Target detection method and device, equipment and medium

Abstract

The invention discloses a target detection method, apparatus and device, and a medium. The target detection method comprises the steps of obtaining information along a road based on a vehicle-mounted visual sensor, and obtaining an original point set of a static target based on a vehicle-mounted millimeter wave radar; performing first screening processing on an original point set of the static target based on the road line information to obtain an initial point set of the static target; performing second screening processing on the original point set of the static target according to the initial point set of the static target to obtain a final point set of the static target; wherein the roadside line information comprises a first left curve representing a roadside line on the left side of a road and / or a first right curve representing a roadside line on the right side of the road; the static target comprises a road left side road edge target and / or a road right side road edge target. The problem of inaccurate static target detection of the millimeter wave radar is solved, and the problem of inaccurate remote target identification of the visual sensor is also solved.

Description

Technical Field

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

[0002] At present, the market demand for autonomous driving functions is growing. Autonomous driving functions can not only reduce the driving burden of drivers, but also assist drivers in controlling their own vehicles to avoid collisions in critical situations, providing safe and comfortable driving and riding experience for passengers in the car.

[0003] The autonomous driving function is mainly composed of four parts: perception, decision-making, planning and control. The main task of the perception part is to provide stable and accurate target information for the autonomous driving system. In order to ensure that the autonomous driving system can receive stable and accurate target information, the perception module needs to effectively delete static false alarm targets to improve the reliability of the autonomous driving system. However, due to the defects of the millimeter wave radar hardware itself, it is impossible to effectively detect and distinguish the identified static targets, resulting in a high target false alarm rate.

[0004] In view of this, the present invention is proposed. Summary of the invention

[0005] A brief summary of one or more aspects is given below to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all conceived aspects, and is neither intended to identify the key or critical elements of all aspects nor to define the scope of any or all aspects. Its only purpose is to give some concepts of one or more aspects in a simplified form as a prelude to a more detailed description that will be given later.

[0006] The present application provides a target detection method, device and medium, which effectively solves the problem of inaccurate static target detection by millimeter wave radar, and also makes up for the defect of inaccurate recognition of long-distance targets by visual sensors.

[0007] In a first aspect, the present invention provides a target detection method, comprising the following steps:

[0008] Obtain roadside information based on vehicle-mounted visual sensors, and obtain original point sets of static targets based on vehicle-mounted millimeter-wave radars;

[0009] Performing a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target;

[0010] According to the initial point set of the static target, the original point set of the static target is subjected to a second screening process to obtain a final point set of the static target;

[0011] Among them, the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

[0012] Furthermore, the target detection method further comprises: determining a reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road, and / or the first right curve representing the roadside line on the right side of the road;

[0013] The static targets located outside the reference road are deleted.

[0014] Further, the determining of the reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road, and / or the first right curve representing the roadside line on the right side of the road comprises:

[0015] Performing curve fitting according to the final point set of the left side curb target of the road to obtain a second left curve representing the left side curb line of the road;

[0016] A curve fitting is performed based on the final point set of the roadside target on the right side of the road to obtain a second right curve representing the roadside line on the right side of the road

[0017] In response to the lateral distance between the second left curve and the vehicle being greater than the lateral distance between the first left curve and the vehicle being greater, the second left curve is determined as a reference roadside line; in response to the lateral distance between the second right curve and the vehicle being greater than the lateral distance between the first right curve and the vehicle being greater, the second right curve is determined as a reference roadside line.

[0018] Furthermore, deleting the static target outside the reference road includes:

[0019] In response to the reference roadside line including the second left curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the second left curve to obtain a first value; in response to the difference between the y coordinate of the current point and the first value being greater than or equal to zero, delete the current point;

[0020] In response to the reference roadside line including the second right curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the second right curve to obtain a second value; in response to the difference between the y coordinate of the current point and the second value being less than or equal to zero, delete the current point;

[0021] Among them, the origin of the preset coordinate system is the center point of the rear axle of the vehicle, the driving direction of the vehicle is the positive direction of the x-axis, and the left side of the vehicle is the positive direction of the y-axis.

[0022] Further, performing a second screening process on the original point set of the static target according to the initial point set of the static target to obtain the final point set of the static target includes:

[0023] Determine the search direction based on the coordinate position of the point in the initial point set of the target on the left side of the road, and recursively search the points in the original point set of the static target according to the search direction and a preset search range to obtain the final point set of the target on the left side of the road;

[0024] The search direction is determined based on the coordinate position of the points in the initial point set of the roadside target on the right side of the road, and the points in the original point set of the static target are recursively searched according to the search direction and the preset search range to obtain the final point set of the roadside target on the right side of the road.

[0025] Further, the search direction is determined based on the coordinate position of the point in the initial point set of the target on the left side of the road, and the points in the original point set of the static target are recursively searched according to the search direction and a preset search range to obtain the final point set of the target on the left side of the road, including:

[0026] Determine the search azimuth according to the coordinate information of the initial point concentration point of the roadside target on the left side of the road;

[0027] Taking the point in the initial point set that is farthest from the vehicle in the longitudinal direction as a reference point for recursive search, generating a search box of a preset size at the reference point according to the search azimuth, so that the azimuth of the search box is consistent with the search azimuth;

[0028] The point encircled by the search box is used as a point in the final point set of the roadside target on the left side of the road;

[0029] The point in the search box that is closest to the reference point is determined as a new reference point to update the position and azimuth of the search box, and the recursive search is stopped until a preset condition is met.

[0030] Furthermore, the first screening process is performed on the original point set of the static target based on the roadside information to obtain the initial point set of the static target, including:

[0031] Traversing the original point set of the left curb target, or traversing the original point set of the static target, substituting the x coordinate of the current point into the equation of the first left curve to obtain a first result; in response to the absolute value of the difference between the y coordinate of the current point and the first result being less than a preset lateral threshold, determining the current point as a point in the initial point set of the left curb target;

[0032] Traverse the original point set of the right side curb target, or traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the first right curve to obtain a second result; in response to the absolute value of the difference between the y coordinate of the current point and the second result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the right side curb target.

[0033] In a second aspect, the present invention further provides a target detection device, comprising:

[0034] An acquisition module is used to acquire roadside information based on the vehicle-mounted visual sensor, and to acquire the original point set of the static target based on the vehicle-mounted millimeter-wave radar;

[0035] A first processing module, configured to perform a first screening process on the original point set of the static target based on the roadside information to obtain an initial point set of the static target;

[0036] A second processing module, configured to perform a second screening process on the original point set of the static target according to the initial point set of the static target, so as to obtain a final point set of the static target;

[0037] Among them, the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

[0038] In a third aspect, the present invention further provides an electronic device, the electronic device comprising:

[0039] one or more processors;

[0040] A storage device for storing one or more programs;

[0041] When the one or more programs are executed by the one or more processors, the one or more processors implement the target detection method as described above.

[0042] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the target detection method as described above.

[0043] The target detection method disclosed in the present invention integrates the detection information of the visual sensor and the detection information of the millimeter wave radar to achieve efficient distinction and deletion of static false alarm targets. Specifically, the roadside is identified by a camera, and the static target is detected by the millimeter wave radar; then, the roadside and the vehicle motion information are used to classify the nearby millimeter wave radar static false alarm targets to form an initial point set; the millimeter wave radar false alarm targets are searched according to the points in the initial point set and the final point set is generated, so that the static false alarm targets are efficiently distinguished and the target detection accuracy is improved. The problem of inaccurate detection of static targets by millimeter wave radar is solved, and the problem of inaccurate recognition of long-distance targets by visual sensors is also solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for describing the embodiments are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0045] Figure 1 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0046] Figure 2 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0047] Figure 3 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0048] Figure 4 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0049] Figure 5 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0050] Figure 6 A schematic diagram of a recursive search provided by an embodiment of the present invention;

[0051] Figure 7 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0052] Figure 8 A schematic diagram of a flow chart of a target detection method provided by an embodiment of the present invention;

[0053] Fig. 9 A schematic diagram of the structure of a target detection device provided by an embodiment of the present invention;

[0054] Fig.10FIG. 4 is a schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.

[0056] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0057] Figure 1 This is a flow chart of a target detection method proposed in this application. The method is applicable to target classification scenarios. Specifically, it integrates the detection information of visual sensors and the detection information of millimeter-wave radars to achieve efficient distinction of static false alarm targets, improve target detection accuracy, and reduce false alarm targets. Specifically, the target detection method can be executed by a target detection device, which can be implemented in the form of software and / or hardware.

[0058] like Figure 1 As shown, the target detection method comprises the following steps:

[0059] S110, obtaining roadside information of the road based on the vehicle-mounted visual sensor, and obtaining an original point set of a static target based on the vehicle-mounted millimeter-wave radar.

[0060] Among them, the vehicle-mounted visual sensor is, for example, a camera, which can be further divided into monocular cameras, multi-camera cameras, etc. The embodiment of the present application does not limit the type of camera.

[0061] The information collected by the vehicle-mounted visual sensor is usually an image, and image recognition technology can be used to further identify the targets contained in the image, such as curbs, fences, lane lines, trash cans, the vehicle itself, pedestrians, etc.

[0062] In some embodiments, the vehicle-mounted visual sensor may be installed in the middle of the front of the vehicle to photograph the curbs on the left and right sides of the front of the vehicle, or to make the curbs on the left and right sides of the front of the vehicle within the field of view of the visual sensor. Similarly, the vehicle-mounted millimeter-wave radar also refers to a millimeter-wave radar installed in front of the vehicle that can scan the curbs on both sides of the road.

[0063] Exemplarily, the curb line information includes a first left curve representing a curb line on the left side of the road and / or a first right curve representing a curb line on the right side of the road, that is, in one embodiment, the curb line information includes a first left curve representing a curb line on the left side of the road; in another embodiment, the curb line information includes a first right curve representing a curb line on the right side of the road; in yet another embodiment, the curb line information includes a first left curve representing a curb line on the left side of the road and a first right curve representing a curb line on the right side of the road.

[0064] The static target includes a target on the left side of the road (or a fence target) and / or a target on the right side of the road (or a fence target); that is, in one embodiment, the static target includes a target on the left side of the road (or a fence target); in another embodiment, the static target includes a target on the right side of the road (or a fence target); in yet another embodiment, the static target includes a target on the left side of the road (or a fence target) and a target on the right side of the road (or a fence target).

[0065] S120: Perform a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target.

[0066] It is understandable that due to the defects of the millimeter wave radar hardware itself, it is unable to effectively detect and distinguish the identified static targets, resulting in a high target false alarm rate. The visual sensor has a higher detection accuracy for targets within a close range. Based on this, the implementation scheme of the present application integrates the detection results of the two, taking advantage of each other's strengths and weaknesses, and ultimately achieving the goal of improving target detection accuracy.

[0067] In some embodiments, a point in the original point set of the static target that is closer to the first left curve (the curve is detected based on a visual sensor) is determined as a point in the initial point set of the left side roadside target, and a point in the original point set of the static target that is closer to the first right curve is determined as a point in the initial point set of the right side roadside target. In this way, the purpose of performing a first screening of the original point set of the static target detected by the millimeter wave radar based on the roadside line detected by the visual sensor is achieved.

[0068] In other embodiments, the left side roadside target may be separated from the static target first (for example, according to the y coordinate value, the point with the y coordinate greater than zero belongs to the possible point of the left side roadside target, and the point with the y coordinate less than zero belongs to the possible point of the right side roadside target), and then the points in the original point set of the left side roadside target are compared with the first left curve, and the points whose distances meet the threshold are determined as the points in the initial point set of the left side roadside target; similarly, the right side roadside target may be separated from the static target first, and then the points in the original point set of the right side roadside target are compared with the first right curve, and the points whose distances meet the threshold are determined as the points in the initial point set of the right side roadside target. Compared with comparing the points in the original point set of the static target with the first left curve and the first right curve respectively, the amount of calculation can be reduced, thereby improving the processing efficiency.

[0069] S130 , performing a second screening process on the original point set of the static target according to the initial point set of the static target to obtain a final point set of the static target.

[0070] It is understandable that, due to the low detection accuracy of the visual sensor for targets at a long distance, for the roadside, the visual sensor cannot clearly detect the roadside information in the distance. Therefore, the original point set of the roadside target in the distance detected by the millimeter-wave radar can no longer be screened and processed based on the roadside information detected by the visual sensor. To address this problem, in the implementation scheme of the present application, the original point set of the static target detected by the millimeter-wave radar is first screened and processed based on the roadside information detected by the visual sensor to obtain the initial point set of the static target, that is, the nearby millimeter-wave radar static false alarm targets are classified to form an initial point set. Afterwards, the original point set of the static target is screened and processed for the second time based on the points in the initial point set to obtain the final point set of the static target.

[0071] Exemplarily, the method of performing a second screening process on the original point set of the static target can be: determining an initial search direction based on the points in the initial point set, searching the points in the original point set using a search box of a preset size, obtaining new points, and updating the search direction and the position of the search box based on the new points (wherein, the initial position of the search box can be determined based on the point in the initial point set that is farthest from the longitudinal distance of the vehicle, or can be determined by the end point of the first left curve or the first right curve far from the vehicle) so as to continue searching and finally obtain the final point set of the static target. That is, the search direction is calculated according to the position of the point in the initial point set, and the millimeter-wave radar false alarm target is recursively searched using the search direction and the customized search range to generate the final point set of the false alarm target.

[0072] It can be understood that before the first screening process is performed on the original point set of the static target based on the roadside information, and before the second screening process is performed on the original point set of the static target based on the initial point set of the static target, the spatial alignment and temporal alignment operations should be completed, that is, the information detected by the visual sensor and the information detected by the millimeter wave radar should be aligned in time and space to achieve information synchronization.

[0073] Optionally, the camera timestamp is used as a reference to search for the nearest millimeter-wave radar information frame before it. The spatial alignment method is to record the ego vehicle motion coordinate system corresponding to the camera timestamp as e1, and the ego vehicle motion coordinate system corresponding to the millimeter-wave radar timestamp as e2. The ego vehicle motion coordinate system e2 is rotated and translated to align it to the ego vehicle motion coordinate system e1, and the spatial alignment is completed. The time alignment method is to calculate the difference between the camera timestamp and the nearest millimeter-wave radar timestamp, and use the constant rate and angular rate model to align the millimeter-wave radar target prediction to the camera timestamp to complete the time alignment.

[0074] The target detection method disclosed in the embodiment of the present invention integrates the detection information of the visual sensor and the detection information of the millimeter wave radar, realizes the efficient distinction of static false alarm targets, and improves the target detection accuracy. It solves the problem of inaccurate static target detection of the millimeter wave radar, and also solves the problem of inaccurate recognition of long-distance targets by the visual sensor.

[0075] In some optional embodiments, refer to Figure 2 FIG. 1 is a flow chart of a target detection method. Based on the above embodiment, this embodiment continues to process the static targets detected by the millimeter wave radar to further delete the targets that will not affect the driving of the vehicle, thereby reducing the interference to the automatic driving system and improving the safety of automatic driving. Figure 2 As shown, the target detection method comprises the following steps:

[0076] S210: Acquire roadside information of the road based on the vehicle-mounted visual sensor, and acquire an original point set of a static target based on the vehicle-mounted millimeter-wave radar.

[0077] S220: Perform a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target.

[0078] S230 , performing a second screening process on the original point set of the static target according to the initial point set of the static target to obtain a final point set of the static target.

[0079] S240: Determine a reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road, and / or the first right curve representing the roadside line on the right side of the road.

[0080] Specifically, the final point set of the static target includes the final point set of the target on the left side of the road and the final point set of the target on the right side of the road, which are selected from the original point set detected by the millimeter wave radar. Further, curve fitting can be performed based on the final point set of the target on the left side of the road to obtain another curve representing the left side of the road, and the final curve of the left side of the road (i.e., the reference roadside) is determined by using the other curve and the curve of the left side of the road determined by the visual sensor, with the purpose of improving the recognition accuracy of the left side of the road, thereby deleting false alarm targets and reducing the interference caused by false alarm targets to the automatic driving system.

[0081] Similarly, a curve fitting can be performed based on the final point set of the right side roadside target to obtain another curve representing the right side roadside line, and the final curve of the right side roadside line (i.e., the reference roadside line) can be determined by using the other curve and the curve of the right side roadside line determined by the visual sensor. The purpose is to improve the recognition accuracy of the right side roadside line, thereby deleting false alarm targets and reducing the interference caused by false alarm targets to the automatic driving system.

[0082] After determining the reference roadside line, all static targets outside the reference roadside line are deleted. By deleting static targets outside the reference roadside line, the interference of irrelevant targets to the autonomous driving system can be reduced, thereby improving the safety of autonomous driving. The problem of inaccurate classification of static targets by millimeter-wave radar is solved. Not only can the fences and curb points on straight roads and curves be effectively deleted, but also the longitudinally arranged obstacles in the drivable area can be avoided from being mistakenly deleted, which can effectively improve the problem of false alarms of vehicles caused by static fences and curb targets.

[0083] Exemplarily, the determining of the reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road, and / or the first right curve representing the roadside line on the right side of the road includes:

[0084] Performing curve fitting according to the final point set of the left side curb target of the road to obtain a second left curve representing the left side curb line of the road; performing curve fitting according to the final point set of the right side curb target of the road to obtain a second right curve representing the right side curb line of the road;

[0085] In response to the lateral distance between the second left curve and the vehicle being greater than the lateral distance between the first left curve and the vehicle being greater, the second left curve is determined as a reference roadside line; in response to the lateral distance between the second right curve and the vehicle being greater than the lateral distance between the first right curve and the vehicle being greater, the second right curve is determined as a reference roadside line;.

[0086] Exemplarily, in a preset coordinate system, the first left curve, the second left curve, the first right curve and the second right curve can be expressed by the following equation (1), wherein the origin of the preset coordinate system is the center point of the rear axle of the vehicle, the driving direction of the vehicle is the positive direction of the x-axis, and the left is the positive direction of the y-axis.

[0087] y=C0+C1*x+C2*x 2 (1)

[0088] For ease of description, the first left curve is represented by the following equation (2), the first right curve is represented by the following equation (3), the second left curve is represented by the following equation (4), and the second right curve is represented by the following equation (5).

[0089] yleftlane=C0 L1 +C1 L1 *x+C2 L1 *x 2 (2)

[0090] yrightlane=C0 R1 +C1 R1 *x+C2 R1 *x 2 (3)

[0091] yleftCurve=C0 L2 +C1 L2 *x+C2 L2 *x 2 (4)

[0092] yrightCurve=C0 R2 +C1 R2 *x+C2 R2 *x 2 (5)

[0093] Among them, parameter C0 L1 、C1 L1 、C2 L1 、C0 R1 、C1 R1 、C2 R1 All are determined based on the detection information of the visual sensor. Parameter C0 L2 、C1 L2 、C2 L2 、C0 R2 、C1 R2 、C2 C2 All are determined based on the detection information of millimeter wave radar. L1 Indicates the lateral distance between the first left curve and the vehicle, C0 L2Indicates the lateral distance between the second left curve and the vehicle, such as Figure 3 As shown, if C0 L2 Greater than C0 L1 , then the second left curve is determined to be the reference roadside line. R2 Less than C0 R1 , the second right curve is determined to be the reference roadside line. Then the static targets outside the reference roadside line are deleted, which solves the problem of inaccurate classification of static targets by millimeter-wave radar. It can not only effectively delete fences and roadside points on straight roads and curves, but also avoid mistakenly deleting longitudinally arranged obstacles in the drivable area, which can effectively improve the problem of false alarms of vehicles caused by static fences and roadside targets.

[0094] Specifically, if the reference roadside line only includes the second left curve, the static target is deleted based on the second left curve. If the reference roadside line only includes the second right curve, the static target is deleted based on the second right curve. If the reference roadside line includes both the second left curve and the second right curve, the static target is deleted based on the second left curve and the second right curve, respectively.

[0095] In some embodiments, a curve fitting may be performed based on the final point set of the left side roadside target of the road based on the random sampling consensus algorithm RANSAC to obtain a second left curve representing the left side roadside line of the road. A curve fitting may be performed based on the final point set of the right side roadside target of the road based on the random sampling consensus algorithm RANSAC to obtain a second right curve representing the right side roadside line of the road.

[0096] Among them, the principle of curve fitting based on the random sampling consensus algorithm RANSAC is as follows:

[0097] The fitting curve function model is set as the above formula (1), and the solution target is the three parameters C0, C1 and C2 in formula (1). Based on the RANSAC algorithm, three points are randomly sampled in the point set to establish the linear equation system shown in the following expression (6):

[0098]

[0099] Convert (6) into matrix form:

[0100]

[0101] make:

[0102] B=[y 1 y 2 y 3 ] T (8)

[0103]

[0104] C=[C0 C1 C2] T (10)

[0105] Combining expressions (7), (8), (9), and (10), we can get the matrix form of expression (6):

[0106] B=AC (11)

[0107] If the matrix A in expression (11) is invertible, the curve coefficient matrix C can be obtained using expression (11):

[0108] C=A -1 B (12)

[0109] For further reference, see Figure 4 The flow chart of a curve fitting process based on the RANSAC algorithm shown in FIG. 1 specifically includes the following steps:

[0110] S410: Determine whether the total number of points in the point set currently requiring curve fitting is less than a minimum threshold required for fitting.

[0111] If yes, then end the fitting process, otherwise go to the next step.

[0112] S420, determine whether the number of loops j is less than the set value max_iters, if so, go to step S430 to continue the loop, otherwise go to step S480.

[0113] The number of loops j is an integer, and its initial value is 0. It can be understood that during the execution of the program, j is defined first, for example, "int j = 0", and the maximum number of inliers "int max_inliers = 0" is defined at the same time.

[0114] S430, randomly sampling from n segments of sub-point sets to obtain n sampling points, and performing curve fitting; and establishing an equation relationship B=AC based on the sampling points obtained by sampling.

[0115] The point set that needs to be fitted with the curve is divided into n segments to obtain n sub-point sets, and then a point is randomly selected from each segment of the point set, and a total of n points are used for curve fitting.

[0116] Among them, n can be 3, or other empirical values ​​such as 4, in order to improve the quality and efficiency of curve fitting. The point set currently required for curve fitting is divided into n segments, which can be specifically divided according to the coordinate values ​​of the points. For example, the fitted curve in this embodiment is used to represent a roadside line, and the point set can be divided into several sub-point sets along the direction of the roadside line.

[0117] S440, determine whether the matrix A is full rank, if not, then j++, if yes, proceed to step S450.

[0118] S450, using C = A -1 B, calculate matrix C, and compare the product of A and the calculated matrix C with matrix B to check whether the accuracy requirements are met.

[0119] If not, the process proceeds to the next cycle. If satisfied, the process proceeds to step S460.

[0120] S460. Apply the formula to calculate B-AC for all points in the point set. If the result is less than the threshold, the point is considered to be an inlier and the number of inliers is num_inliers++, and the global error residual is calculated. If the number of inliers is greater than the maximum number of inliers max_inliers, or the global error is less than the preset threshold, the matrix C is maintained, and the maximum number of inliers and the global error are updated.

[0121] S470. Determine whether max_inliers is greater than N*95%.

[0122] If no, continue the loop, if yes, go to step S490. N represents the total number of points in the point set. If max_inliers is greater than N*95%, it means that the condition for early termination of the loop is met, which can improve the fitting efficiency. Among them, 95% is an exemplary value, and it can also be 96%, 94%, etc., and it is not limited to it.

[0123] S480. Determine whether max_inliers is less than N*67%.

[0124] If yes, the process ends, otherwise, the process goes to step S490. If max_inliers is less than N*67%, it means that the quality of the currently fitted curve is not high and cannot be used, so the fitting quality can be maintained.

[0125] Here, 67% is an exemplary value, and may also be 68%, 66%, etc., and is not limited thereto.

[0126] S490. Store the points that meet the conditions as internal points.

[0127] S250: Delete the static target outside the reference roadside.

[0128] In some implementations, deleting the static target outside the reference road includes:

[0129] In response to the reference road including the second left curve, in a preset coordinate system, the original point set of the static target is traversed, and the x-coordinate of the current point is substituted into the equation of the second left curve to obtain a first value; in response to the difference between the y-coordinate of the current point and the first value being greater than or equal to zero, the current point is deleted.

[0130] In response to the reference roadside line including the second right curve, in a preset coordinate system, the original point set of the static target is traversed, and the x-coordinate of the current point is substituted into the equation of the second right curve to obtain a second value; in response to the difference between the y-coordinate of the current point and the second value being less than or equal to zero, the current point is deleted.

[0131] Among them, the origin of the preset coordinate system is the center point of the rear axle of the vehicle, the driving direction of the vehicle is the positive direction of the x-axis, and the left side of the vehicle is the positive direction of the y-axis.

[0132] Assume that the equation of the reference roadside line is:

[0133] yleft=C0 L +C1 L *x+C2 L *x 2 (13)

[0134] Substitute the x coordinate of the current point into (13) to obtain the first value, and then use the y coordinate y of the current point radar Subtracting the above value from the above value, we can get ydiffl, which can be expressed as:

[0135]

[0136] If ydiffl is greater than or equal to zero, it means that the current point is outside the reference roadside line, and the current point is deleted.

[0137] The same logic is used, assuming that the equation of the reference roadside line is:

[0138] yright=C0 R +C1 R *x+C2 R *x 2 (15)

[0139] Substitute the x coordinate of the current point into (15) to obtain the second value, and then use the y coordinate y of the current point radar Subtracting the second value from the second value, we get ydiffr, which can be expressed as:

[0140]

[0141] If ydiffr is less than or equal to zero, it means that the current point is outside the reference roadside line, and the current point is deleted.

[0142] The target detection method disclosed in the embodiment of the present invention further determines the reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road and / or the first right curve representing the roadside line on the right side of the road, and deletes the static target located outside the reference roadside line, thereby achieving the purpose of deleting targets that will not affect the driving of the vehicle, thereby reducing interference with the automatic driving system and improving the safety of automatic driving. The problem of inaccurate classification of millimeter-wave radar static targets is solved, and not only can fences and roadside points on straight roads and curves be effectively deleted, but also longitudinally arranged obstacles in the drivable area can be avoided from being mistakenly deleted, which can effectively improve the problem of false alarms of vehicles caused by static fences and roadside targets.

[0143] In some optional embodiments, refer to Figure 5 The flowchart of a target detection method shown in FIG. 1 is based on the above embodiment. This embodiment provides an optional specific implementation method for the above step S130 "performing a second screening process on the original point set of the static target according to the initial point set of the static target to obtain the final point set of the static target". Figure 5 As shown, the target detection method comprises the following steps:

[0144] S510: Acquire roadside information of the road based on the vehicle-mounted visual sensor, and acquire an original point set of a static target based on the vehicle-mounted millimeter-wave radar.

[0145] S520: Perform a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target.

[0146] S530, determine the search direction based on the coordinate position of the points in the initial point set of the target on the left side of the road, recursively search the points in the original point set of the static target according to the search direction and the preset search range, and obtain the final point set of the target on the left side of the road; determine the search direction based on the coordinate position of the points in the initial point set of the target on the right side of the road, recursively search the points in the original point set of the static target according to the search direction and the preset search range, and obtain the final point set of the target on the right side of the road.

[0147] Wherein, taking the final point set of the target on the left side of the road as an example, the search direction is determined based on the coordinate position of the point in the initial point set of the target on the left side of the road, and the points in the original point set of the static target are recursively searched according to the search direction and the preset search range to obtain the final point set of the target on the left side of the road, including:

[0148] The search azimuth is determined according to the coordinate information of a point in the initial point set of the target on the left side of the road; the point in the initial point set that is farthest from the vehicle in the longitudinal direction is used as the reference point for recursive search, and a search box of a preset size is generated at the reference point according to the search azimuth, so that the azimuth of the search box is consistent with the search azimuth; the point encircled by the search box is used as a point in the final point set of the target on the left side of the road; the point in the search box that is closest to the reference point is determined as a new reference point to update the position and azimuth of the search box, and the recursive search is stopped until the preset conditions are met.

[0149] Specifically, Figure 6 The schematic diagram of a recursive search shown in the figure assumes that there are five points in the initial point set of the roadside target on the left side of the road, which are recorded as point 1, point 2, point 3, point 4, and point 5 respectively. The azimuth 1 of point 2 relative to point 1 can be calculated based on the coordinate positions of point 1 and point 2. Similarly, the azimuth 2 of point 3 relative to point 2, the azimuth 3 of point 4 relative to point 3, and the azimuth 4 of point 5 relative to point 4 can be calculated. Then, the average of azimuth 1, azimuth 2, azimuth 3 and azimuth 4 is calculated, and the average is used as the search azimuth. With point 5 as the reference point, a search box of a preset size is generated at point 5, and the azimuth of the search box is made consistent with the search azimuth. The points encircled by the search box are used as points in the final point set of the roadside target on the left side of the road. The point in the search box closest to the reference point is determined as a new reference point, and then the new reference point is used as the position of the search box in the next search, and the coordinate position of the new reference point is used to determine the azimuth 5 of the new reference point relative to point 5, and further, the average of azimuth 1, azimuth 2, azimuth 3, azimuth 4 and azimuth 5 is calculated, and the average is used as the new search azimuth, so as to obtain a new search box, perform a second iterative search, and repeat the above operation until the preset conditions are met, stop the recursive search, and obtain the final point set. The preset conditions are, for example: the number of points in the final point set reaches a threshold, or the number of iterations reaches a preset number, or no new points enter the search box, etc.

[0150] The target detection method disclosed in the embodiment of the present invention provides an optional specific implementation method for the second screening process, determines the search direction with the points in the initial point set, recursively searches the original point set based on the search direction, and obtains the final point set, which makes up for the problem of poor target recognition accuracy of the visual sensor at a long distance, and solves the problem of inaccurate target classification of the millimeter wave radar.

[0151] In some optional embodiments, refer to Figure 7The flowchart of a target detection method shown in FIG. 1 is based on the above embodiment. This embodiment provides an optional specific implementation method for the above step S120 "performing a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target". Figure 7 As shown, the target detection method comprises the following steps:

[0152] S710: Acquire roadside information of the road based on the vehicle-mounted visual sensor, and acquire an original point set of a static target based on the vehicle-mounted millimeter-wave radar.

[0153] Among them, the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

[0154] S720, traverse the original point set of the left side curb target, or traverse the original point set of the static target (the static target includes the left side curb target), substitute the x coordinate of the current point into the equation of the first left curve, and obtain a first result; in response to the absolute value of the difference between the y coordinate of the current point and the first result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the left side curb target of the road; traverse the original point set of the right side curb target, or traverse the original point set of the static target (the static target includes the right side curb target), substitute the x coordinate of the current point into the equation of the first right curve, and obtain a second result; in response to the absolute value of the difference between the y coordinate of the current point and the second result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the right side curb target of the road.

[0155] Specifically, the equation of the first left curve can refer to formula (2), and the x-coordinate of the current point is substituted into the equation of the first left curve to obtain the first result. The difference ydiffl between the y-coordinate of the current point and the first result can be expressed as shown in formula (14). If the absolute value of ydiffl is less than the preset lateral threshold, the current point is determined as a point in the initial point set of the roadside target on the left side of the road.

[0156] The equation of the first right curve can refer to formula (3). The x coordinate of the current point is substituted into the equation of the first right curve to obtain the second result. The difference ydiffr between the y coordinate of the current point and the second result can be expressed as shown in formula (16). If the absolute value of ydiffr is less than the preset lateral threshold, the current point is determined as a point in the initial point set of the roadside target on the right side of the road.

[0157] S730: Perform a second screening process on the original point set of the static target according to the initial point set of the static target to obtain a final point set of the static target.

[0158] In some embodiments, reference is made to Figure 8 The flowchart of a target detection method shown in FIG. 1 specifically includes the following steps:

[0159] S810: Acquire forward roadside line and lane line information, acquire forward radar target information, and acquire vehicle motion information.

[0160] Specifically, the front view roadside and lane line information is obtained through the front view camera, the target information is obtained through the millimeter wave radar, and the vehicle motion information, such as the vehicle speed, acceleration, etc., is obtained through other relevant sensors on the vehicle.

[0161] S820, on-board forward-looking and front radar temporal and spatial alignment.

[0162] Specifically, the detection data of the vehicle's front-view camera is time-aligned and space-aligned with the detection data of the millimeter-wave radar installed on the front of the vehicle.

[0163] S830: Construct a search anchor using the front-view roadside line and lane line.

[0164] For example, taking the left side of the road as an example, the left side of the road can be expressed by formula (2). When the road side is complete, the parameter C0 in formula (2) is L1 、C1 L1 、C2 L1 The parameters C0 are determined based on the image information of the roadside. If the roadside is incomplete, the parameter C0 can be determined based on the image information of the limited roadside. L1 , and then determine the parameter C1 based on the image information of the lane line that is closest to the roadside in lateral distance L1 and C2 L1 When determining the parameter C0 L1 、C1 L1 、C2 L1 Then, the curve equation of the left side road is obtained, and the curve equation is the search anchor.

[0165] S840: Based on the search anchor, quickly separate a point set with a high possibility of being a false alarm target from the front radar target point set.

[0166] The essence of this step is: traverse the original point set of the left side roadside target, or traverse the original point set of the static target (static target includes the left side roadside target), substitute the x coordinate of the current point into the equation of the first left curve to obtain a first result; in response to the absolute value of the difference between the y coordinate of the current point and the first result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the left side roadside target; traverse the original point set of the right side roadside target, or traverse the original point set of the static target (static target includes the right side roadside target), substitute the x coordinate of the current point into the equation of the first right curve to obtain a second result; in response to the absolute value of the difference between the y coordinate of the current point and the second result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the right side roadside target.

[0167] The initial point set is a point set with a high possibility of false alarm targets.

[0168] S850: Use improved RANSAC to fit a point set with a high probability of static false alarm targets.

[0169] The specific process of this step can be referred to as Figure 4 shown.

[0170] S860. Obtain a point set of front radar target points that are actually static false alarm targets.

[0171] That is, the final curve of the left side roadside and the final curve of the right side roadside are obtained, and the area between the final curve of the left side roadside and the final curve of the right side roadside is determined as the drivable area. The points falling within the drivable area are also points that are truly static false alarm targets, and the points falling outside the drivable area are deleted as points that are not real false alarm targets.

[0172] S870, obtaining a target point set of static false alarm targets in the radar before filtering out.

[0173] In summary, the technical solution of the present application integrates the information of the camera and the millimeter-wave radar to achieve efficient distinction and deletion of static false alarm targets. Specifically, the roadside and lane lines are identified by the camera, and the static targets are detected by the millimeter-wave radar; then, the roadside, lane lines and vehicle motion information are used to classify the nearby millimeter-wave radar static false alarm targets to form an initial point set; the search direction is calculated according to the position of the point in the initial point set, and the millimeter-wave radar false alarm targets are recursively searched using the search direction and the customized search range to generate the final false alarm target point set; further, the generated point sets are fitted respectively using the RANSAC algorithm to obtain a fitting curve, and the static false alarm targets are distinguished and deleted using the fitting curve. The fitted curve can make up for the problem of inaccurate camera recognition at a long distance, and at the same time make up for the problem of inaccurate classification of static targets by the millimeter-wave radar, and the system development and use costs are low. The invention can not only effectively delete fences and roadside points on straight roads and curves, but also avoid mistaken deletion of longitudinally arranged obstacles in the drivable area, and can effectively improve the problem of vehicle false alarms caused by static fences and roadside targets.

[0174] Fig. 9 Schematic diagram of a target detection device provided by an embodiment of the present invention. Fig. 9 As shown, the device includes: an acquisition module 910, a first processing module 920 and a second processing module 930.

[0175] Among them, the acquisition module 910 is used to obtain the roadside information of the road based on the vehicle-mounted visual sensor, and to obtain the original point set of the static target based on the vehicle-mounted millimeter-wave radar; the first processing module 920 is used to perform a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target; the second processing module 930 is used to perform a second screening process on the original point set of the static target according to the initial point set of the static target to obtain the final point set of the static target; wherein the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

[0176] Furthermore, it also includes: a determination module and a deletion module, the determination module is used to determine the reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road and / or the first right curve representing the roadside line on the right side of the road; the deletion module is used to delete the static target located outside the reference roadside line.

[0177] Furthermore, the determination module includes a fitting unit, which is used to perform curve fitting according to the final point set of the left side curb target of the road to obtain a second left curve representing the left side curb line of the road; perform curve fitting according to the final point set of the right side curb target of the road to obtain a second right curve representing the right side curb line of the road; and a determination unit, which is used to determine a reference curb line based on the first left curve, the second left curve, the first right curve and the second right curve.

[0178] Furthermore, the determination unit is specifically used to: in response to the lateral distance between the second left curve and the own vehicle being greater than the lateral distance between the first left curve and the own vehicle, determine the second left curve as a reference roadside line; in response to the lateral distance between the second right curve and the own vehicle being greater than the lateral distance between the first right curve and the own vehicle, determine the second right curve as a reference roadside line; in response to the lateral distance between the second left curve and the own vehicle being greater than the lateral distance between the first left curve and the own vehicle, determine the second left curve as a left-side final curve; in response to the lateral distance between the second right curve and the own vehicle being greater than the lateral distance between the first right curve and the own vehicle, determine the second right curve as a right-side final curve;.

[0179] Furthermore, the deletion module is specifically used to: in response to the reference roadside line including the second left curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x-coordinate of the current point into the equation of the second left curve, and obtain a first value; in response to the difference between the y-coordinate of the current point and the first value being greater than or equal to zero, delete the current point; in response to the reference roadside line including the second right curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x-coordinate of the current point into the equation of the second right curve, and obtain a second value; in response to the difference between the y-coordinate of the current point and the second value being less than or equal to zero, delete the current point; wherein the origin of the preset coordinate system is the center point of the rear axle of the vehicle, the driving direction of the vehicle is the positive direction of the x-axis, and the left side of the vehicle is the positive square of the y-axis.

[0180] Furthermore, the second processing module 920 includes a search unit, which is used to determine the search direction based on the coordinate position of the points in the initial point set of the target on the left side of the road, and recursively search the points in the original point set of the static target according to the search direction and a preset search range to obtain the final point set of the target on the left side of the road; determine the search direction based on the coordinate position of the points in the initial point set of the target on the right side of the road, and recursively search the points in the original point set of the static target according to the search direction and a preset search range to obtain the final point set of the target on the right side of the road.

[0181] Furthermore, the search unit is specifically used to: determine the search azimuth according to the coordinate information of a point in the initial point set of the target on the left side of the road; use the point in the initial point set that is farthest from the vehicle in the longitudinal direction as the reference point for recursive search, and generate a search box of a preset size at the reference point according to the search azimuth, so that the azimuth of the search box is consistent with the search azimuth; use the point circled by the search box as the point in the final point set of the target on the left side of the road; determine the point in the search box that is closest to the reference point as the new reference point to update the position and azimuth of the search box, and stop the recursive search until the preset conditions are met.

[0182] Furthermore, the first processing module 910 is specifically used to: traverse the original point set of the left side curb target, or traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the first left curve to obtain a first result; in response to the absolute value of the difference between the y coordinate of the current point and the first result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the left side curb target of the road; traverse the original point set of the right side curb target, or traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the first right curve to obtain a second result; in response to the absolute value of the difference between the y coordinate of the current point and the second result being less than a preset lateral threshold, determine the current point as a point in the initial point set of the right side curb target of the road.

[0183] The target detection device provided in the embodiment of the present disclosure can execute the steps in the target detection method provided in the method embodiment of the present disclosure, and the execution steps and beneficial effects are no longer repeated here.

[0184] Fig.10 Schematic diagram of the structure of an electronic device in the embodiment of the present disclosure. Fig.10 , which shows a structural schematic diagram of an electronic device 500 suitable for implementing the embodiments of the present disclosure. Fig.10 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0185] like Fig.10As shown, the electronic device 500 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) or a program loaded from a storage device 508 into a random access memory (RAM) to implement the method of the embodiment described in the present disclosure. In RAM503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, ROM 502, and RAM503 are connected to each other via a bus 504. An I / O interface 505 is also connected to the bus 504. An input device 506, an output device 507, a storage device 508, and a communication device 509 are all connected to the I / O interface 505.

[0186] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart, thereby implementing the target detection method as described above. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-mentioned functions defined in the method of the embodiment of the present disclosure are executed.

[0187] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0188] The computer-readable medium may be included in the electronic device, or may exist independently without being installed in the electronic device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device executes the method steps in the present application.

[0189] Optionally, when the above one or more programs are executed by the electronic device, the electronic device may also execute other steps described in the above embodiments.

[0190] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0191] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0192] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and its core ideas of this application. The above is only the preferred implementation method of this application. It should be pointed out that due to the limitations of textual expression and the objective existence of infinite specific structures, ordinary technicians in this technical field can make several improvements, modifications or changes without departing from the principles of the present invention, and can also combine the above technical features in an appropriate manner; these improvements, modifications, changes or combinations, or the direct application of the inventive concept and technical solution to other occasions without improvement, should be regarded as the scope of protection of this application.

Claims

1. A target detection method, It is characterized in that include: Obtain roadside information based on vehicle-mounted visual sensors, and obtain original point sets of static targets based on vehicle-mounted millimeter-wave radars; Performing a first screening process on the original point set of the static target based on the roadside information to obtain the initial point set of the static target; According to the initial point set of the static target, the original point set of the static target is subjected to a second screening process to obtain a final point set of the static target; Among them, the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

2. The target detection method according to claim 1, It is characterized in that Also includes: Determine a reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road, and / or the first right curve representing the roadside line on the right side of the road; The static targets located outside the reference road are deleted.

3. The target detection method according to claim 2, It is characterized in that The determining of the reference roadside line based on the final point set of the static target, the first left curve representing the roadside line on the left side of the road and / or the first right curve representing the roadside line on the right side of the road comprises: Performing curve fitting according to the final point set of the left side curb target of the road to obtain a second left curve representing the left side curb line of the road; Performing curve fitting according to the final point set of the right side roadside target to obtain a second right curve representing the right side roadside line; In response to the lateral distance between the second left curve and the vehicle being greater than the lateral distance between the first left curve and the vehicle being greater, the second left curve is determined as a reference roadside line; in response to the lateral distance between the second right curve and the vehicle being greater than the lateral distance between the first right curve and the vehicle being greater, the second right curve is determined as a reference roadside line.

4. The target detection method according to claim 3, It is characterized in that Deleting the static target outside the reference road includes: In response to the reference roadside line including the second left curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the second left curve to obtain a first value; in response to the difference between the y coordinate of the current point and the first value being greater than or equal to zero, delete the current point; In response to the reference roadside line including the second right curve, in a preset coordinate system, traverse the original point set of the static target, substitute the x coordinate of the current point into the equation of the second right curve to obtain a second value; in response to the difference between the y coordinate of the current point and the second value being less than or equal to zero, delete the current point; Among them, the origin of the preset coordinate system is the center point of the rear axle of the vehicle, the driving direction of the vehicle is the positive direction of the x-axis, and the left side of the vehicle is the positive direction of the y-axis.

5. The target detection method according to claim 1, It is characterized in that The step of performing a second screening process on the original point set of the static target according to the initial point set of the static target to obtain the final point set of the static target includes: Determine the search direction based on the coordinate position of the point in the initial point set of the target on the left side of the road, and recursively search the points in the original point set of the static target according to the search direction and a preset search range to obtain the final point set of the target on the left side of the road; The search direction is determined based on the coordinate position of the points in the initial point set of the roadside target on the right side of the road, and the points in the original point set of the static target are recursively searched according to the search direction and the preset search range to obtain the final point set of the roadside target on the right side of the road.

6. The target detection method according to claim 5, It is characterized in that The method of determining a search direction based on the coordinate position of a point in the initial point set of the target on the left side of the road, and recursively searching the points in the original point set of the static target according to the search direction and a preset search range to obtain a final point set of the target on the left side of the road includes: Determine the search azimuth according to the coordinate information of the initial point concentration point of the roadside target on the left side of the road; Taking the point in the initial point set that is farthest from the vehicle in the longitudinal direction as a reference point for recursive search, generating a search box of a preset size at the reference point according to the search azimuth, so that the azimuth of the search box is consistent with the search azimuth; The point encircled by the search box is used as a point in the final point set of the roadside target on the left side of the road; The point in the search box that is closest to the reference point is determined as a new reference point to update the position and azimuth of the search box, and the recursive search is stopped until a preset condition is met.

7. The target detection method according to claim 1, It is characterized in that The first screening process is performed on the original point set of the static target based on the roadside information to obtain the initial point set of the static target, including: Traversing the original point set of the left curb target, or traversing the original point set of the static target, substituting the x coordinate of the current point into the equation of the first left curve to obtain a first result; In response to an absolute value of a difference between the y-coordinate of the current point and the first result being less than a preset lateral threshold, determining the current point as a point in an initial point set of a roadside target on the left side of the road; Traversing the original point set of the right curb target, or traversing the original point set of the static target, substituting the x coordinate of the current point into the equation of the first right curve to obtain a second result; In response to the absolute value of the difference between the y-coordinate of the current point and the second result being less than a preset lateral threshold, the current point is determined as a point in the initial point set of the roadside target on the right side of the road.

8. A target detection device, It is characterized in that include: An acquisition module is used to acquire roadside information based on the vehicle-mounted visual sensor, and to acquire the original point set of the static target based on the vehicle-mounted millimeter-wave radar; A first processing module, configured to perform a first screening process on the original point set of the static target based on the roadside information to obtain an initial point set of the static target; A second processing module, configured to perform a second screening process on the original point set of the static target according to the initial point set of the static target, so as to obtain a final point set of the static target; Among them, the roadside information includes a first left curve representing the roadside on the left side of the road and / or a first right curve representing the roadside on the right side of the road; the static target includes a roadside target on the left side of the road and / or a roadside target on the right side of the road.

9. An electronic device, It is characterized in that The electronic device comprises: one or more processors; A storage device for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the target detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, It is characterized in that When the program is executed by a processor, the target detection method as described in any one of claims 1 to 7 is implemented.

Images