection boxes and projection points in the two matching combinations are connected by lines, and there is no intersection between the two lines, such as Figure 3 As shown, the two lines intersect at a point.

[0055] In some scenarios, the lines connecting two matching combinations intersect, but the installation heights of the two corresponding traffic lights are different. In essence, the two matching combinations are correct. To prevent this from happening, the embodiment of the present invention further divides the matching combinations in the matching scheme into multiple height groups based on the z value of the traffic light to which the projection point belongs in the high-precision map. For example, Figure 4 and Figure 5 There are two height groups.

[0056] In some scenarios, traffic lights at the same height but different locations will overlap in the image, resulting in intersections between the lines, but the matching combination is actually correct. Based on this, the embodiment of the present invention divides each height group into multiple plane groups based on the x and y values ​​of the traffic lights to which the projection point belongs in the high-precision map, such as Figure 6As shown, it is divided into two plane groups: far and near.

[0057] like Figure 7 As shown in the figure, there are 5 traffic lights in front of the vehicle. After being projected into the image, they are divided into three groups for processing.

[0058] The embodiment of the present invention can improve the accuracy of the image topology relationship index by grouping and counting the number of intersections, so that it can accurately reflect the matching relationship between the detection frame and the projection point.

[0059] In one embodiment of the present invention, calculating an outlier index based on the length differences between lines of different matching combinations includes:

[0060] Determining the number of target matching combinations in the matching scheme based on the length of the connection line of the matching combinations; wherein the ratio of the length of the connection line of the target matching combination to the length of the connection line of each other matching combination is within a preset ratio range;

[0061] Calculates an outlier metric based on the number of target-matching combinations in a matching scenario.

[0062] Taking into account the possibility of missed detection and false detection, the detection frame may match the wrong projection point, resulting in a difference in length from other correct lines, which is an outlier. The fewer outliers in the matching scheme, the better. In an embodiment of the present invention, the length of the line of the target matching combination is an outlier. For example, if the ratio of the length of the line of matching combination 1 to the length of the line of other matching combinations is greater than 1.3, then matching combination 1 is the target matching combination. The outlier indicator can be the number of target matching combinations in the matching scheme, and matching schemes with fewer target matching combinations are preferred.

[0063] In one embodiment of the present invention, calculating a matching quantity index based on the number of detected traffic lights includes:

[0064] Determine the number of foreground traffic lights in the traffic lights corresponding to the matching solution; wherein the angle between the line connecting the vehicle and the foreground traffic lights and the orientation of the foreground traffic lights is less than a preset angle threshold, the distance between the foreground traffic lights and the vehicle is no greater than the distance between the vehicle and the target traffic light, and the target traffic light is the traffic light located in front of the lane where the vehicle is located and closest to the vehicle;

[0065] Calculate the matching quantity index based on the number of foreground traffic lights.

[0066] Through the embodiments of the present invention, priority can be given to traffic lights that require attention in the lane where the vehicle is located, thereby ensuring the driving safety of the vehicle.

[0067] In one embodiment of the present invention, calculating the matching quantity index based on the number of foreground traffic lights includes:

[0068] Based on the detection frame, count the sum of the number of red lights, yellow lights, and green lights in the image recognition results;

[0069] The matching quantity index is calculated based on the number of red lights, yellow lights, and green lights and the number of foreground traffic lights.

[0070] The match quantity metric can be the sum of the number of foreground traffic lights, red lights, yellow lights, and green lights. The difference between the sum of the number of projection points and the sum of the number of red lights, yellow lights, and green lights includes traffic lights recognized as black and can also include traffic lights that were not identified as a match. Priority is given to matching solutions with a greater sum of the number of foreground traffic lights, red lights, yellow lights, and green lights. Of course, other methods can also be used to calculate the above three metrics, for example, calculating the match quantity metric based solely on the number of foreground traffic lights.

[0071] like Figure 8 As shown, an embodiment of the present invention provides a traffic light matching device in an autonomous driving scenario, comprising:

[0072] The generation module 801 is configured to calculate, based on the detection frames of multiple traffic lights in the image and the projection points of multiple traffic lights in the image in the high-precision map, an offset vector of the projection point relative to each detection frame; and generate multiple matching schemes based on the offset vectors of each projection point; wherein each matching scheme includes: multiple matching combinations consisting of projection points and detection frames;

[0073] The calculation module 802 is configured to: for each matching solution, determine the lines connecting the projection points and the detection box in the same matching combination; calculate the image topological relationship index based on the number of intersections between the lines of different matching combinations; calculate the outlier index based on the length difference between the lines of different matching combinations; count the number of traffic lights detected; and calculate the matching number index based on the number of traffic lights detected;

[0074] The screening module 803 is configured to screen multiple matching solutions using the evaluation indicators in sequence according to the preset evaluation indicator priorities to obtain a target matching solution;

[0075] Among them, the evaluation indicators include: image topology relationship index, outlier index and matching quantity index; the image topology relationship index is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier index is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity index is used to evaluate the matching scheme from the number dimension of the traffic lights.

[0076] In one embodiment of the present invention, the generation module 801 is configured to execute, for each projection point of each traffic light in the image in the high-precision map: calculating the offset vector of the projection point relative to each detection frame based on the coordinates of the projection point and the coordinates of the center of each detection frame.

[0077] In one embodiment of the present invention, the generation module 801 is configured to calculate the offset distance of the projection point relative to each detection frame based on the offset vector of the projection point relative to each detection frame; perform bipartite graph matching based on the offset distance of each projection point to obtain multiple matching schemes.

[0078] In one embodiment of the present invention, the calculation module 802 is configured to count the number of intersections between the lines of each matching combination in the matching scheme; based on the z value of the traffic light to which the projection point belongs in the high-precision map, the matching combinations in the matching scheme are divided into multiple height groups, and the number of intersections between the lines of different matching combinations in each height group is counted respectively; based on the x value and y value of the traffic light to which the projection point belongs in the high-precision map, each height group is divided into multiple plane groups, and the number of intersections between the lines of different matching combinations in each plane group is counted respectively; according to the number of intersections between the lines of each matching combination in the matching scheme, the number of intersections between the lines of different matching combinations in each height group, and the number of intersections between the lines of different matching combinations in each plane group, the image topological relationship index is calculated.

[0079] In one embodiment of the present invention, the calculation module 802 is configured to determine the number of target matching combinations in the matching scheme based on the length of the connection line of the matching combination; wherein the ratio of the length of the connection line of the target matching combination to the length of the connection line of each other matching combination is within a preset ratio range; and calculate the outlier index based on the number of target matching combinations in the matching scheme.

[0080] In one embodiment of the present invention, the calculation module 802 is configured to determine the number of foreground traffic lights in the traffic lights corresponding to the matching scheme; and calculate a matching quantity index based on the number of foreground traffic lights; wherein, the angle between the line connecting the vehicle and the foreground traffic lights and the orientation of the foreground traffic lights is less than a preset angle threshold, the distance between the foreground traffic lights and the vehicle is not greater than the distance between the vehicle and the target traffic light, and the target traffic light is the traffic light located in front of the lane where the vehicle is located and closest to the vehicle.

[0081] In one embodiment of the present invention, the calculation module 802 is configured to count the sum of the number of red lights, yellow lights and green lights in the image recognition results based on the detection frame; and calculate the matching quantity index based on the sum of the number of red lights, yellow lights and green lights and the number of foreground traffic lights.

[0082] An embodiment of the present invention provides an electronic device, including:

[0083] one or more processors;

[0084] a storage device for storing one or more programs,

[0085] When one or more programs are executed by one or more processors, the one or more processors implement the method of any of the above embodiments.

[0086] An embodiment of the present invention provides a computer-readable medium having a computer program stored thereon. When the program is executed by a processor, the method of any of the above embodiments is implemented.

[0087] Reference below Figure 9 , which shows a schematic structural diagram of a computer system 900 of a terminal device suitable for implementing an embodiment of the present invention. Figure 9 The terminal device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0088] like Figure 9 As shown, the computer system 900 includes a central processing unit (CPU) 901, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 902 or a program loaded from a storage unit 908 into a random access memory (RAM) 903. Various programs and data required for the operation of the system 900 are also stored in the RAM 903. The CPU 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to the bus 904.

[0089] The following components are connected to the I / O interface 905: an input section 906 including a keyboard, a mouse, and the like; an output section 907 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 908 including a hard disk and the like; and a communication section 909 including a network interface card such as a LAN card or a modem. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to the I / O interface 905 as needed. A removable medium 911, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 910 as needed, so that computer programs read therefrom can be installed into the storage section 908 as needed.

[0090] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 909, and / or installed from a removable medium 911. When the computer program is executed by the central processing unit (CPU) 901, the above-mentioned functions defined in the system of the present invention are performed.

[0091] It should be noted that the computer-readable medium described in the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media can include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present invention, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wireline, optical fiber cable, RF, or any suitable combination thereof.

[0092] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0093] The modules involved in the embodiments of the present invention may be implemented in software or hardware. The modules described may also be provided in a processor. For example, they may be described as follows: a processor includes a sending module, an acquisition module, a determination module, and a first processing module. The names of these modules do not, in some cases, limit the modules themselves. For example, the sending module may also be described as a "module for sending a picture acquisition request to the connected server."

[0094] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A traffic light matching method in an autonomous driving scenario, characterized in that: include: Calculate, based on the detection frames of multiple traffic lights in the image and the projection points of multiple traffic lights in the high-precision map in the image, the offset vectors of the projection points relative to each detection frame; Generating a plurality of matching schemes according to the offset vectors of the projection points; wherein any of the matching schemes includes: a plurality of matching combinations consisting of projection points and detection frames; For each matching scheme: determining the lines connecting the projection points and the detection frame in the same matching combination; calculating the image topological relationship index based on the number of intersections between the lines of different matching combinations; calculating the outlier index based on the length difference between the lines of different matching combinations; counting the number of traffic lights detected; and calculating the matching number index based on the number of traffic lights detected; According to the preset evaluation index priority, multiple matching solutions are screened using the evaluation indexes in turn to obtain the target matching solution; Among them, the evaluation indicators include: the image topology relationship indicator, the outlier indicator and the matching quantity indicator; the image topology relationship indicator is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier indicator is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity indicator is used to evaluate the matching scheme from the quantity dimension of the traffic lights.

2. The method according to claim 1, wherein Calculating, based on the detection frames of multiple traffic lights in the image and the projection points of multiple traffic lights in the high-precision map in the image, the offset vectors of the projection points relative to the respective detection frames include: For the projection points of each traffic light in the high-precision map in the image, the following steps are performed: based on the coordinates of the projection point and the coordinates of the center of each detection frame, the offset vector of the projection point relative to each detection frame is calculated.

3. The method according to claim 1, wherein According to the offset vectors of the projection points, multiple matching schemes are generated, including: Calculating the offset distance of the projection point relative to each detection frame according to the offset vector of the projection point relative to each detection frame; Bipartite graph matching is performed according to the offset distance of each projection point to obtain multiple matching solutions.

4. The method according to claim 1, wherein Based on the number of intersections between lines of different matching combinations, the image topology relationship indicators are calculated, including: Counting the number of intersections between lines connecting each matching combination in the matching scheme; Based on the z-value of the traffic light to which the projection point belongs in the high-precision map, the matching combinations in the matching scheme are divided into multiple height groups, and the number of intersections between the connecting lines of different matching combinations in each height group is counted respectively; Based on the x-value and y-value of the traffic light to which the projection point belongs in the high-precision map, each height group is divided into multiple plane groups, and the number of intersections between the lines of different matching combinations in each plane group is counted; The image topological relationship index is calculated based on the number of intersections between the lines of each matching combination in the matching scheme, the number of intersections between the lines of different matching combinations in each height group, and the number of intersections between the lines of different matching combinations in each plane group.

5. The method according to claim 1, wherein Based on the length differences between the lines of different matching combinations, outlier indicators are calculated, including: Determining the number of target matching combinations in the matching scheme based on the length of the connection line of the matching combinations; wherein the ratio of the length of the connection line of the target matching combination to the length of the connection line of each other matching combination is within a preset ratio range; The outlier indicator is calculated according to the number of target matching combinations in the matching scheme.

6. The method according to claim 1, wherein According to the number of detected traffic lights, the matching quantity index is calculated, including: Determining the number of foreground traffic lights among the traffic lights corresponding to the matching solution; wherein the angle between a line connecting the vehicle and the foreground traffic lights and the orientation of the foreground traffic lights is less than a preset angle threshold, and the distance between the foreground traffic lights and the vehicle is no greater than the distance between the vehicle and a target traffic light, wherein the target traffic light is the traffic light located in front of the lane where the vehicle is located and closest to the vehicle; The matching quantity index is calculated according to the number of the foreground traffic lights.

7. The method according to claim 6, wherein Calculating the matching quantity index according to the number of the foreground traffic lights includes: Based on the detection frame, counting the sum of the number of red lights, yellow lights, and green lights in the image recognition result; The matching quantity index is calculated based on the number of foreground traffic lights and the sum of the numbers of red lights, yellow lights and green lights.

8. A traffic light matching device in an autonomous driving scenario, characterized in that: include: a generation module configured to calculate, based on the detection frames of the plurality of traffic lights in the image and the projection points of the plurality of traffic lights in the high-precision map in the image, an offset vector of the projection points relative to each detection frame; Generating a plurality of matching schemes according to the offset vectors of the projection points; wherein any of the matching schemes includes: a plurality of matching combinations consisting of projection points and detection frames; a calculation module configured to, for each of the matching schemes, determine the lines connecting the projection points and the detection frame in the same matching combination; calculate an image topological relationship index based on the number of intersections between the lines of different matching combinations; calculate an outlier index based on the length differences between the lines of different matching combinations; count the number of traffic lights detected; and calculate a matching number index based on the number of traffic lights detected; A screening module is configured to screen multiple matching solutions using the evaluation indicators in sequence according to the preset evaluation indicator priorities to obtain a target matching solution; Among them, the evaluation indicators include: the image topology relationship indicator, the outlier indicator and the matching quantity indicator; the image topology relationship indicator is used to evaluate the matching scheme from the coordinate position dimension of the traffic light; the outlier indicator is used to evaluate the matching scheme from the distribution dimension of the traffic light; the matching quantity indicator is used to evaluate the matching scheme from the quantity dimension of the traffic lights.

9. An electronic device, characterized in that: include: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 7.

10. A computer-readable medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images