Target detection-based vehicle control method and apparatus, and vehicle

Abstract

A target detection-based vehicle control method and apparatus, and a vehicle. The method is applied to a vehicle, and comprises: acquiring first point cloud data of an area above the vehicle and second point cloud data of a surrounding area; on the basis of the first point cloud data and size data of the vehicle, determining a first target position on the vehicle for alignment; on the basis of the second point cloud data and size data of a stacker, determining a second target position on the stacker for alignment; and on the basis of the first target position and the second target position, controlling the vehicle to operate, to align the vehicle with the stacker. The alignment positions of the vehicle and the stacker with high precision are obtained by processing the collected point cloud data, thereby accurately controlling the vehicle to complete alignment driving.

Description

Vehicle control method and device based on target detection and vehicle

[0001] The present application claims priority to the Chinese patent application No. 202410735954.0, filed on June 6, 2024, and entitled "Vehicle control method and device based on target detection and vehicle", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the technical field of vehicles, and in particular to a vehicle control method and device based on target detection and a vehicle. BACKGROUND

[0003] In the implementation of container loading and unloading, the vehicle drives to a position where the reach stacker can directly take the container on the vehicle with a spreader, which is a common operation scenario.

[0004] In the prior art, the conventional method is that the control personnel of the vehicle controls the driving state of the vehicle based on experience to meet the above operation requirements.

[0005] However, in the above technical solution, the manual implementation method has the problems of human resource cost and subjective unreliability. Therefore, to ensure the accuracy of vehicle driving, how to accurately, cost-effectively and safely control the vehicle to reach a specific position becomes a technical problem to be solved. SUMMARY

[0006] The present application provides a vehicle control method and device based on target detection and a vehicle to solve the technical problem that the vehicle cannot be accurately, cost-effectively and safely controlled to reach a specific position in the prior art.

[0007] In a first aspect, an embodiment of the present application provides a method applied to a vehicle, the method comprising:

[0008] obtaining first point cloud data of an area above the vehicle and second point cloud data of a surrounding area;

[0009] determining a first target position for alignment on the vehicle according to the first point cloud data and size data of the vehicle;

[0010] determining a second target position for alignment on the reach stacker according to the second point cloud data and size data of the reach stacker;

[0011] controlling the vehicle to run according to the first target position and the second target position, so that the vehicle is aligned with the reach stacker.

[0012] In one or more embodiments, the determining the first target position on the vehicle for alignment based on the first point cloud data and the size data of the vehicle comprises:

[0013] determining a region of interest of the container on the vehicle according to the size data of the vehicle;

[0014] filtering the first point cloud data to determine whether there is third point cloud data located in the region of interest of the container on the vehicle;

[0015] if the third point cloud data exists, performing fitting processing on the third point cloud data to obtain a front plane equation of the container on the vehicle;

[0016] determining the first target position on the vehicle for alignment according to the front plane equation of the container and the size data of the container on the vehicle, the first target position being a center position of the container on the vehicle.

[0017] In one or more embodiments, the method further comprises:

[0018] if the third point cloud data does not exist, determining the first target position on the vehicle for alignment according to the size data of the vehicle, the first target position being a center position of the container on the vehicle for placing the container.

[0019] In one or more embodiments, the determining the second target position on the stacker for alignment based on the second point cloud data and the size data of the stacker comprises:

[0020] determining a side plane equation and a front tire plane equation corresponding to the stacker according to the second point cloud data;

[0021] determining a center position of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker;

[0022] determining a region of interest of the container on the stacker according to the center position of the stacker and the size data of the container;

[0023] determining the second target position according to fourth point cloud data located in the region of interest of the container on the stacker.

[0024] In one or more embodiments, the determining the center position of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker comprises:

[0025] determining a front plane region of interest of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker;

[0026] filtering the second point cloud data to obtain points located in the front plane region of interest;

[0027] fitting the points in the front plane region of interest to obtain a front plane equation of the stacker;

[0028] determining a center position of the stacker according to the side plane equation, the front tire plane equation, and the front plane equation of the stacker.

[0029] In one or more embodiments, the determining the second target position according to fourth point cloud data located in a region of interest of a container on the stacker comprises:

[0030] dividing the fourth point cloud data into points in three dimensional planes according to normal vector directions;

[0031] for each dimensional plane, when the number of points in the dimensional plane is greater than a preset threshold, determining a height value of a spreader on the stacker according to vertical axis values of the points in the dimensional plane;

[0032] determining a horizontal position of the container according to the points in the three dimensional planes and the center position of the stacker;

[0033] when the spreader is grabbing a container, determining the second target position as the horizontal position of the container and the height value of the spreader;

[0034] when the spreader is not grabbing a container, determining the second target position as the center position of the stacker.

[0035] In one or more embodiments, before the determining the horizontal position of the container according to the points in the three dimensional planes and the center position of the stacker, the method further comprises:

[0036] determining whether the spreader on the stacker is grabbing a container according to point cloud distribution of the three dimensional planes.

[0037] In one or more embodiments, the determining the side plane equation and the front tire plane equation of the stacker according to the second point cloud data comprises:

[0038] projecting normal vectors of each point in the second point cloud data onto a parallel ground to obtain a two-dimensional vector corresponding to each point;

[0039] According to a first preset strategy, the two-dimensional vectors corresponding to the points are filtered to obtain points whose normal vectors are parallel to the parallel ground, and the points whose normal vectors are parallel to the parallel ground are fitted to obtain the side plane equation, the first preset strategy being filtering the points corresponding to the two-dimensional vectors whose square sum of data is greater than a preset threshold;

[0040] According to the side plane equation and a second preset strategy, points in the second point cloud data located in a front tire region are determined, and the points located in the front tire region are fitted to obtain the front tire plane equation, the second preset strategy being point-to-side plane distance screening and / or height screening of the points.

[0041] In a second aspect, an embodiment of the present application provides a vehicle control device based on target detection, applied to a vehicle, the device comprising:

[0042] A first point cloud data acquisition module, configured to acquire first point cloud data of an area above the vehicle and second point cloud data of a surrounding area;

[0043] A first determination module, configured to determine a first target position for alignment on the vehicle according to the first point cloud data and size data of the vehicle;

[0044] A second determination module, configured to determine a second target position for alignment on the stacker according to the second point cloud data and size data of the stacker;

[0045] A control module, configured to control the vehicle to run according to the first target position and the second target position.

[0046] In one or more embodiments, the first determination module is specifically configured to:

[0047] Determine an area of interest of a container on the vehicle according to the size data of the vehicle;

[0048] Filter the first point cloud data to determine whether third point cloud data located in the area of interest exists;

[0049] If the third point cloud data exists, perform fitting processing on the third point cloud data to obtain a front plane equation of the container on the vehicle;

[0050] Determine the first target position for alignment on the vehicle according to the front plane equation of the container and size data of the container on the vehicle, the first target position being a center position of the container on the vehicle.

[0051] In one or more embodiments, the first determination module is further configured to:

[0052] If the third point cloud data does not exist, a first target position for alignment on the vehicle is determined according to size data of the vehicle, and the first target position is a center position of the vehicle for placing the container.

[0053] In one or more embodiments, the second determining module is specifically configured to:

[0054] According to the second point cloud data, a side plane equation and a front tire plane equation corresponding to the stacker are determined.

[0055] According to the side plane equation, the front tire plane equation, and size data of the stacker, a center position of the stacker is determined.

[0056] According to the center position of the stacker and size data of the container, a region of interest of the container on the stacker is determined.

[0057] According to fourth point cloud data in the region of interest of the container on the stacker, the second target position is determined.

[0058] In one or more embodiments, the second determining module determines the center position of the stacker according to the side plane equation, the front tire plane equation, and size data of the stacker, and is specifically configured to:

[0059] According to the side plane equation, the front tire plane equation, and size data of the stacker, a front plane region of interest of the stacker is determined.

[0060] The second point cloud data is filtered to obtain points in the front plane region of interest.

[0061] The points in the front plane region of interest are fitted to obtain a front plane equation of the stacker.

[0062] According to the side plane equation, the front tire plane equation, and the front plane equation of the stacker, the center position of the stacker is determined.

[0063] In one or more embodiments, the second determining module determines the second target position according to the fourth point cloud data in the region of interest of the container on the stacker, and is specifically configured to:

[0064] According to a normal vector direction, the fourth point cloud data is divided into points in three dimensional planes.

[0065] For each dimensional plane, when the number of points in the dimensional plane is greater than a preset threshold, a height value of a spreader on the stacker is determined according to a vertical axis value of the points in the dimensional plane.

[0066] determine the horizontal position of the container according to the point on the three-dimensional plane and the center position of the stacker crane;

[0067] when the spreader is not grabbing the container, determine the second target position as the center position of the stacker crane.

[0068] when the spreader is not grabbing the container, determine the second target position as the center position of the stacker crane.

[0069] In one or more embodiments, before the determination of the horizontal position of the container according to the point on the three-dimensional plane and the center position of the stacker crane, the second determination module is further configured to:

[0070] determine whether the spreader on the stacker crane is grabbing a container according to the point cloud distribution of the three-dimensional plane.

[0071] In one or more embodiments, the second determination module 63 determines the side plane equation and the front tire plane equation corresponding to the stacker crane according to the second point cloud data, specifically:

[0072] project the normal vector of each point in the second point cloud data onto the parallel ground to obtain a two-dimensional vector corresponding to each point;

[0073] filter the two-dimensional vector corresponding to each point according to a first preset strategy to obtain points whose normal vector is parallel to the parallel ground, and perform fitting processing on the points whose normal vector is parallel to the parallel ground to obtain the side plane equation, the first preset strategy being to filter points corresponding to the two-dimensional vector whose square sum of data is greater than a preset threshold;

[0074] determine points in the second point cloud data located in the front tire area according to the side plane equation and a second preset strategy, and perform fitting processing on the points located in the front tire area to obtain the front tire plane equation, the second preset strategy being point-to-side plane distance screening and / or point height screening.

[0075] In a third aspect, the embodiments of the present application provide a vehicle, comprising: a processor, and a memory in communication connection with the processor;

[0076] the memory stores computer execution instructions;

[0077] the processor executes the computer execution instructions stored in the memory to implement the method of the first aspect or any of the manners.

[0078] In a fourth aspect, a computer readable storage medium is provided, in which computer executable instructions are stored. The computer executable instructions, when executed by a processor, enable the method of the first aspect or any of the implementations of the first aspect.

[0079] In a fifth aspect, a computer program product is provided, which includes a computer program stored in a computer readable storage medium. At least one processor can read the computer program from the computer readable storage medium. The at least one processor, when executing the computer program, can implement the method of the first aspect or any of the implementations of the first aspect.

[0080] The vehicle control method, device and vehicle based on target detection provided in the application are applied to a vehicle. The first point cloud data of an upper region of the vehicle and the second point cloud data of a surrounding region are acquired. The first target position for alignment on the vehicle is determined according to the first point cloud data and the size data of the vehicle. The second target position for alignment on the stacker is determined according to the second point cloud data and the size data of the stacker. The vehicle is controlled to run according to the first target position and the second target position, so as to align the vehicle with the stacker. In the technical solution, the alignment positions of the vehicle and the stacker are obtained by processing the acquired point cloud data, so that the vehicle is accurately controlled to run for alignment. BRIEF DESCRIPTION OF DRAWINGS

[0081] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments consistent with the application and, together with the description, serve to explain the principles of the application.

[0082] FIG. 1 is a flowchart of a vehicle control method based on target detection according to an embodiment of the application;

[0083] FIG. 2 is a flowchart of a vehicle control method based on target detection according to an embodiment of the application;

[0084] FIG. 3 is a flowchart of a vehicle control method based on target detection according to an embodiment of the application;

[0085] FIG. 4 is a flowchart of a vehicle control method based on target detection according to an embodiment of the application;

[0086] FIG. 5 is a flowchart of a vehicle control method based on target detection according to an embodiment of the application;

[0087] FIG. 6 is a structural diagram of a vehicle control device based on target detection according to an embodiment of the application;

[0088] FIG. 7 is a structural diagram of a vehicle according to an embodiment of the application.

[0089] The specific embodiments of the present disclosure have been shown by the above drawings, and will be described in more detail hereinafter. These drawings and the written description are not intended to restrict the scope of the concept of the present disclosure by any means, but to illustrate the concept of the present disclosure to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0090] In order to make the purposes, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0091] Before introducing the embodiments of the present application, the application background of the embodiments of the present application is first explained:

[0092] The high-precision alignment of the port stacker is simply summarized as follows: after the transport vehicle drives to the specified operation position, the position of the stacker is detected, and the transport vehicle drives to a position facilitating the stacker to take the container on the vehicle directly by the spreader. (Different from low-precision alignment, the low-precision alignment refers to that the vehicle drives to the target operation point, and the range is relatively rough).

[0093] The process in which the transport vehicle takes the container from the transport vehicle by cooperating with the stacker is as follows.

[0094] S1: the vehicle drives to the starting position for operation with the stacker;

[0095] S2: the vehicle detects the surrounding environment of the vehicle and matches with the stacker closest to the vehicle;

[0096] S3: the vehicle drives to the side of the stacker;

[0097] S4: the spreader opens the connection with the container on the trailer of the vehicle, lifts the container, and moves to the safe range of the vehicle;

[0098] S5: the vehicle drives away.

[0099] Among them, the alignment is to make the vehicle accurately stop at a position fixed relative to the position of the stacker. The essence is to obtain a target position a on the stacker and a target position b on the vehicle (such as a trailer), and then make the positions a and b in the expected relative position relationship.

[0100] In the prior art, in order to achieve the above alignment, the conventional method is that the control personnel of the vehicle controls the running state of the vehicle based on experience to achieve the above operation requirements.

[0101] However, in the above technical solutions, the implementation based on manpower has the problems of human resource cost and subjective unreliability.

[0102] Therefore, the embodiments of the present application need to solve the problem of the prior art: how to accurately, low-cost and safely control the vehicle to reach a specific position.

[0103] In view of the technical problems existing in the prior art, the inventor's idea is as follows: in actual port operation, when the vehicle and the stacker are aligned, the start-stop state of the vehicle is related to the position information of the aligned part. If the corresponding position information can be obtained, the running mode of the vehicle can be controlled based on the position between the two, thereby achieving the above-mentioned purpose. In actual scenarios, by collecting point cloud information in real time through sensors, the target position of the stacker for alignment and the target position of the vehicle for alignment are determined, so that the vehicle can be more accurately controlled to reach a specific position, and there is no subjective nature and human resource cost of manpower implementation.

[0104] In the following, the technical solutions of the present application will be described in detail through specific embodiments. It should be noted that the following specific embodiments can be combined with each other, and the same or similar concepts or processes can not be described in detail in some embodiments.

[0105] It is worth noting that the application field of the disclosed method, device and vehicle is not limited.

[0106] Among them, the execution subject of the present application is a vehicle, which can be a server, a terminal device, etc.

[0107] FIG. 1 is a flowchart of a vehicle control method based on target detection according to an embodiment of the present application. As shown in FIG. 1, the method can include the following steps:

[0108] Step 11, acquiring first point cloud data of an area above the vehicle and second point cloud data of a surrounding area;

[0109] In this step, when the vehicle is aligned with the stacker for detection, the vehicle acquires first point cloud data of an area above the vehicle in real time. The first point cloud data can include point clouds of various possible objects. The vehicle also acquires second point cloud data of a surrounding area in real time. The second point cloud data can also include point clouds of various possible objects.

[0110] Among them, the area above the vehicle can include the front and upper area, the directly above area, the rear and upper area, the left side and upper area, the right side and upper area, etc. The actual implementation can adjust the corresponding acquisition angle based on the angle at which the vehicle itself takes pictures of the trailer storing the container. The surrounding area can include the front of the vehicle, the side, etc. The actual implementation can adjust the corresponding acquisition angle based on the angle at which the stacker takes pictures.

[0111] It should be understood that the device for acquiring point cloud data can be one, the angle can be adjusted based on actual needs, or multiple, fixedly installed, or its shooting angle is wider, which can be fully covered.

[0112] Optionally, the manner of acquiring vehicle point cloud data can be achieved by setting a camera, a camera, a radar and the like on the vehicle, and pixel data acquired by some devices can also be converted into point cloud data and converted into a certain coordinate system of a fixed position of the vehicle through coordinate conversion.

[0113] For example, the coordinate system can be that the position of the vehicle head is the origin, the front is the positive direction of the x-axis, the left is the positive direction of the y-axis, and the top is the positive direction of the z-axis.

[0114] It should be understood that the target involved in the embodiments of the present application includes but is not limited to a reach stacker, a container (the embodiments are based on this, and can also be a box, a package, etc.).

[0115] In addition, the above-mentioned point cloud data can also be divided based on the areas corresponding to the reach stacker and the vehicle itself to obtain first point cloud data and second point cloud data.

[0116] Step 12, determining a first target position for alignment on the vehicle according to the first point cloud data and the size data of the vehicle;

[0117] In this step, the size data of the vehicle is known, and the area for placing the container on the vehicle, specifically the trailer part, is also known. The point cloud data of the area for placing the container can be acquired based on the first point cloud data, and the position for alignment on the vehicle can be determined, which is recorded as the first target position.

[0118] Among them, since the reach stacker is needed for the loading and unloading scene of the container on the vehicle, the position of the reach stacker on the vehicle needs to be determined in the implementation of the vehicle control. Therefore, the first target position can be divided into two possibilities according to whether the container is placed on the vehicle, one is the center position for placing the container on the vehicle, and the other is the center position of the container on the vehicle.

[0119] Step 13, determining a second target position for alignment on the reach stacker according to the second point cloud data and the size data of the reach stacker;

[0120] In this step, the point cloud corresponding to the reach stacker in the second point cloud data can be processed to obtain the side plane of the reach stacker, and then the front tire plane is obtained based on the size data of the reach stacker, and then the interest area of the reach stacker container is obtained, and the position for alignment on the reach stacker can be determined, which is recorded as the second target position.

[0121] Among them, due to the need to use the stacker to load and unload the container on the vehicle scene, the implementation of vehicle control needs to determine the position of the stacker itself, and then the second target position can be divided into two possibilities according to whether there is a container on the clamp of the stacker, one is the horizontal position of the container and the height value of the hoist; the second is the center position of the container.

[0122] Step 14, according to the first target position and the second target position, control the vehicle to run to make the vehicle align with the stacker.

[0123] In a possible implementation, according to the two position coordinates obtained, that is, the second target position (x1, y1, z1) of the stacker alignment and the first target position (x2, y2, z2) of the vehicle alignment, the two positions are coordinate transformed through the difference value (English: heading_offset) between the orientation of the vehicle itself positioning and the orientation of the work lane, that is, projected onto the work lane coordinate axis to obtain the target coordinate x3 of the stacker (obtained after coordinate transformation of x1) and the target coordinate x2 of the vehicle.

[0124] Then the high-precision alignment value is x3-x2, wherein the value represents the distance that the vehicle needs to travel along the work lane, and a positive value represents forward, and a negative value represents reverse.

[0125] In addition, another possible implementation is that y3 (obtained after coordinate transformation of y1) and y2 can also be used to adjust the left and right directions of the vehicle.

[0126] The vehicle control method based on target detection provided by the embodiment of the application comprises the following steps: acquiring first point cloud data of an upper region of a vehicle and second point cloud data of a surrounding region; determining a first target position for alignment of the vehicle according to the first point cloud data and size data of the vehicle; determining a second target position for alignment of a stacker according to the second point cloud data and size data of the stacker; and controlling the vehicle to run according to the first target position and the second target position, so that the vehicle aligns with the stacker. In the technical solution, the alignment positions of the vehicle and the stacker are obtained by processing the collected point cloud data, so that the vehicle is accurately controlled to complete the alignment.

[0127] On the basis of the above embodiment, Fig. 2 is a flowchart of the vehicle control method based on target detection provided by the embodiment of the application, as shown in Fig. 2, the above step 12 (i.e., the determination process of the first target position for alignment of the vehicle) can have the following implementation steps:

[0128] Step 21, determining a region of interest of a container on the vehicle according to size data of the vehicle;

[0129] In this step, the size data of the vehicle is used to configure the interested area of the container on the vehicle.

[0130] For example, the size of the vehicle is 100*20, and the interested area is the range of 70*20*8-28 (the height can be determined according to the size data of the placed container) in the trailer part.

[0131] Step 22, filtering the first point cloud data to determine whether there is third point cloud data in the interested area;

[0132] In this step, after determining the interested area of the container on the vehicle, the first point cloud data obtained can be filtered based on the interested area of the container, so as to obtain the third point cloud data in the interested area, that is, the point cloud falling in the container area of the vehicle.

[0133] It should be understood that when there is no container placed on the vehicle, the third point cloud data may not exist or the number of point clouds is small; when the container is placed on the vehicle, the third point cloud data may exist and the number of point clouds is extremely large.

[0134] Here, the case of less point cloud is classified as non-existent third point cloud data.

[0135] Optionally, before the step 22 is implemented, the first point cloud data can be clustered by using the k-means method based on the positional relationship of the point cloud, and the non-noise point cloud class in the expected position range is retained according to the number of point clouds in the container position, that is, the accuracy of subsequent processing is improved and the interference of the noise point cloud class is reduced.

[0136] Step 23, if the third point cloud data exists, fitting the third point cloud data to obtain the front plane equation of the container on the vehicle;

[0137] In this step, if the third point cloud data exists, the random sample consensus (RANSAC) method is used to fit the points in the third point cloud data, which can effectively eliminate the influence of noise points on fitting.

[0138] Then, the fitted container front plane equation a1x+b1y+c1z+d1=0 is obtained, and the inliers in the fitting process are obtained, that is, the points used to determine the container front plane equation in the fitting process.

[0139] In addition, if the third point cloud data does not exist, the first target position for alignment on the vehicle is determined according to the size data of the vehicle;

[0140] The first target position is the center position of the vehicle for placing the container.

[0141] In this implementation, i.e. there is no container on the vehicle at present, the center position of the container set on the vehicle for placing the container can be directly taken as the alignment point, i.e. the first target position of alignment, based on the size data of the vehicle.

[0142] Step 24, determining the first target position for alignment on the vehicle according to the front plane equation of the container and the size data of the container on the vehicle;

[0143] The first target position is the center position of the container on the vehicle.

[0144] In this step, the center position of the container, such as (x2, y2, z2), can be obtained according to the front plane equation of the container and the size data of the container, and is recorded as the first target position of alignment.

[0145] The vehicle control method based on target detection provided by the embodiments of the present application determines the region of interest of the container on the vehicle according to the size data of the vehicle, filters the first point cloud data to determine whether there is third point cloud data located in the region of interest, performs fitting processing on the third point cloud data to obtain the front plane equation of the container on the vehicle, and determines the first target position for alignment on the vehicle according to the front plane equation of the container and the size data of the container on the vehicle. In this technical solution, the center position of the container on the vehicle is more accurately determined based on the size data of the vehicle and the front plane equation of the container, as the position information for alignment on the vehicle.

[0146] Based on the above embodiments, Fig. 3 is a flowchart of the vehicle control method based on target detection provided by the embodiments of the present application, as shown in Fig. 3, the above step 13 (i.e. the determination process of the second target position for alignment on the stacker) can have the following implementation steps:

[0147] Step 31, determining the side plane equation and the front tire plane equation corresponding to the stacker according to the second point cloud data;

[0148] In this step, after obtaining the second point cloud data, the side plane equation corresponding to the stacker can be determined based on the normal vector of the point, and then the front tire plane equation can be determined based on the size data of the stacker.

[0149] Optionally, the implementation of this step 31 can have the following steps:

[0150] Step 1, projecting the normal vector of each point in the second point cloud data onto the parallel ground to obtain the two-dimensional vector corresponding to each point;

[0151] In this implementation, the normal vector (n x , n y , nz ), Project the normal vector to the x-y plane (parallel to the ground), get a two-dimensional vector (n x , n y ).

[0152] Step 2, according to the first preset strategy, filter the two-dimensional vector corresponding to each point, get the point whose normal vector is parallel to the parallel ground, and carry out fitting processing on the point whose normal vector is parallel to the parallel ground, get the side plane equation, the first preset strategy is to filter the two-dimensional vector corresponding to the two-dimensional vector whose data square sum is greater than the preset threshold;

[0153] In this implementation, the first preset strategy can be Use the relationship Filter the points in the point cloud whose normal vector is approximately parallel to the x-y plane, that is, get the points falling on the side plane of the stacker. Wherein T is a preset threshold, such as 0.99.

[0154] Then, the random sample consensus (RANSAC) method is used to perform plane fitting on the points falling on the side plane of the stacker, which can effectively eliminate the influence of outliers on fitting. The stacker side plane equation a2x+b2y+c2z+d2=0 is obtained by fitting, and the inliers in the fitting process are obtained, that is, the points used to determine the stacker side plane equation in the fitting process.

[0155] Step 3, according to the side plane equation and the second preset strategy, determine the points in the second point cloud data located in the front tire area, and perform fitting processing on the points located in the front tire area to obtain the front tire plane equation, the second preset strategy is point-to-side plane distance screening, and / or point height screening.

[0156] In this implementation, based on the stacker side plane equation and the second point cloud data, the point cloud of the stacker front tire area can be obtained, wherein the second preset strategy includes point-to-side plane distance screening, and / or point height screening.

[0157] This is because the points between the front tire plane and the side plane have a certain distance in x, and / or the height of the points is also inconsistent.

[0158] Then, the random sample consensus (RANSAC) method is used to perform plane fitting on the points falling on the stacker front tire, which can effectively eliminate the influence of outliers on fitting. The stacker front tire plane equation a3x+b3y+c3z+d3=0 is obtained by fitting, and the inliers in the fitting process are obtained, that is, the points used to determine the stacker front tire plane equation in the fitting process.

[0159] Optionally, before this step, the second point cloud data can be input to obtain the approximate position of the stacker around the vehicle (which can be in meter level accuracy) through a stacker detection model (such as a deep learning model).

[0160] After that, the approximate position information of the stacker is obtained, and the stacker interest region is determined. The point cloud falling in the stacker region is obtained by filtering the point cloud with the interest region, that is, the updated second point cloud data is obtained.

[0161] In the embodiments of the present application, the number of stackers is not limited, and one of them is described, that is, the second point cloud data can be the point cloud of a stacker region.

[0162] Step 32, determining the center position of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker;

[0163] In this step, after the side plane equation and the front tire plane equation are obtained, the front plane interest region of the stacker can be determined, and then the center position of the stacker is determined based on the size data of the stacker.

[0164] Optionally, the implementation of this step 32 can have the following steps:

[0165] Step 1, determining the front plane interest region of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker;

[0166] In this implementation, the front plane interest region of the stacker can be obtained by obtaining the side plane equation, the front tire plane equation of the stacker, and the size data of the stacker about the front plane.

[0167] The obtaining method is to obtain the interest region pose by the normal vector of the point of the side plane, to obtain the center point of the interest region by the point cloud of the front tire plane, and to obtain the size of the front plane interest region by the size of the stacker.

[0168] Step 2, filtering the second point cloud data to obtain the points in the front plane interest region;

[0169] In this implementation, after the front plane interest region is obtained, the points of the second point cloud data falling in the front plane interest region are retained.

[0170] Step 3, performing fitting processing on the points in the front plane interest region to obtain the front plane equation of the stacker;

[0171] In this implementation, the RANSAC method is used to fit the plane to the points falling in the front plane region of interest, which can effectively eliminate the influence of outliers on the fitting. The front plane equation of the stacker a4x+b4y+c4z+d4=0 is obtained, and the inliers in the fitting process are obtained, i.e. the points used to determine the front plane equation of the stacker in the fitting process.

[0172] Step 4, according to the side plane equation, the front tire plane equation, and the front plane equation of the stacker, the center position of the stacker is determined.

[0173] In this implementation, the center position (x1, y1, z1) of the stacker is obtained by mutual verification of the side plane equation, the tire plane equation, and the front plane equation of the stacker.

[0174] Step 33, according to the center position of the stacker and the size data of the container, the region of interest of the container on the stacker is determined.

[0175] In this step, the container region of interest on the stacker is determined according to the center position of the stacker and the size data of the corresponding container.

[0176] Then, based on the obtained container region of interest and the second point cloud data, the point cloud data in the container region of interest can be determined, which is recorded as the fourth point cloud data.

[0177] Step 34, according to the fourth point cloud data in the region of interest of the container on the stacker, the second target position is determined.

[0178] In this step, the center position of the container on the stacker is determined according to the fourth point cloud data in the container region of interest, which is recorded as the second target position of the alignment.

[0179] Optionally, the implementation of this step 34 can have the following steps:

[0180] Step 1, according to the direction of the normal vector, the fourth point cloud data is divided into points in three-dimensional planes;

[0181] In this implementation, the normal vector of each point in the region of interest of the container on the stacker is estimated. x y z , The points are divided into three-dimensional planes: x-y plane, x-z plane, and y-z plane according to the direction of the normal vector, and the points in each plane are obtained.

[0182] ​​Step 2, for each dimension plane, when the number of points in the dimension plane is greater than a preset threshold, the height value of the spreader on the stacker is determined according to the vertical axis value of the points in the dimension plane;

[0183] In this implementation, the point cloud quantity of each dimension plane is checked respectively, and when the quantity threshold, i.e. the preset threshold, is reached, a random sample consensus (RANSAC) method is used for plane fitting, and the point cloud z mean value of each plane can represent the calculated spreader height.

[0184] One possibility is that the spreader height calculated from the points of the three dimension planes is compared and voted to obtain the container detection height z1 value of the stacker.

[0185] Step 3, according to the points of the three dimension planes and the center position of the stacker, the horizontal position of the container is determined;

[0186] In this implementation, the points of each dimension plane and the center position of the stacker are mutually checked to obtain the stacker container detection x1, y1 values as the horizontal position of the container.

[0187] Before this step, it can also be determined whether the spreader on the stacker is grabbing a container according to the point cloud distribution of the three dimension planes.

[0188] According to whether the three dimension planes exist, the stacker container grabbing state is obtained, for example, if the three dimension planes all exist, it is determined that the spreader is grabbing a container. This method is based on the actual situation.

[0189] Step 4, when the spreader is grabbing a container, the second target position is determined as the horizontal position of the container and the height value of the spreader;

[0190] In this implementation, when the spreader is grabbing a container, the second target position on the stacker is based on the horizontal position (x1, y1) and z1 value.

[0191] Step 5, when the spreader is not grabbing a container, the second target position is determined as the center position of the stacker.

[0192] In this implementation, when the spreader is not grabbing a container, the center position (x1, y1, z1) of the container on the stacker is used as the second target position for positioning on the stacker.

[0193] The vehicle control method based on target detection provided in this application determines the side plane equation and front tire plane equation of the forklift based on second point cloud data; determines the center position of the forklift based on the side plane equation, front tire plane equation, and the size data of the forklift; determines the region of interest (ROI) of the container on the forklift based on the center position of the forklift and the size data of the container; and determines the second target position based on fourth point cloud data located within the ROI of the container on the forklift. In this technical solution, based on the side plane equation, front tire plane equation, and size data of the forklift, the position of the container on the forklift can be determined more accurately, serving as the alignment information on the forklift.

[0194] Based on the above embodiments, Figure 4 is a flowchart illustrating the vehicle control method based on target detection provided in this application. As shown in Figure 4, the process of determining the alignment target position (second target position) corresponding to the forklift can be implemented as follows:

[0195] Step 01: Install the sensor;

[0196] These are typically cameras or lidar, and their installation positions and angles need to cover the area around the vehicle.

[0197] Step 02: Coordinate transformation;

[0198] Specifically, the sensor data is converted to a coordinate system fixed to the vehicle, with the front of the vehicle being the positive x-axis, the left side of the vehicle being the positive y-axis, and the top of the vehicle being the positive z-axis.

[0199] Step 03: Using the sensor point cloud data as input, obtain the approximate location (meter-level accuracy) of the forklifts around the vehicle through the forklift detection model (deep learning model).

[0200] Step 04: Determine the Region of Interest (RoI) for each forklift based on the approximate location information of each forklift obtained in Step 03. Use this RoI to filter the point cloud and obtain the point clouds falling in different forklift regions.

[0201] Step 05: Estimate the normal vector (n) of each point in the point cloud of each forklift area. x ,n y ,n z ), Projecting the normal vector onto the xy plane (parallel to the ground) yields a two-dimensional vector (n). x ,n y ), using relationships Filtering the point cloud to get the points whose normal vectors are approximately parallel to the x-y plane, that is, the points falling on the side plane of the stacker. Where T is a set threshold, such as 0.99.

[0202] Step 06: The RANSAC method is used to fit the plane of the points falling on the side plane of the stacker obtained in step 05, which can effectively eliminate the influence of outliers on fitting. The side plane equation of the stacker a2x+b2y+c2z+d2=0 is obtained, and the inliers in the fitting process are obtained, that is, the points used to determine the side plane equation of the stacker in the fitting process.

[0203] Step 07: According to the side plane equation of the stacker in step 06, the point cloud of the front tire area of the stacker is obtained, and the acquisition method includes point-to-side plane distance screening and point height screening.

[0204] Step 08: The RANSAC method is used to fit the plane of the points falling on the front tire of the stacker obtained in step 07, which can effectively eliminate the influence of outliers on fitting. The front tire plane equation of the stacker a3x+b3y+c3z+d3=0 is obtained, and the inliers in the fitting process are obtained, that is, the points used to determine the front tire plane equation of the stacker in the fitting process.

[0205] Step 09: According to the side plane and front tire plane equations and point cloud in steps 06 and 08, the front plane region of interest (RoI) of the stacker is obtained, and the acquisition method is to obtain the RoI pose through the side plane normal vector, the RoI center point through the front tire plane point cloud, and the RoI size through the size of the stacker.

[0206] Step 010: The front plane point cloud is obtained according to the front plane RoI of the stacker in step 09.

[0207] Step 011: The RANSAC method is used to fit the plane of the points falling on the front plane of the stacker obtained in step 010, which can effectively eliminate the influence of outliers on fitting. The front plane equation of the stacker a4x+b4y+c4z+d4=0 is obtained, and the inliers in the fitting process are obtained, that is, the points used to determine the front plane equation of the stacker in the fitting process.

[0208] Step 012: According to the side plane, tire plane, and front plane in steps 06, 08, and 011, the center position (x1, y1, z1) of the stacker is obtained.

[0209] Step 013: The container region of interest of the stacker is determined according to the center position of the stacker in step 012, and the point cloud of the region of interest is obtained. The normal vector of each point in the point cloud of the region of interest is estimated, (n xn y n z ), The points are divided into three categories by the normal vector direction: x-y plane, x-z plane, and y-z plane.

[0210] Step 014: The number of points in each type of plane is checked respectively, and when the number threshold is reached, the random sample consensus (RANSAC) method is used for plane fitting. The z mean value of the point cloud of each plane can represent the calculated height of the spreader.

[0211] Step 015: According to whether the three types of planes exist in step 014, the grab state of the container is obtained.

[0212] Step 016: According to the calculated height of the spreader in step 014, the container detection height z1 value is obtained by mutual comparison and voting; according to the feature plane point cloud in step 014 and the center position of the spreader in step 012, the container detection x1 and y1 values are obtained by mutual verification.

[0213] Step 017: The target position of the spreader is obtained, which is divided into a target position with a container on the spreader and a target position without a container on the spreader. The target position of the spreader with a container is based on the container position (x1, y1, z1) in step 016; the target position of the spreader without a container is based on the center position (x1, y1, z1) of the spreader in step 012.

[0214] On the basis of the above embodiment, Fig. 5 is a flowchart of a vehicle control method based on target detection according to an embodiment of the present application, as shown in Fig. 5, the determination process of the corresponding target position (first target position) of the vehicle can be implemented as follows:

[0215] Step 51: Install a sensor;

[0216] Typically, it is a camera or a laser radar, and the installation position and angle need to cover the area above the vehicle body.

[0217] Step 52: Coordinate conversion;

[0218] The sensor data is converted to a certain coordinate system fixed to the ego vehicle, the front of the ego vehicle is the positive direction of the vehicle x axis, the left of the ego vehicle is the positive direction of the y axis, and the top of the ego vehicle is the positive direction of the z axis.

[0219] Step 53: Based on the position relationship of the point cloud, the k-means method is used to cluster the point cloud, and the non-noise point cloud class within the expected position range is reserved according to the number of point clouds of each type of container position.

[0220] Step 54: Obtain the region of interest (RoI) of the container on the trailer through the vehicle size configuration, filter the point cloud based on the region of interest, and obtain the point cloud falling in the container region of the trailer.

[0221] Step 55: Use the random sample consensus (RANSAC) method to perform plane fitting on the points falling in the container region obtained in step 54, which can effectively eliminate the influence of outliers on fitting. The fitted container front plane equation is a1x+b1y+c1z+d1=0, and the inliers in the fitting process are obtained, i.e., the points used to determine the container front plane equation in the fitting process.

[0222] Step 56: Obtain the container center position (x2, y2, z2) according to the container front plane equation in step 55 and the container size.

[0223] Step 57: Obtain the trailer alignment target position, which includes the target position with a container on the trailer and the target position without a container on the trailer. The alignment target position with a container on the trailer is obtained based on the container position (x2, y2, z2) in step 56, and the alignment target position without a container on the trailer is obtained based on the vehicle size configuration.

[0224] In combination with FIGS. 4 and 5, the final alignment implementation can be: according to the obtained two position coordinates, the stacker alignment target position (x1, y1, z1) and the trailer alignment target position (x2, y2, z2), the two positions are coordinate transformed through the difference value (English: heading_offset) between the orientation of the vehicle itself positioning and the orientation of the working lane, i.e., projected onto the working lane coordinate axis to obtain the stacker target coordinate x3 and the trailer target coordinate x2. The high-precision alignment value is x3-x2, wherein the value represents the distance the vehicle needs to travel along the working lane, a positive value indicates forward movement, and a negative value indicates reverse movement.

[0225] The technical solutions and technical effects of the embodiments corresponding to FIGS. 4-5 can be referred to the above embodiments, which will not be described here again.

[0226] The following is an apparatus embodiment related to the present application, which can be used to execute the method embodiments of the present application. For details not disclosed in the apparatus embodiments of the present application, please refer to the method embodiments of the present application.

[0227] FIG. 6 is a structural schematic diagram of a vehicle control device based on target detection provided by an embodiment of the present application. As shown in FIG. 6, the vehicle control device based on target detection includes:

[0228] The acquisition module 61 is configured to acquire first point cloud data of an area above a vehicle and second point cloud data of a surrounding area.

[0229] The first determining module 62 is configured to determine a first target position on the vehicle for alignment according to the first point cloud data and size data of the vehicle.

[0230] The second determining module 63 is configured to determine a second target position on the stacker for alignment according to the second point cloud data and size data of the stacker.

[0231] The control module 64 is configured to control the vehicle to run according to the first target position and the second target position.

[0232] In one or more embodiments, the first determining module 62 is specifically configured to:

[0233] determine a region of interest of the container on the vehicle according to the size data of the vehicle;

[0234] filter the first point cloud data to determine whether there is third point cloud data located in the region of interest;

[0235] if the third point cloud data exists, perform fitting processing on the third point cloud data to obtain a front plane equation of the container on the vehicle;

[0236] determine the first target position on the vehicle for alignment according to the front plane equation of the container and size data of the container on the vehicle, and the first target position is a center position of the container on the vehicle.

[0237] In one or more embodiments, the first determining module 62 is further configured to:

[0238] if the third point cloud data does not exist, determine the first target position on the vehicle for alignment according to the size data of the vehicle, and the first target position is a center position of the container arranged on the vehicle for placing the container.

[0239] In one or more embodiments, the second determining module 63 is specifically configured to:

[0240] determine a side plane equation and a front tire plane equation corresponding to the stacker according to the second point cloud data;

[0241] determine a center position of the stacker according to the side plane equation, the front tire plane equation, and size data of the stacker;

[0242] determine a region of interest of the container on the stacker according to the center position of the stacker and size data of the container;

[0243] determine the second target position according to fourth point cloud data located in the region of interest of the container on the stacker.

[0244] In one or more embodiments, the second determining module 63 determines the center position of the stacker according to the side plane equation, the front tire plane equation, and size data of the stacker, specifically:

[0245] determines a front plane region of interest of the stacker according to the side plane equation, the front tire plane equation, and the size data of the stacker;

[0246] filters the second point cloud data to obtain points located in the front plane region of interest;

[0247] fits the points in the front plane region of interest to obtain a front plane equation of the stacker;

[0248] determines the center position of the stacker according to the side plane equation, the front tire plane equation, and the front plane equation of the stacker.

[0249] In one or more embodiments, the second determining module 63 determines the second target position according to fourth point cloud data located in a region of interest of a container on the stacker, specifically:

[0250] divides the fourth point cloud data into points in three dimensional planes according to the normal vector direction;

[0251] for each dimensional plane, when the number of points in the dimensional plane is greater than a preset threshold, determines a height value of a spreader on the stacker according to the vertical axis values of the points in the dimensional plane;

[0252] determines a horizontal position of the container according to the points in the three dimensional planes and the center position of the stacker;

[0253] when the spreader is grabbing a container, determines the second target position as the horizontal position of the container and the height value of the spreader;

[0254] when the spreader is not grabbing a container, determines the second target position as the center position of the stacker.

[0255] In one or more embodiments, before determining the horizontal position of the container according to the points in the three dimensional planes and the center position of the stacker, the second determining module 63 is further configured to:

[0256] determines whether the spreader on the stacker is grabbing a container according to the point cloud distribution of the three dimensional planes.

[0257] In one or more embodiments, the second determining module 63 determines the side plane equation and the front tire plane equation corresponding to the stacker according to the second point cloud data, specifically:

[0258] projects the normal vector of each point in the second point cloud data onto a parallel ground to obtain a two-dimensional vector corresponding to each point;

[0259] According to the first preset strategy, the two-dimensional vectors corresponding to each point are filtered to obtain points whose normal vectors are parallel to the parallel ground in each point, and the points whose normal vectors are parallel to the parallel ground are fitted to obtain the side plane equation, and the first preset strategy is to filter the two-dimensional vectors corresponding to the two-dimensional vectors whose square sum of data is greater than a preset threshold;

[0260] According to the side plane equation and the second preset strategy, the points in the second point cloud data located in the front tire area are determined, and the points located in the front tire area are fitted to obtain the front tire plane equation, and the second preset strategy is point-to-side plane distance screening and / or point height screening.

[0261] The device provided in the embodiments of the present application can be used to execute the determination method in any of the above-mentioned embodiments, and has similar implementation principles and technical effects, which will not be described here again.

[0262] It should be noted that the division of each module of the above device is only a logical functional division, and all or part of the modules can be integrated into one physical entity, or can be physically separated. And these modules can all be implemented in the form of software called by a processing element; or all can be implemented in the form of hardware; or part of the modules can be implemented in the form of software called by a processing element, and part of the modules can be implemented in the form of hardware. In addition, all or part of the modules can be integrated together, or can be independently implemented. The processing element mentioned here can be an integrated circuit with signal processing capability. In the implementation process, each step of the above method or each module can be completed by the integrated logic circuit of hardware or the instruction of software in the processing element.

[0263] FIG. 7 is a structural schematic diagram of a vehicle provided by the embodiments of the present application, as shown in FIG. 7, the vehicle can include a processor 71, a memory 72, and computer program instructions stored on the memory 72 and executable on the processor 71, and the processor 71 executes the computer program instructions to implement the method provided by any of the preceding embodiments.

[0264] Optionally, the above-mentioned various devices of the vehicle can be connected through a system bus.

[0265] The memory 72 can be a separate storage unit, or can be a storage unit integrated in the processor 71. The number of processors 71 is one or more.

[0266] It should be understood that the processor 71 can be a central processing unit (CPU), and can also be other general-purpose processors 71, digital signal processors 71 (DSP), application specific integrated circuits (ASIC), etc. The general-purpose processor 71 can be a microprocessor 71 or can also be any conventional processor 71, etc. The steps of the method disclosed in the present application can be directly embodied as execution of the hardware processor 71, or execution of a combination of hardware and software modules in the processor 71.

[0267] The system bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or only one type of bus. The memory 72 can include a random access memory 72 (RAM), and can also include a non-volatile memory 72 (NVM), such as at least one disk memory 72.

[0268] All or part of the steps of the above-mentioned method embodiments can be completed by program instruction related hardware. The foregoing program can be stored in a readable memory 72. When the program is executed, the steps of the above-mentioned method embodiments are executed; and the foregoing memory 72 (storage medium) includes a read-only memory 72 (ROM), a RAM, a flash memory 72, a hard disk, a solid state disk, a magnetic tape (English: magnetic tape), a floppy disk (English: floppy disk), an optical disc (English: optical disc), and any combination thereof.

[0269] The vehicle provided by the embodiments of the present application can be used to execute the method provided by any one of the above-mentioned method embodiments, and has similar implementation principles and technical effects, which will not be described here.

[0270] The embodiments of the present application provide a computer readable storage medium, which stores computer instructions, when the computer instructions run on a computer, the computer executes the above-mentioned method.

[0271] The computer readable storage medium described above can be implemented by any type of volatile or nonvolatile storage devices or a combination thereof, such as static random access memory, electrically erasable programmable read-only memory, erasable programmable read-only memory, programmable read-only memory, read-only memory, magnetic storage, flash memory, magnetic disk or optical disk. The readable storage medium can be any available medium that can be accessed by a general or special purpose computer.

[0272] Optionally, the readable storage medium is coupled to the processor, so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in the device.

[0273] The embodiments of the present application also provide a computer program product, which comprises a computer program stored in a computer readable storage medium, at least one processor can read the computer program from the computer readable storage medium, and the at least one processor executes the computer program to implement the above method.

[0274] It should be understood that the present disclosure is not limited to the precise construction that has been described above and shown in the accompanying drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is only limited by the appended claims.

Claims

1. A target detection-based vehicle control method, characterized by, The method is applied to a vehicle and comprises: acquiring first point cloud data of an area above the vehicle and second point cloud data of a surrounding area; determining a first target position on the vehicle for alignment according to the first point cloud data and size data of the vehicle; determining a second target position on the stacker for alignment according to the second point cloud data and size data of the stacker; controlling the vehicle to run so as to align the vehicle with the stacker according to the first target position and the second target position.

2. The method of claim 1, wherein, The determination of the first target position on the vehicle for alignment according to the first point cloud data and the size data of the vehicle comprises: determining an area of interest of a container on the vehicle according to the size data of the vehicle; filtering the first point cloud data to determine whether third point cloud data located in the area of interest exists; if the third point cloud data exists, performing fitting processing on the third point cloud data to obtain a front plane equation of the container on the vehicle; determining the first target position on the vehicle for alignment according to the front plane equation of the container and size data of the container on the vehicle, the first target position being a center position of the container on the vehicle.

3. The method of claim 2, wherein, The method further comprises: if the third point cloud data does not exist, determining the first target position on the vehicle for alignment according to the size data of the vehicle, the first target position being a center position of a container placed on the vehicle.

4. The method of claim 1, wherein, The determination of the second target position on the stacker for alignment according to the second point cloud data and size data of the stacker comprises: determining a side plane equation and a front tire plane equation corresponding to the stacker according to the second point cloud data; determining a center position of the stacker according to the side plane equation, the front tire plane equation and size data of the stacker; determining an area of interest of a container on the stacker according to the center position of the stacker and size data of the container; determining the second target position according to fourth point cloud data located in the area of interest of the container on the stacker.

5. The method of claim 4, wherein, The determination of the center position of the stacker according to the side plane equation, the front tire plane equation and size data of the stacker comprises: determining a front plane area of interest of the stacker according to the side plane equation, the front tire plane equation and size data of the stacker; filtering the second point cloud data to obtain points located in the front plane area of interest; performing fitting processing on the points in the front plane area of interest to obtain a front plane equation of the stacker; determining the center position of the stacker according to the side plane equation, the front tire plane equation and the front plane equation of the stacker.

6. The method of claim 4, wherein, The determination of the second target position according to the fourth point cloud data located in the area of interest of the container on the stacker comprises: dividing the fourth point cloud data into points in three-dimensional planes according to the direction of a normal vector; For each dimension plane, when the number of points in the dimension plane is greater than a preset threshold, a height value of a spreader on the stacker is determined according to a vertical axis value of the point in the dimension plane; A horizontal position of the container is determined according to the points of the three dimension planes and the center position of the stacker; When the spreader is grabbing a container, the second target position is determined as the horizontal position of the container and the height value of the spreader; When the spreader is not grabbing a container, the second target position is determined as the center position of the stacker.

7. The method of claim 6, wherein, Before the horizontal position of the container is determined according to the points of the three dimension planes and the center position of the stacker, the method further comprises: According to the point cloud distribution of the three dimension planes, it is determined whether the spreader on the stacker is grabbing a container.

8. The method of claim 4, wherein, The determination of the side plane equation and the front tire plane equation corresponding to the stacker according to the second point cloud data comprises: The normal vector of each point in the second point cloud data is projected onto the parallel ground to obtain a two-dimensional vector corresponding to each point; According to a first preset strategy, the two-dimensional vector corresponding to each point is filtered to obtain a point whose normal vector is parallel to the parallel ground, and the point whose normal vector is parallel to the parallel ground is fitted to obtain the side plane equation, and the first preset strategy is to filter the two-dimensional vector whose square sum of data is greater than a preset threshold; According to the side plane equation and a second preset strategy, the points in the second point cloud data located in the front tire area are determined, and the points located in the front tire area are fitted to obtain the front tire plane equation, and the second preset strategy is point-to-side plane distance screening and / or point height screening.

9. A vehicle control device based on target detection, characterized by, The device is applied to a vehicle, and the device comprises: An acquisition module is configured to acquire first point cloud data of an area above the vehicle and second point cloud data of a surrounding area; A first determination module is configured to determine a first target position for alignment on the vehicle according to the first point cloud data and size data of the vehicle; A second determination module is configured to determine a second target position for alignment on a stacker according to the second point cloud data and size data of the stacker; A control module is configured to control operation of the vehicle according to the first target position and the second target position.

10. A vehicle characterized by comprising: The device comprises: a processor and a memory connected to the processor in communication; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method according to any one of claims 1 to 8.

Images