Target position estimation method and device, computer equipment and storage medium

Abstract

The invention relates to a target position estimation method and device, computer equipment and a storage medium. The method comprises the following steps: based on at least two pieces of target detection information of a to-be-estimated target, determining a target detection frame corresponding to each piece of target detection information in a preset coordinate system; the target detection information comprises a detection position, a detection course angle and a detection size; fitting based on the at least two target detection frames to obtain a target fitting frame of the to-be-estimated target; and determining a position estimation result of the to-be-estimated target based on the target fitting frame. By adopting the method, the data transmission and calculation amount requirements can be reduced, and the target position estimation result can be timely and accurately provided.

Description

Technical Field

[0001] The present application relates to the technical field of target detection, and in particular to a target position estimation method, device, computer equipment and storage medium. Background Art

[0002] With the development of the field of autonomous driving, more and more OEMs have adopted at least two rear-corner radars to achieve target detection in blind spots and lane changes. Correspondingly, due to the large viewing angle of the rear-corner radar, there will be overlapping detection areas. Now, for the overlapping detection areas, the detection results of the same target are different, and the detection position of the target cannot be accurately provided.

[0003] In traditional technology, the point cloud data of all millimeter-wave radars are fused and calculated to jointly estimate the position of the detection target. However, this method requires the transmission of a large amount of point cloud data, so the data transmission and calculation pressure is high, which affects the real-time and integrity of the data.

[0004] It can be seen that the existing technology for estimating the target position in overlapping detection areas still has the problems of large data transmission and calculation requirements and low timeliness. Summary of the invention

[0005] Based on this, it is necessary to provide a target position estimation method, device, computer equipment and storage medium that can reduce data transmission and calculation requirements and provide target position estimation results in a timely and accurate manner to address the above technical problems.

[0006] In a first aspect, the present application provides a target position estimation method, which is applied to a target position estimation device, wherein the target position estimation device includes at least two radar detection modules, and the detection ranges of the at least two radar detection modules overlap in whole or in part. The target position estimation method includes:

[0007] Based on at least two target detection information of the target to be estimated, determining a target detection frame corresponding to each of the target detection information in a preset coordinate system; the target detection information includes a detection position, a detection heading angle, and a detection size;

[0008] Perform fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated;

[0009] Based on the target fitting frame, a position estimation result of the target to be estimated is determined.

[0010] In one of the embodiments, determining a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated includes:

[0011] Determining the length and width of the target detection frame based on the detection size of the target detection information;

[0012] Based on the detected position, determining the coordinates of the center point of the target detection frame in a preset coordinate system;

[0013] Based on the length and width, the center point coordinates and the heading angle, a target detection frame corresponding to the current target detection information in the preset coordinate system is determined.

[0014] In one embodiment, the target detection frame includes a plurality of detection frame target points; the detection frame target points include detection frame corner points; and determining the target detection frame corresponding to the current target detection information in the preset coordinate system based on the length and width, the center point coordinates, and the heading angle includes:

[0015] Based on the heading angle and the preset coordinate point, a detection frame coordinate system is determined; the preset coordinate point is the origin of the preset coordinate system;

[0016] Based on the relative position of each detection frame target point and the center point and the length and width, determine a first distance difference between each detection frame target point and the center point in the detection frame coordinate system;

[0017] Based on the heading angle, the first distance difference is rotated to obtain a second distance difference between the target point of each detection frame and the center point in a preset coordinate system;

[0018] Based on the second distance difference of the plurality of detection frame target points and the center point coordinates, determining the coordinates of the plurality of detection frame corner points corresponding to the current target detection information in the preset coordinate system;

[0019] The target detection frame is determined based on the coordinates of the multiple detection frame corner points.

[0020] In one embodiment, rotating the first distance difference based on the heading angle to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system includes:

[0021] Based on the heading angle, determining a rotation matrix corresponding to the heading angle;

[0022] A matrix multiplication operation is performed based on the rotation matrix and the first distance difference to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system.

[0023] In one embodiment, performing fitting based on at least two of the target detection frames to obtain the target fitting frame of the target to be estimated includes:

[0024] Determining the heading angle of the target fitting frame based on the heading angles of at least two targets to be estimated;

[0025] Determine a fitting frame coordinate system based on the heading angle of the target fitting frame and preset coordinate points;

[0026] Respectively calculating the coordinates of a plurality of detection frame target points of each of the target detection frames in the fitting frame coordinate system; the detection frame target points include detection frame corner points;

[0027] Based on the multiple detection frame target points, determine multiple fitting frame target points of the target fitting frame;

[0028] Based on the multiple fitting box target points, a target fitting box of the target to be estimated is determined.

[0029] In one embodiment, determining a plurality of fitting frame target points of a target fitting frame based on a plurality of detection frame target points comprises:

[0030] Based on the horizontal coordinates and vertical coordinates of the plurality of detection frame target points, respectively determine the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate of the target to be estimated in the fitting frame coordinate system;

[0031] Based on the maximum abscissa, the minimum abscissa, the maximum ordinate and the minimum ordinate, a plurality of the fitting frame target points of the target fitting frame are determined.

[0032] In one embodiment, the position estimation result includes a geometric center position, and determining the position estimation result of the target to be estimated based on the target fitting frame includes:

[0033] Based on the heading angle of the target fitting frame, determining an update point of the target to be estimated; the update point includes one of a left front corner point, a right front corner point and a center point of the front side surface;

[0034] Based on the update point, the geometric center position of the target to be estimated is determined.

[0035] In one embodiment, determining the geometric center position of the target to be estimated based on the update point includes:

[0036] Obtaining the actual length and actual width of the target to be estimated;

[0037] Based on the update point, the actual length and the actual width, determining a distance difference between the update point and a geometric center position in a coordinate system corresponding to the target fitting frame;

[0038] Based on the distance difference, determining a fitting center position of the target to be estimated in a coordinate system corresponding to the target fitting frame;

[0039] Based on the heading angle of the fitting frame, the fitting center position is rotated to obtain the geometric center position of the target to be estimated in the preset coordinate system.

[0040] In a second aspect, the present application provides a target position estimation device, comprising a position estimation component and at least two radar detection modules, wherein the detection ranges of the at least two radar detection modules overlap in whole or in part, and the position estimation component comprises a detection frame determination module, a fitting frame determination module and a position estimation module, wherein:

[0041] A detection frame determination module, configured to determine a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated; the target detection information includes a detection position, a detection heading angle, and a detection size;

[0042] A fitting frame determination module, used for performing fitting based on at least two target detection frames to obtain a target fitting frame of the target to be estimated;

[0043] The position estimation module is used to determine the position estimation result of the target to be estimated based on the target fitting frame.

[0044] In a third aspect, the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned method when executing the computer program.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the method described above when executed by a processor.

[0046] The above-mentioned target position estimation method, device, computer equipment and storage medium determine the target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated; the target detection information includes a detection position, a detection heading angle and a detection size; fitting is performed based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated; based on the target fitting frame, the position estimation result of the target to be estimated is determined, using the target level output, and only processing is performed through the detection position, detection heading angle and detection size in the target detection information, and the target detection frame is calculated separately, thereby avoiding the amount of calculation brought by point cloud data fusion and ensuring real-time performance. At the same time, it realizes the combination of target detection information of at least two radars to improve the accuracy of backward target position detection, which can reduce the requirements for data transmission and calculation, and achieve the effect of improving the timeliness and accuracy of the target position estimation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A diagram showing an application environment of a target position estimation method in an embodiment;

[0048] Figure 2 is a schematic diagram of a flow chart of a target position estimation method in one embodiment;

[0049] Figure 3 is a schematic flow chart of a target position estimation method in another embodiment;

[0050] Figure 4 is a schematic diagram of a target detection frame in a vehicle coordinate system in one embodiment;

[0051] Figure 5 is a schematic diagram of multiple target detection frames in the vehicle coordinate system in one embodiment;

[0052] Figure 6 is a schematic diagram of a target fitting frame in one embodiment;

[0053] Figure 7 is a schematic diagram of reference points of a fitting frame in one embodiment;

[0054] Figure 8 A schematic diagram of the angle between the target fitting frame and the origin of the vehicle coordinate system in one embodiment;

[0055] Fig. 9 A schematic diagram of selecting the center point of the front side surface based on the included angle in one embodiment;

[0056] Fig.10 A schematic diagram of selecting a right front corner point based on an included angle in one embodiment;

[0057] Fig.11 A diagram showing a position estimation result of a target obliquely crossing in one embodiment;

[0058] Fig.12 A diagram showing a position estimation result of a target moving straight ahead in one embodiment;

[0059] Fig.13 is a structural block diagram of a target position estimation device in one embodiment;

[0060] Fig.14 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] The target position estimation method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. The terminal 102 communicates with at least two radar detection modules 210 through a wired or wireless network. The radar detection module 210 is used to obtain target detection information of the target to be estimated; the terminal 102 determines the target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated; the target detection information includes a detection position, a detection heading angle and a detection size; fitting is performed based on at least two target detection frames to obtain a target fitting frame of the target to be estimated; based on the target fitting frame, the position estimation result of the target to be estimated is determined. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers and Internet of Things devices, and the Internet of Things device can be a smart vehicle-mounted device.

[0063] In one embodiment, Figure 2 As shown, a target position estimation method is provided, which is applied to a target position estimation device, wherein the target position estimation device includes at least two radar detection modules, and the detection ranges of the at least two radar detection modules are completely or partially overlapped. Figure 1 The terminal 102 in the example is used as an example to illustrate, and the following steps are included:

[0064] Step S100: determining a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated.

[0065] Among them, the target detection information can be collected and acquired by the radar detection module, including the detection position, detection heading angle and detection size. The detection position can be the position of the target to be estimated relative to the radar detection module or the preset coordinate system; the detection heading angle can be the orientation angle of the target to be estimated relative to the radar detection module or the preset coordinate system; the detection size can be the size of the target to be estimated based on the detection, which can be used to define the size of the target detection frame.

[0066] The preset coordinate system may be the coordinate system of the subject on which the method is executed. Exemplarily, when the target position estimation method of this embodiment is applied to a vehicle, the preset coordinate system may be the coordinate system of the vehicle. When the target position estimation method is applied to other terminal devices, the preset coordinate system may also be set accordingly according to the usage mode and application purpose of the terminal device, which is not limited in this embodiment.

[0067] Furthermore, the origin of the preset coordinate system may be pre-set according to the execution subject. For example, when the target position estimation method of this embodiment is applied to the vehicle end, the origin of the preset coordinate system may be any position of the front axle or rear axle of the vehicle end, such as the center of the rear axle.

[0068] By determining a target detection frame corresponding to each target detection information in a preset coordinate system, the target detection information detected by each radar module can be converted into a standardized representation in the same coordinate system.

[0069] Step S200: performing fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated.

[0070] The fitting of at least two of the target detection frames may be performed by fitting all the target detection frames obtained above, so as to obtain a target fitting frame of the target to be estimated.

[0071] Fitting the target detection frame may be fitting to obtain the most likely actual detection frame of the target to be estimated. Exemplarily, the detection heading angle, detection size and detection position of at least two target detection frames may be fitted to obtain a fitted target fitting frame. When the number of target detection information is not less than three, the radar detection module closest to the target to be estimated may be determined based on the distribution of detection positions in the target detection information, and a higher weight may be assigned based on the target detection information provided by the radar detection module, and the target detection information with different weights may be fitted to obtain a target fitting frame of the target to be estimated.

[0072] The target detection frame may be fitted to obtain a minimum fitting frame including multiple target detection frames. Exemplarily, the range of the target fitting frame may be a minimum range including at least two of the target detection frames.

[0073] By comprehensively considering the detection information from different radar modules and eliminating the inconsistencies caused by sensor errors, viewing angle differences and other factors, a target fitting frame that is closer to the actual target position and size is finally generated. This can improve the accuracy of target position estimation, avoid the need for direct processing of raw point cloud data, and reduce computational complexity.

[0074] Step S300: determining a position estimation result of the target to be estimated based on the target fitting frame.

[0075] Among them, the position estimation result of the target to be estimated is determined by the target fitting frame. It can be that the point with the smallest distance between the target to be estimated and the terminal 102 or the target position estimation device is determined according to the target fitting frame, or the geometric center point or center of mass point of the target to be estimated is determined as the position estimation result of the target to be estimated.

[0076] A target position estimation method provided in this embodiment determines a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of a target to be estimated; the target detection information includes a detection position, a detection heading angle and a detection size; fitting is performed based on at least two target detection frames to obtain a target fitting frame of the target to be estimated; based on the target fitting frame, a position estimation result of the target to be estimated is determined, which uses target-level output, processes only the detection position, the detection heading angle and the detection size in the target detection information, and calculates the target detection frame separately, thereby avoiding the amount of calculation caused by point cloud data fusion and ensuring real-time performance. At the same time, it realizes the combination of target detection information of at least two radars to improve the accuracy of backward target position detection, which can reduce the requirements for data transmission and calculation, and achieve the effect of improving the timeliness and accuracy of target position estimation results.

[0077] In one of the embodiments, determining a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated includes:

[0078] Determining the length and width of the target detection frame based on the detection size of the target detection information;

[0079] Based on the detected position, determining the coordinates of the center point of the target detection frame in a preset coordinate system;

[0080] Based on the length and width, the center point coordinates and the heading angle, a target detection frame corresponding to the current target detection information in the preset coordinate system is determined.

[0081] Among them, based on the detection size of the target detection information, the length and width of the target detection frame are determined, and the length and width of the target detection frame can be directly set or calculated according to the detection size provided by the radar detection module.

[0082] Based on the detection position, the coordinates of the center point of the target detection frame in the preset coordinate system are determined, and the detection position can be converted in the preset coordinate system to determine the coordinates of the center point of the target detection frame. Further, the detection position provided by the radar detection module may be the nearest point between the target to be estimated and the terminal 102 or the target position estimation device, or the center point of the nearest edge, which is called an anchor point. Since the anchor point and the center point of the target detection frame may not coincide, the coordinates of the center point of the target detection frame can also be calculated based on the distance difference between the anchor point and the center point.

[0083] Based on the length and width dimensions, the center point coordinates and the heading angle, a target detection frame corresponding to the current target detection information in the preset coordinate system is determined. The occupied range of the target detection frame in the preset coordinate system can be determined according to the length and width dimensions and the center point coordinates, and the target detection frame is rotated according to the heading angle, so as to determine the position, size and angle of the target detection frame in the preset coordinate system.

[0084] Furthermore, the target detection frame can be represented by connecting four corner points. Therefore, the coordinate positions of the four corner points of the target detection frame relative to the center point coordinates can be determined based on the length and width dimensions and the center point coordinates, and the coordinate positions of the four corner points relative to the center point coordinates can be rotated based on the heading angle to obtain the relative coordinate positions of the four corner points after rotation, and then the coordinate positions of the four corner points in the preset coordinate system can be converted.

[0085] A target position estimation method provided in this embodiment can create a target detection frame that conforms to the actual target size and orientation by combining the length and width dimensions, center point coordinates and heading angle, and realize the conversion of target detection information to the same reference coordinate system, thereby achieving the effect of improving the timeliness and accuracy of the target position estimation results.

[0086] In one embodiment, determining the target detection frame corresponding to the current target detection information in the preset coordinate system based on the length and width, the center point coordinates, and the heading angle includes:

[0087] Determine a detection frame coordinate system based on the heading angle and preset coordinate points;

[0088] Based on the relative position of each detection frame target point and the center point and the length and width, determine a first distance difference between each detection frame target point and the center point in the detection frame coordinate system;

[0089] Based on the heading angle, the first distance difference is rotated to obtain a second distance difference between the target point of each detection frame and the center point in a preset coordinate system;

[0090] Based on the second distance difference of the plurality of detection frame target points and the center point coordinates, determining the coordinates of the plurality of detection frame corner points corresponding to the current target detection information in the preset coordinate system;

[0091] The target detection frame is determined based on the coordinates of the multiple detection frame corner points.

[0092] The preset coordinate point is the origin of the preset coordinate system. The selection of the origin can refer to the method for selecting the origin in the above embodiment, which will not be described in detail in this embodiment.

[0093] Based on the heading angle and the preset coordinate point, the detection frame coordinate system is determined. The preset coordinate point can be used as the origin, but the corresponding coordinate axis is determined according to the heading angle, so as to obtain the detection frame coordinate system with the same origin as the preset coordinate system. For example, the X axis of the detection frame coordinate system can be determined by the heading angle, and the direction to the left of the target to be estimated and perpendicular to the X axis is used as the Y axis.

[0094] The target detection frame may include a plurality of detection frame target points; the detection frame target points include detection frame corner points. Furthermore, the detection frame target points may also include detection frame edge midpoints.

[0095] The first distance difference may be the relative distance between the detection frame target point and the center point coordinates. The relative position of the detection frame target point and the center point may include left front, right front, left rear, right rear, front side, right side, rear side, left side, etc. Therefore, based on the relative position of each detection frame target point and the center point, the length and width dimensions and the center point coordinates, the first distance difference between each detection frame target point and the center point in the detection frame coordinate system is determined. The relative distance to the center point coordinates is determined based on the position of the detection frame target point relative to the center point, combined with the length and width dimensions, and expressed as the first distance difference. The first distance difference may be expressed in the form of relative coordinates.

[0096] Based on the heading angle, the first distance difference is rotated to obtain the second distance difference between each detection frame target point and the center point in the preset coordinate system. The relative coordinates can be converted according to the heading angle to convert the first distance difference in the detection frame coordinate system into the second distance difference in the preset coordinate system.

[0097] Based on the second distance difference of the multiple detection frame target points and the center point coordinates, the coordinates of the multiple detection frame corner points corresponding to the current target detection information in the preset coordinate system are determined. This can be done by combining the center point coordinates in the preset coordinate system with the second distance difference of each detection frame target point, so that the coordinates of each detection frame target point of the target detection frame corresponding to the target detection information in the preset coordinate system, that is, the coordinates of the detection frame corner points, can be determined.

[0098] In this embodiment, the target detection frame may be composed of a plurality of detection frame target points, and further, may be composed of four detection frame corner points. When the coordinates of the detection frame corner points in the preset coordinate system are determined, the target detection frame may be determined.

[0099] The present embodiment provides a method for estimating a target position, which establishes a detection frame coordinate system, calculates a first distance difference between a target point in each detection frame and a center point, and then converts the first distance difference to a second distance difference in a preset coordinate system, thereby determining a target detection frame in the preset coordinate system. This method can achieve high-precision and efficient detection frame calculation, avoid direct processing of a large amount of point cloud data, and improve the timeliness and accuracy of the target position estimation results.

[0100] In one embodiment, rotating the first distance difference based on the heading angle to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system includes:

[0101] Based on the heading angle, determining a rotation matrix corresponding to the heading angle;

[0102] A matrix multiplication operation is performed based on the rotation matrix and the first distance difference to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system.

[0103] In this embodiment, the first distance difference may be converted by a rotation matrix. The rotation matrix may be constructed according to a given heading angle. The rotation matrix may be an orthogonal matrix for realizing the rotation transformation of a point, line or figure in a plane or space.

[0104] In one embodiment, the rotation matrix R θ It can be expressed as:

[0105]

[0106] Where θ is the opposite of the heading angle yaw.

[0107] A target position estimation method provided in this embodiment can realize the conversion of the distance difference from the detection frame coordinate system to the preset coordinate system by performing matrix multiplication operation based on the rotation matrix and the first distance difference, thereby obtaining the second distance difference between each detection frame target point and the center point in the preset coordinate system. The position of each detection frame target point in the preset coordinate system can be determined more accurately, the real-time performance of the system is enhanced, and the quality of multi-sensor data fusion is improved, so as to achieve the effect of improving the timeliness and accuracy of the target position estimation results.

[0108] In one embodiment, performing fitting based on at least two of the target detection frames to obtain the target fitting frame of the target to be estimated may include:

[0109] Determine a fitting frame coordinate system based on the heading angles and preset coordinate points of at least two targets to be estimated;

[0110] Based on multiple detection frame target points of each target detection frame, the target fitting frame of the target to be estimated is calculated. The fitting frame coordinate system can be based on the preset coordinate point as the origin, and the X axis and Y axis of the fitting frame coordinate system are determined in combination with the heading angles of at least two targets to be estimated. The preset coordinate point and the determination of the X axis and Y axis based on the heading angle are not described in detail in this embodiment.

[0111] In this embodiment, by first determining the fitting frame coordinate system and obtaining multiple detection frame target points in the fitting frame coordinate system, the fitting frame in the fitting frame coordinate system can be further determined, and the fitting frame can be rotated to the preset coordinate system, that is, the target fitting frame of the target to be estimated in the preset coordinate system can be obtained.

[0112] Furthermore, the target fitting box may be a minimum detection box including multiple target detection boxes.

[0113] In one embodiment, performing fitting based on at least two of the target detection frames to obtain the target fitting frame of the target to be estimated includes:

[0114] Determining the heading angle of the target fitting frame based on the heading angles of at least two targets to be estimated;

[0115] Determine a fitting frame coordinate system based on the heading angle of the target fitting frame and preset coordinate points;

[0116] Respectively calculating the coordinates of a plurality of detection frame target points of each of the target detection frames in the fitting frame coordinate system; the detection frame target points include detection frame corner points;

[0117] Based on the multiple detection frame target points, determine multiple fitting frame target points of the target fitting frame;

[0118] Based on the multiple fitting box target points, a target fitting box of the target to be estimated is determined.

[0119] Among them, based on the heading angles of at least two of the targets to be estimated, the heading angle of the target fitting frame is determined, which can be achieved by taking the average value, median value, etc. of at least two heading angles. When there is enough target detection information, clustering and other algorithms can also be used to determine the fitting frame heading angle that best matches the actual heading angle.

[0120] The preset coordinate point may be the origin of the preset coordinate system. Based on the heading angle of the target fitting frame and the preset coordinate point, a fitting frame coordinate system may be established with the preset coordinate point as the origin and the heading angle of the target fitting frame.

[0121] The coordinates of multiple detection frame target points of each target detection frame in the fitting frame coordinate system are calculated respectively; the detection frame target points include detection frame corner points, and the multiple detection frame target points under each target detection frame can be converted from a preset coordinate system to a fitting frame coordinate system. Exemplarily, each detection frame target point can be rotated based on the heading angle of the target fitting frame to obtain the rotated detection frame target point in the fitting frame coordinate system.

[0122] Based on the multiple detection frame target points, multiple fitting frame target points of the target fitting frame are determined, which can be determined according to the horizontal coordinates and vertical coordinates of the multiple detection frame target points. The fitting frame target points can be used to select the boundaries of the detection frame target points.

[0123] After obtaining multiple fitting frame target points, the target fitting frame of the target to be estimated can be determined. Further, the multiple fitting frame target points can be rotated based on the heading angle of the fitting frame to obtain the target fitting frame in a preset coordinate system.

[0124] A target position estimation method provided in this embodiment establishes a fitting frame coordinate system, and determines multiple fitting frame target points based on multiple detection frame target points, thereby determining the target fitting frame in a preset coordinate system. This can achieve high-precision and efficient fitting frame calculation, avoid direct processing of a large amount of point cloud data, and achieve the effect of improving the timeliness and accuracy of the target position estimation result.

[0125] In one embodiment, determining a plurality of fitting frame target points of a target fitting frame based on a plurality of detection frame target points comprises:

[0126] Based on the horizontal coordinates and vertical coordinates of the plurality of detection frame target points, respectively determine the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate of the target to be estimated in the fitting frame coordinate system;

[0127] Based on the maximum abscissa, the minimum abscissa, the maximum ordinate and the minimum ordinate, a plurality of the fitting frame target points of the target fitting frame are determined.

[0128] Among them, based on the horizontal coordinates and vertical coordinates of multiple detection frame target points, the maximum horizontal coordinate, minimum horizontal coordinate, maximum vertical coordinate and minimum vertical coordinate of the target to be estimated in the fitting frame coordinate system are respectively determined. The horizontal coordinates of multiple detection frame target points can be compared to obtain the minimum horizontal coordinate and the maximum horizontal coordinate, and the vertical coordinates of multiple detection frame target points can be compared to obtain the minimum vertical coordinate and the maximum vertical coordinate.

[0129] By combining the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate, a plurality of fitting frame target points of the target fitting frame can be determined. Further, the fitting frame target points are the lower left corner point, the lower right corner point, the upper left corner point and the upper right corner point.

[0130] A target position estimation method provided in this embodiment determines the boundary of the target fitting box by determining the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate, so as to ensure that the target fitting box tightly wraps all the target points of the detection box, and at the same time avoids unnecessary redundant space, simplifies the calculation process, reduces processing time and resource consumption, and can achieve the effect of improving the timeliness and accuracy of the target position estimation result.

[0131] In one embodiment, the position estimation result includes a geometric center position, and determining the position estimation result of the target to be estimated based on the target fitting frame includes:

[0132] Determining an update point of the target to be estimated based on the heading angle of the target fitting frame;

[0133] Based on the update point, the geometric center position of the target to be estimated is determined.

[0134] Wherein, the update point includes one of the left front corner point, the right front corner point and the center point of the front side. The update point can be used to calculate the geometric center position of the target to be estimated, so as to determine the position of the target to be estimated. Based on the heading angle of the target fitting frame, the update point of the target to be estimated is determined, which can be based on the size of the heading angle, to determine the position of the point on the target fitting frame of the target to be estimated that is closest to the terminal 102 or the target position estimation device. Exemplarily, when the absolute value of the heading angle of the target to be estimated is greater than a preset threshold, it means that the heading angle of the target to be estimated deviates too much from the heading angle of the terminal 102 or the target position estimation device. At this time, the distance between the upper left corner point or the upper right corner point on the target fitting frame of the target to be estimated is closer, and the upper left corner point or the upper right corner point can be used as the update point; if the absolute value of the heading angle is less than the preset threshold, it means that the heading angle of the target to be estimated is closer to the heading angle of the terminal 102 or the target position estimation device, and the center point of the front side of the target to be estimated can be used as the update point.

[0135] Based on the update point, the geometric center position of the target to be estimated is determined, and the coordinates of the center point of the actual detection point of the target to be estimated are converted according to the update point as the geometric center position.

[0136] A target position estimation method provided in this embodiment can effectively improve the accuracy of geometric center position calculation by combining the heading angle and size information of the target fitting frame, optimize the calculation process, reduce unnecessary complexity, and achieve the effect of improving the timeliness and accuracy of the target position estimation results.

[0137] In one embodiment, determining the geometric center position of the target to be estimated based on the update point includes:

[0138] Obtaining the actual length and actual width of the target to be estimated;

[0139] Based on the update point, the actual length and the actual width, determining a distance difference between the update point and a geometric center position in a coordinate system corresponding to the target fitting frame;

[0140] Based on the distance difference, determining a fitting center position of the target to be estimated in a coordinate system corresponding to the target fitting frame;

[0141] Based on the heading angle of the fitting frame, the fitting center position is rotated to obtain the geometric center position of the target to be estimated in the preset coordinate system.

[0142] The actual length and actual width of the target to be estimated can be provided based on prior knowledge. For example, in addition to obtaining the detected size of the target to be estimated by radar detection, the type or model of the target to be estimated can also be obtained by user input, image recognition, wireless communication with the target to be estimated, etc., and the actual length and actual width of the target of the corresponding type or model in the database can be obtained, so as to determine the actual length and actual width of the target to be estimated.

[0143] Based on the update point, the actual length and the actual width, the distance difference between the update point and the geometric center position in the coordinate system corresponding to the target fitting frame is determined. It can be understood that in the coordinate system corresponding to the target fitting frame, when the actual length and the actual width are determined, that is, the occupied range of the actual detection frame is determined, then the distance between the corner point or side center point on the actual detection frame and the center point of the actual detection frame, that is, the geometric center position, can be determined.

[0144] In this embodiment, the known update point is the left front corner point, the right front corner point or the front side center corner point, so the relative distance between the update point and the geometric center position can be calculated based on the position of the update point in the actual detection frame.

[0145] Based on the distance difference, the fitting center position of the target to be estimated in the coordinate system corresponding to the target fitting frame is determined. This can be done by calculating the coordinates of the update point in the fitting frame coordinate system and then calculating the geometric center position of the target to be estimated in the fitting frame coordinate system based on the update point coordinates and the distance difference, which is recorded as the fitting center position.

[0146] The fitting center position is rotated to convert it from the fitting frame coordinate system to the preset coordinate system, so that the geometric center position of the target to be estimated in the preset coordinate system can be obtained.

[0147] A target position estimation method provided in this embodiment effectively improves the accuracy of geometric center position calculation, optimizes the calculation process, reduces unnecessary complexity, and achieves a more accurate assessment of the position of the target to be estimated, thereby achieving the effect of improving the timeliness and accuracy of the target position estimation results.

[0148] In order to more clearly illustrate the technical solution of the present application, the present application also provides a detailed embodiment.

[0149] In one embodiment, a target position estimation method is provided. Taking the method applied to a vehicle end as an example, the vehicle end includes at least two rear corner radar modules, and the detection ranges of the at least two rear corner radar modules overlap.

[0150] In the intelligent driving assistance system of automobiles, some auxiliary functions such as rear collision warning, door opening warning and lane change assistance have high requirements for the accuracy of the target position detected by the rear vehicle and the horizontal overlap length of the detected target with the vehicle. In the overlapping area of ​​the left and right corner radars, the detection of the same target by the two radars is usually different, and the data of only one corner radar cannot meet the back-end requirements for the accuracy of the detected target position. In actual engineering applications, target detection accuracy, computing resources and costs need to be taken into consideration. Traditional solutions use images or lidar data to obtain accurate target position information, but such solutions have complex system architecture, high cost and high computing power. In some solutions, all millimeter-wave radar point cloud data are collected together for calculation. Although this method can realize the fusion of multiple millimeter-wave radar data from the point cloud level to jointly estimate the detection target position, it requires the transmission of a large amount of point cloud data, which puts a lot of pressure on data transmission and consumes a lot of computing power, affecting the real-time and integrity of the data.

[0151] This embodiment provides a method for estimating the target position, which connects the target-level output data of the backward millimeter-wave radar to the back-end fusion processing module, combines the detection position of the same target output by two millimeter-wave radars in the calculation, detects the heading angle and the detection size, etc., and makes an accurate estimate of the position of the detected target. It only needs to install the millimeter-wave radar, and does not need to use image or laser radar data. It is low-cost and can be applied to low-cost solutions; and it only needs to use the target output of the millimeter-wave radar, and does not need point cloud data, which has little pressure on system data transmission and low computing power requirements. In addition, by summarizing the output of millimeter-wave radar detection, it is not only highly flexible, but also can be quickly adapted to mass production projects with different configurations.

[0152] In this embodiment, if Figure 3 As shown, the target position estimation method includes:

[0153] Step S1, define a preset coordinate system, that is, define the ego vehicle coordinate system (ego_coord): establish a coordinate system with the center point of the rear axle of the ego vehicle as the origin, the direction of the front of the ego vehicle as the positive direction of the x-axis, and the left direction of the ego vehicle when the ego vehicle faces forward as the positive direction of the y-axis. This embodiment uses a two-dimensional coordinate system.

[0154] In step S2, the rear corner radar module (or corner radar perception module) outputs the target detection information to the back-end fusion processing module (or multi-sensor fusion module), wherein the target detection information includes the detection position, detection heading angle and detection size of the target. The target center point position is used as the geometric center of the target detection frame, and the detection size is used as the length and width of the target detection frame. Then, the orientation of the detection frame is rotated to the detection heading angle direction, so that the target detection frame of the target to be estimated in the vehicle coordinate system output by the corner radar perception module can be restored through the above target detection information.

[0155] For the target detection frame, the center point position p0 (x0, y0) of the detection frame is needed. Common corner radar products on the market usually output the target detection point not as the center point of the detection frame, but as the closest point of the detection frame to the origin of the vehicle coordinate system, or as the center point of the closest edge of the detection frame to the origin of the vehicle coordinate system, collectively referred to as the anchor point p (x, y). The distance difference (dx, dy) between the anchor point and the center point of the detection frame in the vehicle coordinate system is also output. At this time, the anchor point needs to be converted to the center point of the detection frame first:

[0156] x0=x-dx

[0157] y0=y-dy

[0158] Thus, the center point coordinates p0, heading angle yaw, and length and width dimensions Length and Width under the current target detection information can be obtained.

[0159] Step S3, obtain the fusion target of the left rear and right rear corner radars. Traverse the left rear and right rear corner radar detection targets fused in the fusion target, and determine whether the target is in the overlapping area of ​​the rear corner radar detection. If not, the process ends. If so, calculate the coordinates of the four corner points of each corner radar output target frame in the vehicle coordinate system. For each corner radar detection target traversed, calculate the position coordinates of the four corner points of the detection frame in the vehicle coordinate system, such as Figure 4 As shown, p0 is the center point of the detection frame, and the four corner points are the left front, right front, right back and left back corner points of the detection frame. The calculation process is as follows:

[0160] For each corner radar detection target, first define a detection frame coordinate system (obj_coord), with the midpoint of the rear axle of the vehicle as the origin, that is, the same origin as the origin of the vehicle coordinate system, the direction of the detected heading angle as the positive direction of the x-axis, and the direction rotated 90 degrees counterclockwise from the x-axis in the two-dimensional plane as the positive direction of the y-axis, that is, the left side of the target to be estimated; in the detection frame coordinate system, calculate the first distance difference from each corner point to the center point of the detection frame, and the distance difference calculation formula is:

[0161]

[0162] Among them, Δp obj_coord It is the distance difference between the target point target and the reference point base in the current coordinate system. x and sign y They are used to represent the relative position relationship (not the relative distance) between a point within the detection frame and the center point of the detection frame. Length and Width are the length and width of the detection frame, respectively. x (target) -sign x (base) It can be used to indicate the horizontal relative position of the target point target and the reference point base. y (target) -sign y (base) It can be used to indicate the vertical relative position of the target point target and the reference point base. x and sign y The value is {0, 1, -1}, as shown in the following table:

[0163] Table 1 Sign of target point and detection box center point x and sign y Value table

[0164] Point location <![CDATA[sign x ]]> <![CDATA[sign y ]]> left front corner point 1 1 Right front corner 1 -1 Back right corner point -1 -1 Left rear corner -1 1 Front side center point 1 0 Right side center point 0 -1 Rear side center point -1 0 Left side center point 0 1 Center Point 0 0

[0165] Through the distance difference calculation formula, the target point target can be the corner point of the detection frame, and the reference point base can be the center point of the detection frame. The distance difference between the four corner points and the center point of the detection frame can be obtained, that is, the first distance difference. Suppose a corner point is p A (x A ,y A ), since the sign value of the center point is 0, it can be ignored. Therefore, the calculation formula can be:

[0166]

[0167] The calculated distance difference Δp in the detection frame coordinate system 0A，obj_coord , rotate to the vehicle coordinate system: convert the relative coordinates of the distance difference of the points in the detection frame coordinate system to the coordinate system in the vehicle coordinate system. Since the origin is the same, the translation amount is 0, and the rotation calculation formula is as follows:

[0168] p'=R θ ·p

[0169] Among them, R θ is the rotation matrix, the value of θ is the opposite of the heading angle yaw -yaw:

[0170]

[0171] It can be obtained that the second distance difference Δp 0A，ego_coord =R (-yaw) ·Δp 0A，obj_coord , and then we can get the position of each detection box corner point in the vehicle coordinate system as:

[0172] p A =p0+Δp 0A，ego_coord

[0173] =p0+R (-yaw) ·Δp 0A，obj_coord

[0174] Step S4, convert the coordinates of the four corner points of each corner radar output frame from the vehicle coordinate system to the fitting frame coordinate system. For each corner radar detection target fused in the fusion target, calculate the positions of the four detection frame corner points in turn. After traversing all corner radar detection targets, calculate the average heading angle of the left rear and right rear corner radar detection targets fused in the fusion target. Then define a coordinate system, which is used to calculate a new target fitting box (Fitting Box), so it is named the fitting box coordinate system (FB_coord). The coordinate system is defined as the center of the rear axle of the vehicle as the origin and the average heading angle The direction is the positive direction of the x-axis, and the direction rotated 90 degrees counterclockwise from the x-axis in the two-dimensional plane is the positive direction of the y-axis; after defining the above coordinate system, all the corner points calculated in the previous step are rotated to the fitting frame coordinate system according to the same rotation method as the above embodiment, and p is obtained. A，FB_coord , according to the rotation calculation formula: p A，FB_coord =R θ ·p A , where θ is the average value of the detection heading angle yaw of multiple target detection information but:

[0175]

[0176] Find the corner points with the maximum, minimum, maximum and minimum y values ​​in the fitting frame coordinate system, and fit the new target frame with the four corner points, that is, the target fitting frame. Specifically, the corner point with the maximum x value, the minimum x value, the maximum y value and the minimum y value in the fitting frame coordinate system can be found among all the corner points, and a new detection frame can be fitted using the found points.

[0177] Take a fusion target that combines one left rear corner radar point and one right rear corner radar point as an example. Figure 5 As shown, P01 is the target detection of the left rear corner radar, and P02 is the target detection of the right rear corner radar. Assuming yaw2 is slightly smaller than yaw1, it can be obtained that corner point p1 (x1, y1) and corner point p4 (x4, y4) are both corner points with the largest x value, corner point p2 (x2, y2) and corner point p3 (x3, y3) are both corner points with the largest y value, corner point p4 (x4, y4) is the corner point with the smallest y value, and corner point p2 (x2, y2) is the corner point with the smallest x value.

[0178] Using the corner points p1, p2, p3 and p4 calculated above, a target fitting box can be obtained, such as Figure 6 As shown, p 0_FB is the center point of the new target fitting frame, and the heading angle of the target fitting frame is In the fitting frame coordinate system, the center point p 0_FB The coordinates in the fitting frame coordinate system FB_coord are:

[0179]

[0180] The newly obtained target fitting frame is used as the joint detection of the left and right rear corner radars on the same target, and the measurement value of the radar on the target position detection is obtained with this new target fitting frame.

[0181] Step S5, using the target fitting frame, calculates the radar measurement value of the target detection position.

[0182] In the newly obtained target fitting box, except for the center point p 0_FB In addition, there are 4 corner points and 4 side center points, such as Figure 7 As shown, they are the left front corner point p FLe , right front corner point p FRi , right rear corner point p RRi , the left rear corner point p RLe , the center point of the front side p FC , the center point p of the right side CR , the center point of the rear side p RC and the left side center point p CL , naming these points as fitting box reference points.

[0183] Furthermore, one of the reference points of the fitting frame can be selected, and it is considered that this point is the most accurate point for the radar to jointly detect the target. The selection method includes:

[0184] According to the detection characteristics of millimeter-wave radar, the detection of target vehicles by millimeter-wave radar is usually more accurate for the side position detection of the target close to the radar side. In other words, for the rear-facing vehicle target located in the overlapping area of ​​the rear corner radar detection, the position detection of the front side of the vehicle is usually accurate. Therefore, for the target fitting frame, the center point of the front side, the left front corner point or the right front corner point of the detection frame is the most accurate point for detection. Select one of the points and use this point as the target point to calculate the position of the actual target geometric center point as the estimated position of the target. For example, if the left front corner point of the fitting detection frame is selected, it is considered that the left front corner point of the actual target vehicle coincides with this point, and then based on the unchanged position of this point, the geometric center point of the actual target is calculated as the estimated position of the target, and this value can be used to track and update the target state in the tracking algorithm later. The selected reference point p ref，FB_coord , recorded as the update point. The method of selecting the update point is as follows:

[0185] Calculate the angle α between the center of the front side of the new fitting frame and the line connecting the origin of the vehicle and the heading direction of the new fitting frame. Connect the center point p of the front side of the fitting detection frame FC and the midpoint O of the rear axle of the vehicle, set as line segment P FC O, from the center point p of the front side of the fitting detection box FC The heading angle of the fitting frame Direction is ray, set as P FC H, the heading angle of the target fitting detection frame With P FC The angle ∠OP of O FC H is α, counterclockwise is positive, such as Figure 8 shown.

[0186] When the absolute value of the angle α is small, it is considered that the target vehicle is close to moving towards the self-vehicle, and the corner radar only has a good detection of the front side of the target vehicle. At this time, the center point of the front side of the fitting box is considered to be accurate; when the absolute value of the angle α exceeds the preset threshold α0, it is considered that the corner radar has a good detection of the front side and the corresponding left or right side of the target vehicle. At this time, the corresponding left or right front corner point of the fitting box is considered to be accurate. The value of the preset threshold α0 can be determined by experimental data, and the selection method of the update point is shown in the following table:

[0187] Table 2 Update point selection strategy

[0188] Update Point Angle α condition Front side center point <![CDATA[|α|<α0]]> left front corner point <![CDATA[α<=-α0]]> Right front corner <![CDATA[α>=α0]]>

[0189] That is to say, when the target vehicle is in a different position and moving direction relative to the self-vehicle, the selected update points will be different.

[0190] It can be that first it is determined whether the angle |α|<α0 holds. If so, the center point of the front side surface of the fitting box is selected as the update point. If not, it is determined whether the angle α>=α0 holds. If so, the right front corner point of the fitting box is selected as the update point. If not, the left front corner point of the fitting box is selected as the update point.

[0191] like Fig. 9 As shown in the figure, when the angle |α|<α0, the center point of the front side of the fitting frame is selected as the update point, that is, the center point of the front side of the actual target coincides with the center point of the front side of the fitting frame, p obj_FB is the geometric center point of the actual target; Fig.10 As shown in the figure, when the angle α>=α0, the right front corner of the fitting frame is selected as the update point, that is, the right front corner of the actual target coincides with the right front corner of the fitting frame. The above point O is the center point of the rear axle of the vehicle. The actual position of point O is far away, and the length of the line connecting the center point of the front side of the fitting frame and the center point of the rear axle of the vehicle is much larger than the size of the fitting frame.

[0192] Step S6: Based on the selected update point, the actual target frame can be constructed with the update point. The geometric center position of the actual target frame is calculated. First, in the fitting frame coordinate system, the center point p of the fitting frame obtained in step S4 is obtained. 0_FB，FB_coord Calculate the selected update point p ref,FB_coord The coordinates of the actual target frame are calculated, and then the geometric center point position is calculated, which is recorded as p obj,FB_coordo

[0193] In the fitting frame coordinate system, calculate the updated point coordinates. The selected reference point p can be calculated by the distance difference calculation formula ref，FB_coord With the center point p 0,FB_coord The distance difference Δp in the fitting frame coordinate system 0-ref，FB_coordAt this time, the length of the fitting box in the distance difference calculation formula is Length = (x1-x2), and the width of the fitting box is Width = (y3-y4), then:

[0194]

[0195] It can be seen that in the fitting frame coordinate system, the coordinates of the selected reference point can be completely calculated from x1, x2, y3 and y4.

[0196] In the fitting frame coordinate system, the coordinates of the actual target geometric center are calculated and converted to the vehicle coordinate system. Further, the coordinates of the update point can be used to calculate the coordinates of the geometric center of the actual target frame, that is, the geometric center position, and the distance difference between the update point coordinates and the geometric center coordinates in the fitting frame coordinate system, that is, the third distance difference is Δp ref-obj，FB_coord , can be expressed as:

[0197]

[0198] Among them, Length obj and Width obj are the actual length and width of the target, respectively. The actual length and width can be obtained through user input, wireless communication, image recognition, etc. Thus, the coordinates of the target geometric center position in the fitting frame coordinate system can be obtained as:

[0199]

[0200] The target geometric center point coordinates are transformed from the fitting frame coordinate system to the vehicle coordinate system through the rotation calculation formula, where the rotation matrix R θ The value of θ is the heading angle of the fitting frame The opposite of - but:

[0201]

[0202] Thus, the target geometric center position p is calculated obj，ego_coord As an estimated value of the target position, the target position can be subsequently filtered and updated through filtering update methods such as Kalman filtering.

[0203] In a specific embodiment, in the implementation scenario where the driving vehicle is stationary and the target vehicle behind is approaching the driving vehicle at a certain speed, such as Fig.11 The figure shows the estimated result of the target's oblique position. Fig.12The figure shows the result of the target's straight-line position estimation. The target is within the detection angle of the left rear corner radar and the right rear corner radar. The left rear corner radar detection target box is orange, and the right rear corner radar detection target box is magenta. There are differences between the two corner radar detection positions and boxes. After the implementation of the present invention, the two corner radars with different detections are fitted into a fitting box, as shown in the blue box in the figure. According to the reference point selection method, the nearest center point of the fused target box is finally output as the position of the target vehicle.

[0204] A target position estimation method provided by the present embodiment establishes a fitting frame coordinate system, and determines multiple fitting frame target points based on multiple detection frame target points, thereby determining a target fitting frame in a preset coordinate system, which can achieve high-precision and efficient fitting frame calculation, avoid direct processing of a large amount of point cloud data, and achieve the effect of improving the timeliness and accuracy of the target position estimation result; by determining the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate to determine the boundary of the target fitting frame, it can ensure that the target fitting frame tightly wraps all the detection frame target points, while avoiding unnecessary redundant space, simplifying the calculation process, reducing processing time and resource consumption, and achieving the effect of improving the timeliness and accuracy of the target position estimation result; by combining the heading angle and size information of the target fitting frame, the accuracy of the geometric center position calculation can be effectively improved, the calculation process is optimized, and the reduction Unnecessary complexity can achieve the effect of improving the timeliness and accuracy of the target position estimation results; by obtaining the actual size information and coordinate transformation of the target to be estimated, the accuracy of the geometric center position calculation is effectively improved, the calculation process is optimized, and unnecessary complexity is reduced, so that a more accurate evaluation of the position of the target to be estimated can be achieved, which can achieve the effect of improving the timeliness and accuracy of the target position estimation results; the target-level output data of the backward millimeter-wave radar is connected to the back-end fusion processing module, and the detection position, heading angle and detection size of the same target output of two millimeter-wave radars are combined in the calculation to make an accurate estimate of the position of the detected target. Only the millimeter-wave radar needs to be installed, and no image or lidar data is needed. The cost is low and it can be applied to low-cost solutions; and only the target output of the millimeter-wave radar is needed, and no point cloud data is needed, which has little pressure on the system data transmission and low requirements on computing power. In addition, by summarizing the output of millimeter-wave radar detection, it is not only highly flexible, but also can be quickly adapted to mass production projects with different configurations.

[0205] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0206] Based on the same inventive concept, the embodiment of the present application also provides a target position estimation device for implementing the target position estimation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more target position estimation device embodiments provided below can refer to the limitations of the target position estimation method above, and will not be repeated here.

[0207] In one embodiment, Fig.13 As shown, a target position estimation device is provided, including a position estimation component 100, and at least two radar detection modules 200 whose detection ranges are fully or partially overlapped, wherein the position estimation component 100 includes a detection frame determination module 110, a fitting frame determination module 120 and a position estimation module 130, wherein:

[0208] A detection frame determination module 110 is used to determine a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated; the target detection information includes a detection position, a detection heading angle, and a detection size;

[0209] A fitting frame determination module 120, configured to perform fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated;

[0210] The position estimation module 130 is used to determine the position estimation result of the target to be estimated based on the target fitting frame.

[0211] In some embodiments, the detection frame determination module 110 is further configured to:

[0212] Determining the length and width of the target detection frame based on the detection size of the target detection information;

[0213] Based on the detected position, determining the coordinates of the center point of the target detection frame in a preset coordinate system;

[0214] Based on the length and width, the center point coordinates and the heading angle, a target detection frame corresponding to the current target detection information in the preset coordinate system is determined.

[0215] In some embodiments, the target detection frame includes a plurality of detection frame target points; the detection frame target points include detection frame corner points; and the detection frame determination module 110 is further configured to:

[0216] Based on the heading angle and the preset coordinate point, a detection frame coordinate system is determined; the preset coordinate point is the origin of the preset coordinate system;

[0217] Based on the relative position of each detection frame target point and the center point and the length and width, determine a first distance difference between each detection frame target point and the center point in the detection frame coordinate system;

[0218] Based on the heading angle, the first distance difference is rotated to obtain a second distance difference between the target point of each detection frame and the center point in a preset coordinate system;

[0219] Based on the second distance difference of the plurality of detection frame target points and the center point coordinates, determining the coordinates of the plurality of detection frame corner points corresponding to the current target detection information in the preset coordinate system;

[0220] The target detection frame is determined based on the coordinates of the multiple detection frame corner points.

[0221] In some embodiments, the detection frame determination module 110 is further configured to:

[0222] Based on the heading angle, determining a rotation matrix corresponding to the heading angle;

[0223] A matrix multiplication operation is performed based on the rotation matrix and the first distance difference to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system.

[0224] In some embodiments, the fitting frame determination module 120 is further configured to:

[0225] Determining the heading angle of the target fitting frame based on the heading angles of at least two targets to be estimated;

[0226] Determine a fitting frame coordinate system based on the heading angle of the target fitting frame and preset coordinate points;

[0227] Respectively calculating the coordinates of a plurality of detection frame target points of each of the target detection frames in the fitting frame coordinate system; the detection frame target points include detection frame corner points;

[0228] Based on the multiple detection frame target points, determine multiple fitting frame target points of the target fitting frame;

[0229] Based on the multiple fitting box target points, a target fitting box of the target to be estimated is determined.

[0230] In some embodiments, the fitting frame determination module 120 is further configured to:

[0231] Based on the horizontal coordinates and vertical coordinates of the plurality of detection frame target points, respectively determine the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate of the target to be estimated in the fitting frame coordinate system;

[0232] Based on the maximum abscissa, the minimum abscissa, the maximum ordinate and the minimum ordinate, a plurality of the fitting frame target points of the target fitting frame are determined.

[0233] In some embodiments, the position estimation result includes a geometric center position, and the position estimation module 130 is further used to:

[0234] Based on the heading angle of the target fitting frame, determining an update point of the target to be estimated; the update point includes one of a left front corner point, a right front corner point and a center point of the front side surface;

[0235] Based on the update point, the geometric center position of the target to be estimated is determined.

[0236] In some embodiments, the location estimation module 130 is further configured to:

[0237] Obtaining the actual length and actual width of the target to be estimated;

[0238] Based on the update point, the actual length and the actual width, determining a distance difference between the update point and a geometric center position in a coordinate system corresponding to the target fitting frame;

[0239] Based on the distance difference, determining a fitting center position of the target to be estimated in a coordinate system corresponding to the target fitting frame;

[0240] Based on the heading angle of the fitting frame, the fitting center position is rotated to obtain the geometric center position of the target to be estimated in the preset coordinate system.

[0241] Each module in the above target position estimation method can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0242] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig.14As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a target position estimation method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0243] Those skilled in the art will understand that Fig.14 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0244] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the target position estimation method of any of the above embodiments is implemented:

[0245] Based on at least two target detection information of the target to be estimated, determining a target detection frame corresponding to each of the target detection information in a preset coordinate system; the target detection information includes a detection position, a detection heading angle, and a detection size;

[0246] Perform fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated;

[0247] Based on the target fitting frame, a position estimation result of the target to be estimated is determined.

[0248] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the target position estimation method of any of the above embodiments is implemented:

[0249] Based on at least two target detection information of the target to be estimated, determining a target detection frame corresponding to each of the target detection information in a preset coordinate system; the target detection information includes a detection position, a detection heading angle, and a detection size;

[0250] Perform fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated;

[0251] Based on the target fitting frame, a position estimation result of the target to be estimated is determined.

[0252] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0253] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0254] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0255] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for estimating a target position, characterized in that: Applied to a target position estimation device, the target position estimation device comprises at least two radar detection modules, the detection ranges of the at least two radar detection modules are completely or partially overlapped, and the target position estimation method comprises: Based on at least two target detection information of the target to be estimated, determining a target detection frame corresponding to each of the target detection information in a preset coordinate system; the target detection information includes a detection position, a detection heading angle, and a detection size; Perform fitting based on at least two of the target detection frames to obtain a target fitting frame of the target to be estimated; Based on the target fitting frame, a position estimation result of the target to be estimated is determined.

2. The target position estimation method according to claim 1, characterized in that: The determining, based on at least two pieces of target detection information of the target to be estimated, a target detection frame corresponding to each piece of target detection information in a preset coordinate system comprises: Determining the length and width of the target detection frame based on the detection size of the target detection information; Based on the detected position, determining the coordinates of the center point of the target detection frame in a preset coordinate system; Based on the length and width, the center point coordinates and the heading angle, a target detection frame corresponding to the current target detection information in the preset coordinate system is determined.

3. The target position estimation method according to claim 2, characterized in that: The target detection frame includes a plurality of detection frame target points; the detection frame target points include detection frame corner points; and the determining, based on the length and width, the center point coordinates, and the heading angle, of the target detection frame corresponding to the current target detection information in the preset coordinate system includes: Based on the heading angle and the preset coordinate point, a detection frame coordinate system is determined; the preset coordinate point is the origin of the preset coordinate system; Based on the relative position of each detection frame target point and the center point and the length and width, determine a first distance difference between each detection frame target point and the center point in the detection frame coordinate system; Based on the heading angle, the first distance difference is rotated to obtain a second distance difference between the target point of each detection frame and the center point in a preset coordinate system; Based on the second distance difference of the plurality of detection frame target points and the center point coordinates, determining the coordinates of the plurality of detection frame corner points corresponding to the current target detection information in the preset coordinate system; The target detection frame is determined based on the coordinates of the multiple detection frame corner points.

4. The target position estimation method according to claim 3, characterized in that: The step of rotating the first distance difference based on the heading angle to obtain a second distance difference between the target point of each detection frame and the center point in a preset coordinate system includes: Based on the heading angle, determining a rotation matrix corresponding to the heading angle; A matrix multiplication operation is performed based on the rotation matrix and the first distance difference to obtain a second distance difference between each detection frame target point and the center point in a preset coordinate system.

5. The target position estimation method according to claim 1, characterized in that: The performing fitting based on at least two of the target detection frames to obtain the target fitting frame of the target to be estimated comprises: Determining the heading angle of the target fitting frame based on the heading angles of at least two targets to be estimated; Determine a fitting frame coordinate system based on the heading angle of the target fitting frame and preset coordinate points; Respectively calculating the coordinates of a plurality of detection frame target points of each of the target detection frames in the fitting frame coordinate system; the detection frame target points include detection frame corner points; Based on the multiple detection frame target points, determine multiple fitting frame target points of the target fitting frame; Based on the multiple fitting box target points, a target fitting box of the target to be estimated is determined.

6. The target position estimation method according to claim 5, characterized in that: The determining of a plurality of fitting frame target points of a target fitting frame based on a plurality of detection frame target points comprises: Based on the horizontal coordinates and vertical coordinates of the plurality of detection frame target points, respectively determine the maximum horizontal coordinate, the minimum horizontal coordinate, the maximum vertical coordinate and the minimum vertical coordinate of the target to be estimated in the fitting frame coordinate system; Based on the maximum abscissa, the minimum abscissa, the maximum ordinate and the minimum ordinate, a plurality of the fitting frame target points of the target fitting frame are determined.

7. The target position estimation method according to claim 1, characterized in that: The position estimation result includes a geometric center position, and determining the position estimation result of the target to be estimated based on the target fitting frame includes: Based on the heading angle of the target fitting frame, determining an update point of the target to be estimated; the update point includes one of a left front corner point, a right front corner point and a center point of the front side surface; Based on the update point, the geometric center position of the target to be estimated is determined.

8. The target position estimation method according to claim 7, characterized in that: Determining the geometric center position of the target to be estimated based on the update point includes: Obtaining the actual length and actual width of the target to be estimated; Based on the update point, the actual length and the actual width, determining a distance difference between the update point and a geometric center position in a coordinate system corresponding to the target fitting frame; Based on the distance difference, determining a fitting center position of the target to be estimated in a coordinate system corresponding to the target fitting frame; Based on the heading angle of the fitting frame, the fitting center position is rotated to obtain the geometric center position of the target to be estimated in the preset coordinate system.

9. A target position estimation device, characterized in that: The method comprises a position estimation component and at least two radar detection modules, wherein the detection ranges of the at least two radar detection modules overlap in whole or in part, and the position estimation component comprises a detection frame determination module, a fitting frame determination module and a position estimation module, wherein: A detection frame determination module, configured to determine a target detection frame corresponding to each target detection information in a preset coordinate system based on at least two target detection information of the target to be estimated; the target detection information includes a detection position, a detection heading angle, and a detection size; A fitting frame determination module, used for performing fitting based on at least two target detection frames to obtain a target fitting frame of the target to be estimated; The position estimation module is used to determine the position estimation result of the target to be estimated based on the target fitting frame.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 8 is implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 8 is implemented.

Images