Mobile target-based mapping anchor calibration method, system and electronic device

Abstract

The application discloses a mapping anchor point calibration method and system based on a mobile target and electronic equipment. Image data of an intersection area collected by a two-dimensional sensing device and point cloud data of the intersection area collected by a three-dimensional sensing device are obtained. The mobile target in the image data and the point cloud data and the position coordinates of the mobile target are determined according to the image data, the point cloud data and a trained target detection model. Then, a user's selection operation on the mobile target in the image data and / or the point cloud data is obtained, and the mobile target in the image data and the point cloud data is matched based on the selection operation to obtain a matched mobile target. Finally, the mapping anchor point is determined according to the position coordinates of the matched mobile target. In this way, the mobile target in the image data and the point cloud data is determined by automatic detection, manual frame selection is not required, and the cumbersome process of manually finding points is avoided, so that the position of the mapping anchor point is more accurate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine vision, and in particular to a mapping anchor point calibration method and system based on a moving target and an electronic device. BACKGROUND

[0002] In the application process of machine vision technology, sometimes two-dimensional and three-dimensional images of a target scene need to be collected at the same time, and the information of the two is fused. In the process of projecting the information of the two-dimensional image to the three-dimensional coordinate system, because of the natural lack of the dimension of the two-dimensional image information, it is necessary to supplement the mapping relationship of the two-dimensional coordinates to the three-dimensional coordinates. When the three-dimensional scene is approximately a plane, the mapping can be completed through the internal and external parameters of the camera, but when the three-dimensional scene is relatively complex and cannot be approximated as a plane, it is necessary to supplement a plurality of anchor points of two-dimensional coordinates to three-dimensional coordinates to assist the mapping. The more the number of anchor points, the better the effect of coordinate mapping.

[0003] In the prior art, the calibration of the mapping anchor points from two-dimensional to three-dimensional images is generally completed by fixed targets in the scene such as street lamps, flowerpots, zebra crossings, etc. The coordinates of the anchor point objects in the two-dimensional image and the three-dimensional point cloud are determined by manual framing or point selection. However, this anchor point calibration method is limited by the number of fixed targets with distinctive features in the scene. If the number of fixed targets is small, the generated mapping dictionary is relatively sparse, and the mapping accuracy is not high. Moreover, since the coordinates of the fixed targets in the image do not change, even with a large amount of image data, more mapping anchor points cannot be obtained. At the same time, the manual acquisition of the mapping anchor point coordinates often varies from person to person and is not accurate enough. SUMMARY

[0004] In view of the above problems, the present application is proposed to provide a mapping anchor point calibration method, system and electronic device based on a moving target, which can automatically detect the moving target, without manual framing, and eliminate the tedious process of manual point finding, so that the position of the mapping anchor point is more accurate.

[0005] According to a first aspect of the present application, a mapping anchor point calibration method based on a moving target is provided, comprising:

[0006] obtaining image data of a road intersection region collected by a two-dimensional perception device, and point cloud data of the road intersection region collected by a three-dimensional perception device;

[0007] determining the moving target in the image data, the position coordinates of the moving target, and the moving target in the point cloud data, and the position coordinates of the moving target according to the image data, the point cloud data, and the trained target detection model;

[0008] Obtaining a matching operation of a user on a moving target in image data and / or point cloud data, and matching the moving target in the image data and the point cloud data based on the matching operation to obtain a matched moving target;

[0009] According to a position coordinate of the matched moving target, a mapping anchor point is determined.

[0010] Optionally, obtaining the matching operation of the user on the moving target in the image data and / or the point cloud data, and matching the moving target in the image data and the point cloud data based on the matching operation includes:

[0011] Obtaining a moving target selected by the user in the image data and a first order of selecting the moving target;

[0012] Obtaining a moving target selected by the user in the point cloud data and a second order of selecting the moving target;

[0013] According to the first order and the second order, the moving target selected by the user in the image data and the moving target selected by the user in the point cloud data are matched.

[0014] Optionally, obtaining the matching operation of the user on the moving target in the image data and / or the point cloud data, and matching the moving target in the image data and the point cloud data based on the matching operation includes:

[0015] Obtaining a first matching operation of the user in the image data to obtain a plurality of moving targets;

[0016] Obtaining a second matching operation of the user in the point cloud data corresponding to the first matching operation to obtain a plurality of moving targets;

[0017] According to a position relationship logic between the plurality of moving targets of the first matching operation and the plurality of moving targets of the second matching operation, the matching is performed.

[0018] Optionally, before the matching is performed according to the position relationship logic between the plurality of moving targets of the first matching operation and the plurality of moving targets of the second matching operation, the method further includes:

[0019] In a preset position relationship logic table, all position relationship logics are traversed for the plurality of moving targets of the first matching operation and the plurality of moving targets of the second matching operation;

[0020] A position relationship logic satisfying the plurality of moving targets of the first matching operation and the plurality of moving targets of the second matching operation is determined.

[0021] Optionally, obtaining the matching operation of the user on the moving target in the image data and / or the point cloud data, and matching the moving target in the image data and the point cloud data based on the matching operation includes:

[0022] determining target data from the image data and the point cloud data, and performing a first sorting on moving targets in the target data;

[0023] determining non-target data from the image data and the point cloud data, and obtaining a second sorting of moving targets in the non-target data by a user;

[0024] matching the moving targets in the image data and the point cloud data according to the first sorting and the second sorting.

[0025] Optionally, the mapping anchor point is determined according to the position coordinates of the matched moving targets, comprising:

[0026] determining the center point coordinates of the matched moving targets according to the position coordinates of the matched moving targets;

[0027] determining the mapping anchor point according to the center point coordinates of the matched moving targets.

[0028] Optionally, before determining the moving targets in the image data and the position coordinates of the moving targets, and the moving targets in the point cloud data and the position coordinates of the moving targets according to the image data, the point cloud data and the trained target detection model, the method further comprises:

[0029] taking the image data and the point cloud data with the moving target label as sample data;

[0030] training the initial target detection model through the sample data to obtain the trained target detection model.

[0031] Optionally, the method further comprises:

[0032] reacquiring the image data of the intersection region collected by the two-dimensional perception device and the point cloud data of the intersection region collected by the three-dimensional perception device at different time instants to obtain new mapping anchor points.

[0033] According to a second aspect of the present application, a mapping anchor point calibration system based on moving targets is provided, comprising:

[0034] a two-dimensional perception device configured to collect image data of an intersection region;

[0035] a three-dimensional perception device configured to collect point cloud data of the intersection region;

[0036] a first processing module configured to determine moving targets in the image data and position coordinates of the moving targets, and moving targets in the point cloud data and position coordinates of the moving targets according to the image data, the point cloud data and a trained target detection model;

[0037] The second processing module is configured to obtain a selection operation of a user on the moving target in the image data and / or the point cloud data, and match the moving target in the image data and the point cloud data based on the selection operation, to obtain a matched moving target.

[0038] The anchor point obtaining module is configured to determine a mapping anchor point according to the position coordinates of the matched moving target.

[0039] According to a third aspect of the present application, an electronic device is provided, which comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the above-mentioned mapping anchor point calibration method based on a moving target when executing the computer program.

[0040] According to a fourth aspect of the present application, a computer readable storage medium is provided, which stores a computer program, and the program is executed by a processor to implement the above-mentioned mapping anchor point calibration method based on a moving target.

[0041] The above-mentioned one or more technical solutions in the embodiments of the present application have at least the following technical effects:

[0042] The mapping anchor point calibration method, system and electronic device based on a moving target provided by the embodiments of the present application, by acquiring image data of a road intersection region collected by a two-dimensional perception device and point cloud data of the road intersection region collected by a three-dimensional perception device, then determining a moving target in the image data, position coordinates of the moving target, a moving target in the point cloud data, and position coordinates of the moving target according to the image data, the point cloud data and a trained target detection model, then obtaining a selection operation of a user on the moving target in the image data and / or the point cloud data, and matching the moving target in the image data and the point cloud data based on the selection operation, to obtain a matched moving target, and finally determining a mapping anchor point according to the position coordinates of the matched moving target. In this way, the moving target in the image data and the point cloud data is determined by automatic detection, without manual frame selection, and the tedious process of manual point finding is avoided, so that the position of the mapping anchor point is more accurate. Compared with fixed targets such as street lamps and flowerpots, moving targets are more widely distributed in a scene, cover more points, and their position coverage area is also a region of interest for data analysis.

[0043] The above description is only a summary of the technical solutions of the present application, in order to more clearly understand the technical means of the present application, the specific embodiments of the present application can be implemented according to the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0044] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments with reference made to the accompanying drawings. The drawings provided herein are for illustrative purposes only and, as such, are not intended to limit the present application. Further, in the drawings, like reference numerals are used to designate like parts throughout the several views.

[0045] In the drawings:

[0046] Figure 1 A flow chart of a mobile target based mapping anchor point calibration method in an embodiment of the present application is shown.

[0047] Figure 2 A schematic diagram of mobile target matching in image data and point cloud data in an embodiment of the present application is shown.

[0048] Figure 3 A block schematic diagram of a mobile target based mapping anchor point calibration system in an embodiment of the present application is shown.

[0049] Figure 4 A schematic diagram of an electronic device in an embodiment of the present application is shown.

[0050] Icon:

[0051] 100 - electronic device; 10 - mapping anchor point calibration apparatus; 20 - memory; 30 - processor; 40 - communication unit. DETAILED DESCRIPTION

[0052] So that the manner in which the above recited features and advantages of the present application are attained and utilized can be understood in detail, a brief description of the application summarized above will be rendered by reference to the detailed description of the preferred embodiments which are illustrated in the following drawings. The application will be described with reference to the attached drawings, wherein like reference numerals represent identical or corresponding elements throughout the several views.

[0053] Accordingly, the detailed description of the embodiments of the application as provided in the following is not intended to limit the scope of the application, but merely to set forth a preferred embodiment of the application. Set forth herein is a description of the best attempts at a complete disclosure of the application in such full, clear, concise, and exact terms as to enable one of ordinary skill in the art to which the application pertains, to make and use any needed

[0054] It is to be noted that like reference numerals and letters refer to like items in the drawings and, as such, definitions of those items need not be repeated with respect to each of the following drawings.

[0055] In the description of the present application, it should also be noted that, unless specifically defined and limited otherwise, the terms "set", "install", "connect", "connect" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium, or it can be connected inside two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0056] The present embodiment provides an application scenario, which includes an intersection, which can be a three-way intersection, a crossroad, etc. The present embodiment takes a crossroad as an example for description. A two-dimensional perception device, such as a camera, is installed at the crossroad, and a three-dimensional perception device, such as a laser radar, is also installed. It should be noted that the central intersection of the crossroad is the intersection area, the effective shooting area of the two-dimensional perception device covers the intersection area, and the effective shooting area of the three-dimensional perception device also covers the intersection area. There are characteristic moving objects, such as vehicles, non-motor vehicles, and pedestrians, etc. in the crossroad. In the present embodiment, vehicles are taken as moving targets, and through two-dimensional collection and three-dimensional collection of the moving targets, image data and point cloud data are obtained, and then the moving targets and the position coordinates of the moving targets in the image data and the point cloud data are detected. The same moving targets are selected in the image data and the point cloud data for matching, and the mapping anchor points are obtained accordingly. There are usually multiple moving targets in the image data and the point cloud data, and accordingly multiple groups of mapping anchor points can be obtained. According to these mapping anchor points, the image data and the point cloud data can be better fused, so as to monitor the vehicles running on the road. For example, the license plate, the running speed, the running track, the road congestion situation, etc. of the vehicle are detected. According to these monitoring information, the illegal situation of a single vehicle can be obtained; and traffic optimization can also be performed, such as changing the length of the red light to relieve traffic congestion, etc.

[0057] Based on the above application scenario, combined with Figure 1 The present embodiment provides a mapping anchor point calibration method based on moving targets, which includes steps 101 to 104:

[0058] Step 101: obtaining image data of the intersection area collected by a two-dimensional perception device, and point cloud data of the intersection area collected by a three-dimensional perception device;

[0059] In the present embodiment, the two-dimensional perception device and the three-dimensional perception device are installed at the crossroad. The effective areas shot by the two-dimensional perception device and the three-dimensional perception device are not necessarily completely the same. However, the effective area shot by the two-dimensional perception device covers the intersection area, and similarly, the effective area shot by the three-dimensional perception device covers the intersection area. The two-dimensional perception device can be a camera, and the three-dimensional perception device can be a laser radar.

[0060] The two-dimensional perception device captures the intersection area to obtain image data. Meanwhile, the three-dimensional perception device captures the intersection area to obtain point cloud data.

[0061] Step 102: According to the image data, the point cloud data, and the trained target detection model, the moving target in the image data, the position coordinates of the moving target, and the moving target in the point cloud data, and the position coordinates of the moving target are determined.

[0062] In this embodiment, the image data and the point cloud data generally include a moving target, which can be a vehicle, a pedestrian, etc. The recognition and detection of the moving target can be automatically recognized by the trained target detection model. The position coordinates of the moving target are obtained. For the recognition of the moving target in the image data, the target detection model can realize the detection of the moving target by using a single CNN model through the Yo l o algorithm. For the recognition of the moving target in the point cloud data, the target detection model can use pi l l ar / centerpoi nt / voxe l net.

[0063] In this embodiment, the initial target detection model needs to be trained by sample data. Specifically, the image data and the point cloud data with the moving target label can be used as sample data, and the more sample data, the better. Then the initial target detection model is trained by the sample data, and the model parameters are adjusted during the training, and finally the trained target detection model is obtained. Through the target detection model, the moving target in the data can be automatically recognized quickly and effectively without manual frame selection by the user, so that the detection and recognition of the moving target are more accurate. The moving target is different from the fixed target, and the distribution of the moving target in the scene is more extensive, the coverage points are more, and the coverage area where the position is located is also the area of interest for data analysis.

[0064] Step 103: Obtain the selection operation of the user on the moving target in the image data and / or the point cloud data, and match the moving target in the image data and the point cloud data based on the selection operation to obtain the matched moving target.

[0065] The selection operation refers to an operation of selecting or marking the moving target in the image data or the point cloud data by the user. The user can select one moving target or multiple moving targets. It should be noted that if the user needs to perform the selection operation on the moving target in the image data and the point cloud data respectively, the moving target selected by the user in the image data should be the same as the moving target selected by the user in the point cloud data; if the user needs to perform the selection operation on the moving target in the image data or the point cloud data only once, the moving target selected by the user in the image data or the point cloud data can be the same as or different from the moving target selected in the other data.

[0066] In an embodiment, the image data and the point cloud data are combined. Figure 2 As shown in the figure, the upper half of the figure shows image data of a certain intersection area, and the lower half of the figure shows point cloud data of the same intersection area at the same time. Taking a vehicle as a moving target, a plurality of moving targets are automatically identified in the image data and are marked with a box. A plurality of moving targets are also automatically identified in the point cloud data and are marked with a box. Since the image data and the point cloud data have different shooting ranges, the moving targets identified in the image data and the point cloud data can not be exactly the same. In the example shown in the figure, the number of moving targets identified in the image data is greater than the number of moving targets identified in the point cloud data. Figure 2 In the example shown in the figure, the number of moving targets identified in the image data is greater than the number of moving targets identified in the point cloud data. Therefore, if the user needs to perform the selection operation on the moving target in the image data and the point cloud data respectively, the user should pay attention to the fact that the moving target selected by the user should exist in both the image data and the point cloud data. For example, the user clicks three moving targets in the image data in sequence by point selection, and the number 1, 2 and 3 will appear in sequence on the boxes of the three selected moving targets in the display interface. Then the user clicks three moving targets in the point cloud data in the same order, and the number 1, 2 and 3 will appear in sequence on the boxes of the three selected moving targets in the display interface. Then, the three moving targets in the image data and the three moving targets in the point cloud data are matched according to the selection order. For example, the moving target 1 in the image data is matched with the moving target 1 in the point cloud data, so that the moving target 1 in the image data and the moving target 1 in the point cloud data form a matched moving target.

[0067] Step 104: determining a mapping anchor point according to the position coordinates of the matched moving target.

[0068] After the matched mobile target is obtained, a center point and / or a feature point can be determined according to the position coordinates of the matched mobile target, and the mapping anchor point is determined based on the center point and / or the feature point. Specifically, the center point coordinates of the matched mobile target can be determined according to the position coordinates of the matched mobile target, and then the mapping anchor point is determined based on the center point coordinates of the matched mobile target. Alternatively, the feature point coordinates of the matched mobile target, such as the head / tail corner point and the wheel contact point, can be determined by performing feature point detection in the range of the position coordinates of the matched mobile target, and then the mapping anchor point is determined based on the feature point coordinates of the matched mobile target.

[0069] It should be noted that the matched mobile target refers to a mobile target in the image data and a corresponding mobile target in the point cloud data. The position coordinates of the two mobile targets can be determined, and then the center point of the mobile target can be determined based on the position coordinates of the mobile target. The coordinates of the center point are then obtained, and finally the mapping anchor point is determined based on the center point. In this way, manual detection of the bounding box of the mobile target and determination of the center point are not required, the tedious process of manually finding the point is avoided, and the position of the mapping anchor point is more accurate.

[0070] In summary, the mapping anchor point calibration method based on a mobile target provided by the embodiments of the present specification automatically detects the mobile target in the image data and the point cloud data, does not require manual bounding box selection, and avoids the tedious process of manually finding the point, so that the position of the mapping anchor point is more accurate. Compared with fixed targets such as street lamps and flowerpots, mobile targets are more widely distributed in the scene, cover more points, and their position coverage area is also the area of interest for data analysis.

[0071] In this embodiment, in order to facilitate understanding and implementation by those skilled in the art, three implementation manners are given for the matching operation.

[0072] In an alternative implementation manner, the mobile target selected by the user in the image data and the first order of selecting the mobile target can be obtained first. Then, the mobile target selected by the user in the point cloud data and the second order of selecting the mobile target are obtained. The mobile target selected by the user in the image data and the mobile target selected by the user in the point cloud data are matched according to the first order and the second order.

[0073] When the user selects a point in the image data, the number of selected mobile targets can be one or more. In addition, the mobile target selected by the user in the image data and the mobile target selected by the user in the point cloud data should be the same mobile target. The mobile target selected by the user in the image data and the mobile target selected by the user in the point cloud data are matched according to the first order and the second order.

[0074] In another optional implementation, the first selection operation of the user in the image data is obtained to obtain a plurality of moving targets; then the second selection operation of the user in the point cloud data corresponding to the first selection operation is obtained to obtain a plurality of moving targets; and the position relationship logic between the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation is matched.

[0075] In the first selection operation, the user frames a plurality of moving targets, and then frames the same moving targets in the second selection operation. After the two selection operations are completed, the position relationship logic between the framed moving targets is determined. Specifically, the position relationship logic between the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation is determined.

[0076] In the preset position relationship logic table, all position relationship logics are traversed for the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation.

[0077] The position relationship logic satisfying the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation is determined.

[0078] In the preset logic relationship table, a limited number of position relationship logics are set, such as from left to right, from top to bottom, from top right to bottom left, and the like. Then, all position relationship logics in the preset position relationship logic table are sequentially tried for the moving targets framed twice, and finally a position relationship logic is determined, i.e., the position relationship logic satisfying the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation, and the moving targets of the two selection operations are matched according to the position relationship logic.

[0079] In another optional implementation, the user only needs to perform a selection operation on the moving targets in one of the image data or the point cloud data, and the other data is automatically labeled and sorted. Specifically, the target data is first determined from the image data and the point cloud data, and the moving targets in the target data are first sorted; then the non-target data is determined from the image data and the point cloud data, and the second sorting of all moving targets in the non-target data is obtained; and then the moving targets in the image data and the point cloud data are matched according to the first sorting and the second sorting.

[0080] If the mobile target identified by the image data is the same as the mobile target identified by the point cloud data, the user only needs to follow the first sorting to sequentially perform the second sorting in the non-target data. If the mobile target identified by the image data is not the same as the mobile target identified by the point cloud data, when the user performs the second sorting in the non-target data, if the mobile target with a certain label in the target data does not appear in the non-target data, an empty selection operation needs to be performed. The empty selection operation can be a double-click of the mouse in the blank area of the non-target data or clicking the empty selection operation button. The above implementation further reduces the operation steps.

[0081] For example, the image data is taken as the target data and the point cloud data is taken as the non-target data. The image data includes three mobile targets A, B and C, and the point cloud data includes four mobile targets a, b, d and e. The three mobile targets of the image data are automatically sorted in the first sorting, i.e., the mobile target A is labeled 1, the mobile target B is labeled 2, and the mobile target C is labeled 3. Then, the user clicks the mobile target a in the point cloud data with reference to the first sorting, and the mobile target a is labeled 1. The mobile target b is clicked, and the mobile target b is labeled 2. Double-clicking in the blank area indicates empty selection, i.e., the mobile target C is not matched. Finally, the matched mobile target A and the mobile target a, the mobile target B and the mobile target b are obtained.

[0082] It is worth mentioning that, unlike fixed objects, the position of the mobile target changes over time, and the data collected at different times of the same scene can provide more mapping anchors. Based on this, the image data of the intersection area collected by the two-dimensional perception device and the point cloud data of the intersection area collected by the three-dimensional perception device can be reacquired at different times, and then the above steps are followed to obtain new mapping anchors. In this way, by obtaining more mapping anchors, the mapping accuracy can be greatly improved.

[0083] In summary, the mapping anchor calibration method based on a mobile target provided by the embodiments of the present specification obtains image data of an intersection area collected by a two-dimensional perception device and point cloud data of the intersection area collected by a three-dimensional perception device; then determines the mobile target in the image data, the position coordinates of the mobile target, and the mobile target in the point cloud data and the position coordinates of the mobile target according to the image data, the point cloud data, and a trained target detection model; then obtains the selection operation of the user on the mobile target in the image data and / or the point cloud data, and matches the mobile target in the image data and the point cloud data based on the selection operation to obtain the matched mobile target; finally, determines the mapping anchor according to the position coordinates of the matched mobile target. In this way, the mobile target in the image data and the point cloud data is determined by automatic detection, without manual framing and the tedious process of manual point finding, so that the position of the mapping anchor is more accurate. Compared with fixed targets such as street lamps and flowerpots, mobile targets are more widely distributed in the scene, cover more points, and their position coverage area is also the area of interest for data analysis.

[0084] Based on the above application scenarios, combined with Figure 3 The embodiments of the present specification also provide a mapping anchor calibration system based on a mobile target, which is applicable to the mapping anchor calibration method described above, and the system comprises:

[0085] A two-dimensional perception device is used to collect image data of an intersection area;

[0086] A three-dimensional perception device is used to collect point cloud data of the intersection area;

[0087] A first processing module is used to determine the mobile target in the image data, the position coordinates of the mobile target, and the mobile target in the point cloud data and the position coordinates of the mobile target according to the image data, the point cloud data, and a trained target detection model;

[0088] A second processing module is used to obtain the selection operation of the user on the mobile target in the image data and / or the point cloud data, and match the mobile target in the image data and the point cloud data based on the selection operation to obtain the matched mobile target;

[0089] An anchor point acquisition module is used to determine the mapping anchor according to the position coordinates of the matched mobile target.

[0090] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the mapping anchor calibration system based on a mobile target described above can refer to the corresponding process in the foregoing method, and will not be described in more detail here.

[0091] In summary, the embodiment of the present specification provides a mapping anchor calibration system based on a moving target. The system obtains image data of an intersection area collected by a two-dimensional perception device and point cloud data of the intersection area collected by a three-dimensional perception device. Then, the system determines the moving target in the image data, the position coordinates of the moving target, and the moving target in the point cloud data and the position coordinates of the moving target according to the image data, the point cloud data, and a trained target detection model. Next, the system obtains a selection operation of a user on the moving target in the image data and / or the point cloud data, and matches the moving target in the image data and the point cloud data based on the selection operation to obtain a matched moving target. Finally, the system determines the mapping anchor according to the position coordinates of the matched moving target. In this way, the moving target in the image data and the point cloud data is determined by automatic detection, without manual framing and manual point finding, so that the position of the mapping anchor is more accurate. Compared with fixed targets such as street lamps and flowerpots, the moving target is more widely distributed in the scene, covers more points, and its position coverage area is also the area of interest for data analysis.

[0092] Please refer to Figure 4 , Figure 4 A structural block diagram of an electronic device 100 is provided for the embodiment. The electronic device can include a mapping anchor calibration apparatus 10, a memory 20, a processor 30, and a communication unit 40. The memory 20 stores machine-readable instructions executable by the processor 30. When the electronic device 100 is running, the processor 30 and the memory 20 communicate through a bus. The processor 30 executes the machine-readable instructions and performs a mapping anchor calibration method based on a moving target.

[0093] The memory 20, the processor 30, and the communication unit 40 are electrically connected to each other directly or indirectly to realize the transmission or interaction of signals. For example, these elements can be electrically connected to each other through one or more communication buses or signal lines. The mapping anchor calibration apparatus 10 based on a moving target includes at least one software function module stored in the memory 20 in the form of software or firmware. The processor 30 is configured to execute the executable modules (such as software function modules or computer programs included in the mapping anchor calibration apparatus 10) stored in the memory 20.

[0094] The memory 20 can be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), an Electric Erasable Programmable Read-Only Memory (EEPROM), etc.

[0095] In some embodiments, the processor 30 is configured to perform one or more functions described in the present embodiments. In some embodiments, the processor 30 can include one or more processing cores (e.g., a single-core processor (S) or a multi-core processor (S)). For example only, the processor 30 can include a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), an Application Specific Instruction-set Processor (ASIP), a Graphics Processing Unit (GPU), a Physics Processing Unit (PPU), a Digital Signal Processor (DSP), a Field-Programmable Gate Array (FPGA), a Programmable Logic Device (PLD), a controller, a microcontroller unit, a Reduced Instruction Set Computer (RISC), or a microprocessor, etc., or any combination thereof.

[0096] For ease of illustration, only one processor is described in the electronic device 100. However, it should be noted that the electronic device 100 in the present embodiment can also include multiple processors, and thus the steps performed by one processor described in the present embodiment can also be performed jointly by multiple processors or individually. For example, if a processor of a server performs steps A and B, it should be understood that steps A and B can also be performed jointly by two different processors or individually in one processor. For example, a processor performs step A, a second processor performs step B, or the processor and the second processor jointly perform steps A and B.

[0097] In the present embodiment, the memory 20 is configured to store a program, and the processor 30 is configured to execute the program upon receiving an execution instruction. The method defined by the flow disclosed in any of the embodiments of the present embodiment can be applied in the processor 30 or implemented by the processor 30.

[0098] The communication unit 40 is configured to establish a communication connection between the electronic device 100 and other devices through a network, and configured to transceive data through the network.

[0099] In some embodiments, the network can be any type of wired or wireless network, or a combination thereof. By way of example only, the network can include a wired network, a wireless network, a fiber optic network, a telecommunications network, an intranet, the Internet, a Local Area Network (LAN), a Wide Area Network (WAN), a wireless Local Area Network (WLAN), a Metropolitan Area Network (MAN), a Wide Area Network (WAN), a Public Switched Telephone Network (PSTN), a Bluetooth network, a ZigBee network, or a Near Field Communication (NFC) network, etc., or any combination thereof.

[0100] In the present embodiment, the electronic device 100 can be, but is not limited to, a notebook computer, an ultra-mobile personal computer (UMPC), a netbook, a Personal Digital Assistant (PDA), and the like. The present embodiment does not make any limitation on the specific type of the electronic device.

[0101] It can be understood that, Figure 4The structure shown is only schematic. The electronic device 100 can also have more or fewer components than shown, or have them in different configurations or arrangements. Figure 4 The components shown can be implemented in hardware, software, or a combination thereof. Figure 4 The components shown can be implemented in hardware, software, or a combination thereof. Figure 4 The components shown can be implemented in hardware, software, or a combination thereof.

[0102] On the basis of the foregoing, the embodiment provides a readable storage medium, and the readable storage medium stores a computer program. The computer program is executed by a processor to implement the mapping anchor calibration method based on a moving target according to any one of the foregoing embodiments.

[0103] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the readable storage medium described above can refer to the corresponding process in the foregoing method, and will not be described in detail here.

[0104] To sum up, the mapping anchor calibration method, system, and electronic device based on a moving target provided by the embodiments of the present disclosure are provided. The image data of the intersection region collected by the two-dimensional perception device and the point cloud data of the intersection region collected by the three-dimensional perception device are obtained. Then, the moving target in the image data, the position coordinates of the moving target, and the moving target in the point cloud data, the position coordinates of the moving target are determined according to the image data, the point cloud data, and the trained target detection model. Then, the matching operation of the user on the moving target in the image data and / or the point cloud data is obtained, and the moving target in the image data and the point cloud data is matched based on the matching operation to obtain the matched moving target. Finally, the mapping anchor point is determined according to the position coordinates of the matched moving target. In this way, the moving target in the image data and the point cloud data is determined by automatic detection, without manual framing, and the tedious process of manual point finding is avoided, so that the position of the mapping anchor point is more accurate. Compared with fixed targets such as street lamps and flowerpots, moving targets are more widely distributed in the scene, cover more points, and their position coverage area is also the area of interest for data analysis.

[0105] The above is only various embodiments of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present disclosure, which should be covered within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be subject to the protection scope of the claims.

Claims

1. A method for mapping anchor point calibration based on a moving target, characterized in that, The method comprises: acquiring image data of a road intersection region collected by a two-dimensional perception device and point cloud data of the road intersection region collected by a three-dimensional perception device; determining a moving target in the image data, a position coordinate of the moving target, a moving target in the point cloud data, and a position coordinate of the moving target according to the image data, the point cloud data, and a trained target detection model; acquiring a user's selection operation on the moving target in the image data and / or the point cloud data, and matching the moving target in the image data and the point cloud data based on the selection operation to obtain a matched moving target; determining a mapping anchor point according to a position coordinate of the matched moving target; the acquiring of the user's selection operation on the moving target in the image data and / or the point cloud data and the matching of the moving target in the image data and the point cloud data based on the selection operation comprises: acquiring a moving target selected by the user in the image data and a first order of selecting the moving target; acquiring a moving target selected by the user in the point cloud data and a second order of selecting the moving target; matching the moving target selected by the user in the image data and the moving target selected by the user in the point cloud data according to the first order and the second order; the acquiring of the user's selection operation on the moving target in the image data and / or the point cloud data and the matching of the moving target in the image data and the point cloud data based on the selection operation comprises: acquiring a first selection operation of the user in the image data to obtain a plurality of moving targets; acquiring a second selection operation of the user in the point cloud data corresponding to the first selection operation to obtain a plurality of moving targets; matching according to a position relationship logic between the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation; the acquiring of the user's selection operation on the moving target in the image data and / or the point cloud data and the matching of the moving target in the image data and the point cloud data based on the selection operation comprises: determining target data from the image data and the point cloud data, and first sorting moving targets in the target data; determining non-target data from the image data and the point cloud data, and acquiring second sorting of moving targets in the non-target data by the user; matching the moving targets in the image data and the point cloud data according to the first sorting and the second sorting; the determining of the mapping anchor point according to the position coordinate of the matched moving target comprises: determining a center point coordinate of the matched moving target according to the position coordinate of the matched moving target; determining the mapping anchor point according to the center point coordinate of the matched moving target.

2. The mobile object-based mapping anchor point calibration method according to claim 1, wherein, Before the matching according to the position relationship logic between the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation, the method further comprises: In the preset position relationship logic table, all position relationship logics of the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation are traversed; The position relationship logic of the plurality of moving targets of the first selection operation and the plurality of moving targets of the second selection operation is determined.

3. The mobile object-based mapping anchor point calibration method of claim 1, wherein, The method further comprises: At different time instants, the image data of the intersection region collected by the two-dimensional perception device and the point cloud data of the intersection region collected by the three-dimensional perception device are reacquired to obtain new mapping anchor points.

4. A mobile target based mapping anchor calibration system, comprising: The system comprises the mapping anchor point calibration method based on moving targets according to any one of claims 1-3. The two-dimensional perception device is configured to collect image data of the intersection region; The three-dimensional perception device is configured to collect point cloud data of the intersection region; The first processing module is configured to determine, according to the image data, the point cloud data and the trained target detection model, the moving targets in the image data, the position coordinates of the moving targets, and the moving targets in the point cloud data, and the position coordinates of the moving targets; The second processing module is configured to acquire the selection operation of the user on the moving targets in the image data and / or the point cloud data, and match the moving targets in the image data and the point cloud data based on the selection operation to obtain matched moving targets; The anchor point acquisition module is configured to determine, according to the position coordinates of the matched moving targets, the mapping anchor points.

5. An electronic device, comprising: The electronic device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, and the processor implements the mapping anchor point calibration method based on moving targets according to any one of claims 1-3 when executing the computer program.

6. A computer-readable storage medium, characterized in that, The computer program is stored on the memory and executable on the processor, and the processor implements the mapping anchor point calibration method based on moving targets according to any one of claims 1-3 when executing the computer program.

Images