Vehicle control method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention discloses a vehicle control method and device, electronic equipment and a storage medium, and can solve the problems of high calculation complexity and limited accuracy when a large number of obstacles exist around a vehicle. The method comprises the following steps: acquiring an original image, wherein the original image comprises at least one target object; determining relative position information between the at least one target object and the target vehicle; according to the relative position information between the at least one target object and the target vehicle, mapping the at least one target object to a preset voxel coordinate system to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed by taking the target vehicle as an original point, the preset voxel coordinate system comprises a plurality of voxel blocks, and the sizes of the voxel blocks and the coordinates of the voxel blocks have a corresponding relationship; the target voxel information at least comprises an object identifier, and the object identifier is used for indicating whether a target object exists in the voxel block; and controlling the target vehicle according to the target voxel information.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of vehicle control, and particularly relate to a vehicle control method and device, an electronic device, and a storage medium. BACKGROUND

[0002] In the field of automatic driving, it is extremely important to ensure that the intelligent driving system accurately perceives the relative distance between the ego vehicle and the surrounding objects and the current state of each object. Most current intelligent driving systems rely on the spatial representation of objects in the environment to understand the position and layout of the objects in the traffic environment. For spatial representation, an occupancy network divides 3D space into voxels and determines whether each voxel is occupied by an obstacle, thereby restoring the 3D environment in which the vehicle is located. However, when there are a large number of obstacles around the vehicle, not only is the calculation complexity high, but also the accuracy is limited. SUMMARY

[0003] Therefore, it is necessary to provide a vehicle control method and device, an electronic device, and a storage medium to solve the above technical problems.

[0004] In a first aspect, embodiments of the present application provide a vehicle control method, which includes: obtaining an original image, the original image including at least one target object;

[0005] determining relative position information between the at least one target object and a target vehicle, the relative position information including: a relative distance and a relative orientation;

[0006] mapping the at least one target object into a preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle to obtain target voxel information; the preset voxel coordinate system being a coordinate system constructed with the target vehicle as the origin, the preset voxel coordinate system including a plurality of voxel blocks, the size of the voxel block and the coordinates of the voxel block having a corresponding relationship; the target voxel information at least including an object identifier, the object identifier being used to indicate whether the target object exists in the voxel block;

[0007] controlling the target vehicle according to the target voxel information.

[0008] In a second aspect, embodiments of the present application provide a vehicle control device, which includes: an obtaining module configured to obtain an original image, the original image including at least one target object;

[0009] a processing module configured to determine relative position information between the at least one target object and a target vehicle, the relative position information including: a relative distance and a relative orientation;

[0010] The processing module is further configured to map the at least one target object into a preset voxel coordinate system according to relative position information between the at least one target object and the target vehicle, to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed with the target vehicle as an origin, the preset voxel coordinate system includes a plurality of voxel blocks, and a size of the voxel block and a coordinate of the voxel block have a corresponding relationship; the target voxel information at least includes an object identifier, and the object identifier is used to indicate whether the target object exists in the voxel block.

[0011] The processing module is further configured to control the target vehicle according to the target voxel information.

[0012] In a third aspect, an electronic device is provided, and the electronic device includes:

[0013] a memory storing executable program code;

[0014] a processor coupled to the memory;

[0015] The processor invokes the executable program code stored in the memory to execute the vehicle control method in the first aspect.

[0016] In a fourth aspect, a computer readable storage medium storing a computer program is provided, and the computer program causes a computer to execute the vehicle control method in the first aspect. The computer readable storage medium includes ROM / RAM, a magnetic disk, an optical disk, and the like.

[0017] In a fifth aspect, a computer program product is provided, and when the computer program product runs on a computer, the computer program product causes the computer to execute part or all steps of any one method in the first aspect.

[0018] In a sixth aspect, an application publishing platform is provided, and the application publishing platform is used to publish a computer program product, and when the computer program product runs on a computer, the computer program product causes the computer to execute part or all steps of any one method in the first aspect.

[0019] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0020] The embodiment of the application provides a vehicle control method and device, electronic equipment and storage medium. An original image is acquired, and the original image comprises at least one target object; relative position information between the at least one target object and a target vehicle is determined, the relative position information comprising: relative distance and relative direction; the at least one target object is mapped into a preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle, to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed with the target vehicle as an origin, and the preset voxel coordinate system comprises a plurality of voxel blocks, and the size of the voxel block and the coordinates of the voxel block have a corresponding relationship; the target voxel information at least comprises an object identifier, and the object identifier is used to indicate whether the target object exists in the voxel block; and the target vehicle is controlled according to the target voxel information. In the scheme, the voxel is represented by the coordinate, so that the calculation amount of analyzing the voxel is greatly reduced, and the scheme is easy to understand; in addition, the voxel of different sizes is set according to different distances from the target vehicle, so that the priority of attention required by the object at different distances in vehicle driving is fully considered, so that the object that needs to be focused or avoided can be quickly captured when the vehicle is controlled, and the accuracy of vehicle control is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0021] The accompanying drawings, which are incorporated herein and form part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application.

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the application, the accompanying drawings needed in the embodiments will be briefly introduced. Obviously, the accompanying drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0023] Figure 1 is a flowchart of a vehicle control method provided by the embodiment of the application Figure 1 ;

[0024] Figure 2 is a schematic diagram of an original image provided by the embodiment of the application

[0025] Figure 3 is a schematic diagram of a preset voxel coordinate system provided by the embodiment of the application Figure 1 ;

[0026] Figure 4 is a schematic diagram of a preset voxel coordinate system provided by the embodiment of the application Figure 2 ;

[0027] Figure 5 is a flowchart of a vehicle control method provided by the embodiment of the applicationFigure 2

[0028] Figure 6 is a flowchart of a vehicle control method provided by an embodiment of the application Figure 3

[0029] Figure 7 is a planar schematic diagram of an original image provided by an embodiment of the application

[0030] Figure 8 is a flowchart of a vehicle control method provided by an embodiment of the application Figure 4

[0031] Figure 9 is a flowchart of a vehicle control method provided by an embodiment of the application Figure 5

[0032] Figure 10 is a structural schematic diagram of a vehicle control device provided by an embodiment of the application

[0033] Figure 11 is a structural schematic diagram of an electronic device provided by an embodiment of the application. DETAILED DESCRIPTION

[0034] In order to more clearly understand the above objectives, features and advantages of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. It should be explained that, in the case of no conflict, the embodiments and the features in the embodiments of the present application can be combined with each other. Obviously, the described embodiments are some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0035] The terms “first” and “second” and the like in the specification and claims of the present application are used to distinguish different objects, and are not used to describe a specific order of the objects.

[0036] The terms “include” and “have” and any variations thereof in the embodiments of the present application are intended to cover the inclusions without exclusivity, for example, a process, a method, a system, a product or an apparatus that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, products or apparatuses.

[0037] ​​​​It should be noted that in the embodiments of the present application, the words such as "exemplary" or "for example" are used to mean serving as an example, instance, or illustration. Any embodiment or design presented as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or advantageous than other embodiments or design solutions. Rather, the use of such terms is intended to present concepts in a concrete manner.

[0038] At present, in order to promote the development of the automatic driving industry, continue to improve the regulations, and clearly define the vehicle function safety, data security and other standards, provide legal basis for technology landing, it is planned to issue L3 level automatic driving vehicle certification system. In addition, it is also determined that a number of enterprises are intelligent networked automobile access and on-road traffic pilot consortium, and they are given the on-road qualification of L3 level automatic driving vehicles, and the technology verification and commercialization exploration are promoted, which fully reflects the importance of the development of automatic driving technology.

[0039] In the field of automatic driving, it is extremely important to ensure that the intelligent driving system accurately perceives the relative distance between the vehicle and the surrounding objects, as well as the current state of each object. Intelligent driving technology trains an AI model to fit human driving behavior by using a large set of human driving data, and learns the driving logic of human beings. Unlike the experience learning method of human beings, the AI model learns the patterns contained in the driving behavior through data. In order to enable the model to quickly and efficiently learn the potential patterns in the data, it is crucial to organize the data into a form that is easy and efficient to understand. Spatial representation technology is a means to achieve this purpose. Most current intelligent driving systems rely on spatial representation of objects in the environment to understand the position and layout of objects in the traffic environment. In the field of automatic driving, spatial representation is one of the core technologies for realizing environment perception and decision planning, and is also an important component of spatial intelligence.

[0040] Domestic and foreign research institutions and enterprises have proposed a variety of methods for spatial representation. Bird's Eye View (BEV) converts the 2D image features of multiple cameras or other sensors into a top-down 3D coordinate system, forming an environment representation with a global perspective. Its advantage lies in constructing a globally coherent two-dimensional spatial cognition, and it is good at processing structured road information such as lane lines and road edges. However, due to the characteristics of planar projection, it is difficult to perceive the height of three-dimensional obstacles, and it is also difficult to perceive the existence of distant objects. In the process of feature transformation, fine-grained structural information is easily lost, resulting in reduced detection accuracy. High-definition map (HD Map) can provide lane-level topological information, but has the problem of high cost of use.

[0041] Currently, the most commonly used method is the Occupancy Network (ON), a deep learning-based environmental perception technique. Its core idea is to divide the 3D space into a voxel grid and use a neural network to predict whether each voxel is occupied by an obstacle (occupied, free, or unobserved). This is combined with semantic labels and temporal dynamic information to construct a fine-grained scene representation. Its advantages lie in providing height and depth information of objects, accurately reconstructing the 3D environment, and significantly improving the perception capabilities of autonomous driving for unknown obstacles and complex geometries. Compared to traditional bounding boxes, it does not rely on predefined object categories but directly models geometric occupancy, thus solving the perception problem of long-tail obstacles (such as irregularly shaped objects or unlabeled categories). However, it also has disadvantages such as high computational complexity, high computational requirements for deployment platforms, and difficulty in deployment on low- to mid-range computing power platforms. Furthermore, because it processes all voxels equally, it leads to an imbalance in the allocation of computational resources between critical near-field regions and low-value far-field regions, making it difficult to improve the analysis accuracy of near-field dynamic targets and generating redundant computational load.

[0042] To address some or all of the aforementioned technical problems, embodiments of this application provide a vehicle control method, apparatus, electronic device, and storage medium. The method involves acquiring an original image, which includes at least one target object; determining relative position information between the at least one target object and a target vehicle, the relative position information including relative distance and relative orientation; mapping the at least one target object to a preset voxel coordinate system based on the relative position information between the at least one target object and the target vehicle to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed with the target vehicle as the origin, and includes multiple voxel blocks, with a correspondence between the size and coordinates of the voxel blocks; the target voxel information includes at least an object identifier, which indicates whether a target object exists within the voxel block; and controlling the target vehicle based on the target voxel information. In this scheme, voxels are represented by coordinates, which significantly reduces the computational load of voxel analysis and makes it easier to understand. In addition, by setting voxels of different sizes at different distances from the target vehicle, the different attention priorities required for objects at different distances during vehicle driving are fully considered. This allows for the rapid capture of objects that need to be focused on or avoided when controlling the vehicle, effectively improving the accuracy of vehicle control.

[0043] like Figure 1 As shown, Figure 1 A flowchart of a vehicle control method provided in this application embodiment, the method may include the following steps:

[0044] 101. Obtain the original image.

[0045] In the embodiments of the present application, the original image can be acquired by a camera installed on the vehicle. The camera can be located at multiple positions of the vehicle, such as the front, the rear, the left side, the right side, the roof and the bottom of the vehicle.

[0046] In some embodiments, the original image can include at least one target object, which can be any object around the vehicle, such as other vehicles, pedestrians, traffic lights, power poles and other road users. Image recognition can be performed on the original image to determine the at least one target object included in the original image, as shown in the figure. Figure 2 The at least one target object is circled by a frame in the original image.

[0047] 102. Determine the relative position information between the at least one target object and the target vehicle.

[0048] In the embodiments of the present application, after the at least one target object is determined, the relative position information between the at least one target object and the target vehicle can be determined. The target vehicle is the vehicle that acquires the original image, and the at least one target object can be considered as an object around the target vehicle.

[0049] It should be noted that the relative position information includes relative distance and relative direction. The relative distance can refer to the distance between the target object and the target vehicle, and the relative direction can refer to the angle of the target object relative to the target vehicle, such as the target object being located in the south of the target vehicle or the target object being located in the northeast direction of 45 degrees of the target vehicle. The relative distance and the relative direction can accurately determine the position of the target object relative to the target vehicle.

[0050] Each target object can be considered to have different relative position information, and objects with the same relative position information can be considered as the same target object. For example, there is a garbage can on the roadside, and a paper box is placed on the garbage can. The relative distance and the relative direction of the garbage can and the paper box relative to the target vehicle are the same, so the garbage can and the paper box can be considered as one target object.

[0051] In some embodiments, when determining the relative position information between at least one target object and a target vehicle, it can be determined by sensors on the vehicle or based on the original image. Specifically, the relative distance can be directly acquired by a distance sensor on the vehicle, or it can be determined by combining the distance between the target object and the target vehicle in the original image with the image scale. The relative orientation can be acquired by an angle sensor on the vehicle, or it can be determined by the angle between the target object and the target vehicle in the original image, as well as the position and shooting range of the camera on the vehicle. Of course, other methods can also be used to determine the relative position information, and this application embodiment does not specifically limit this method.

[0052] 103. Based on the relative position information between at least one target object and the target vehicle, map at least one target object onto a preset voxel coordinate system to obtain target voxel information.

[0053] In this embodiment, the preset voxel coordinate system can be a coordinate system constructed with the target vehicle as the origin. This preset voxel coordinate system can include multiple voxel blocks, and the size of each voxel block corresponds to its coordinates. That is, the regions of each quadrant of the preset voxel coordinate system can be divided into multiple voxel blocks, and the size of each voxel block can be different. Since the origin of the preset voxel coordinate system is the target vehicle, smaller coordinates indicate a closer proximity to the origin and require more attention; therefore, the size of voxels with smaller coordinates can be set smaller. Conversely, larger coordinates indicate a greater distance from the origin and require less attention; therefore, the size of voxels with larger coordinates can be set larger. Figure 3 As shown, the size of the blocks (voxels) closer to the vehicle (target vehicle) is smaller, the size of the blocks (voxels) slightly farther away from the vehicle (target vehicle) (medium) is larger than the size of the blocks (voxels) closer to the vehicle (target vehicle), and the size of the blocks (voxels) farther away from the vehicle (target vehicle) is larger than the size of the blocks (voxels) slightly farther away from the vehicle (target vehicle) (medium).

[0054] For example, such as Figure 3 As shown, the target vehicle can be considered to occupy the area between coordinates (-1, -2), (-1, 2), (1, 2), and (1, -2) in the preset voxel coordinate system. Voxels closer to the target vehicle are represented by 1x1 squares, each voxel representing a 1dm x 1dm area; voxels at a moderate distance are represented by 1x2 squares, each voxel representing a 5dm x 10dm area; and voxels farther away are represented by 2x4 squares, each voxel representing a 10dm x 30dm area. These coordinates and areas are merely examples. Figure 3Only a part of the division of the voxels is shown, and the division of voxels of more regions can also be included, and is not specifically limited.

[0055] In the embodiment of the present application, when mapping the target object to the preset voxel coordinate system, the target object can be mapped to a unique position in the preset voxel coordinate system according to the relative position and the relative orientation of the target object, as shown in Figure 4 It can be seen that the target object exists in at least one voxel in the preset voxel coordinate system, and correspondingly, the target object exists in the voxel, at this time, the object identifier corresponding to the voxel can be updated, the object identifier is used to indicate whether the target object exists in the voxel block, for example, the object identifier can include 0 and 1, the object identifier is 0, which indicates that there is no target object in the voxel, and the object identifier is 1, which indicates that there is a target object in the voxel, after mapping all target objects to the voxels in the preset voxel coordinate system, the object identifiers of all voxels can be summarized, so as to obtain the target voxel information.

[0056] 104. Control the target vehicle according to the target voxel information.

[0057] In the embodiment of the present application, after determining the target voxel information, the positions of each target object around the target vehicle are represented through the coordinate system, so the target vehicle can be controlled through the target voxel information, and the control can be understood as various behaviors such as advancing, turning, braking or avoiding.

[0058] The embodiment of the present application provides a vehicle control method, which represents the voxel through the coordinate, so that the calculation amount of analyzing the voxel is greatly reduced, and is easy to understand; in addition, different sizes of voxels are set according to different distances from the target vehicle, so that the different attention priorities of objects at different distances in vehicle driving are fully considered, so that the object which needs to be focused or avoided can be quickly captured when the vehicle is controlled, and the accuracy of vehicle control is effectively improved.

[0059] In the embodiment of the present application, after determining the relative position relationship between the at least one target object and the target vehicle, the at least one target object needs to be mapped to the preset voxel coordinate system, so the preset voxel coordinate system can be constructed in advance, as shown in Figure 5 Before step 103, the vehicle control method provided by the embodiment of the present application can further include the following steps:

[0060] 501. Create an original coordinate system with the target vehicle as the origin.

[0061] In the embodiment of the present application, an original coordinate system can be created with the target vehicle as the origin, and the coordinate axis direction of the original coordinate system can be set according to the absolute direction, that is, the x axis corresponds to the positive east direction and the positive west direction, and the y axis corresponds to the positive north direction and the positive south direction. Of course, the coordinate axis direction of the original coordinate system can also be set according to the direction of the road, that is, the y axis corresponds to the forward direction and the backward direction of the road, and the x axis corresponds to the direction of the road along the two sides, and of course, other setting modes are also possible, which are not limited in the embodiment of the present application.

[0062] 502. Map the preset voxel grid into the original coordinate system to obtain a preset voxel coordinate system.

[0063] In the embodiment of the present application, after the original coordinate system is created, in order to display the voxel information through the coordinate system, the preset voxel grid can be mapped into the original coordinate system, so that each voxel block can be represented by a coordinate value, thereby obtaining a preset voxel coordinate system.

[0064] In some embodiments, the preset voxel grid can be obtained by dividing the space where the target vehicle is located in advance, for example, as shown in the following table. Figure 5 As shown in the step 502, the vehicle control method can further include the following steps:

[0065] 503. Divide the space where the target vehicle is located into a plurality of distance regions.

[0066] In the embodiment of the present application, during the driving of the vehicle on the road, in order to ensure safety, more attention is paid to the objects closer to the vehicle, and the objects far away are not paid much attention to. Therefore, the space around the vehicle can be divided into distance regions, and each distance region has a different distance from the target vehicle. In addition, more attention is needed for the objects closer to the vehicle, and therefore the distance span of each distance region can also be different, for example, 0-1 meters, 1-3 meters, 3-10 meters, 10-30 meters, 30-150 meters, etc.

[0067] 504. In each distance region, a plurality of voxel blocks are created according to the size corresponding to the distance region.

[0068] In the embodiment of the present application, since the intelligent driving model controls the target vehicle according to the target objects in each voxel, more attention is needed for the objects closer to the target vehicle. Therefore, when creating the voxel blocks, the size of the voxel blocks can be determined according to the distance from the target vehicle, that is, the correspondence between the distance regions and the sizes of the voxel blocks can be set in advance. After the plurality of distance regions are divided, the size of the voxel blocks corresponding to each distance region can be determined according to the correspondence, and then a plurality of voxel blocks are created in the distance region according to the size.

[0069] It should be noted that, since more attention needs to be paid to closer objects, and less attention needs to be paid to distant objects, the size of the voxel block corresponding to the area closer to the vehicle can be smaller, and the size of the voxel block corresponding to the area farther away from the vehicle can be larger.

[0070] For example, the space within 1 m from the vehicle can be divided into cubes with a square base of 1 dm*1 dm; the space 1-3 m from the vehicle can be divided into cubes with a square base of 5 dm*5 dm; the space 3-10 m from the vehicle can be divided into cubes with a square base of 10 dm*10 dm; the space 10-30 m from the vehicle can be divided into cubes with a square base of 30 dm*30 dm; and the space 30-150 m from the vehicle can be divided into cubes with a square base of 50 dm*50 dm. Of course, the above division of different distances and the setting of the base area are exemplary and can be set and adjusted as needed, and the embodiments of the present application are not limited thereto.

[0071] In some embodiments, the size of the voxel block in each distance region can be the same or different, for example, the size of the voxel block in the distance region also gradually increases as the distance from the target vehicle increases, that is, the size of the voxel block close to the target vehicle is small, and the size of the voxel block far from the target vehicle is large, which is not limited specifically.

[0072] In some embodiments, since each distance region includes not only one distance value, there is a certain distance span, and the distance span of the distance region farther away from the target vehicle is larger, for example: as can be seen from the above example, the distance region closest to the target vehicle is within 1 m, and the distance span is 1 m; the next distance region is 1-3 m, and the distance span is 2 m; the next distance region is 3-10 m, and the distance span is 7 m; and the farthest distance region is 30-150 m, and the distance span is 120 m.

[0073] Therefore, when determining the correspondence between the distance region and the size of the voxel block, the distance span existing in the distance region can be calculated, and therefore, in each distance region, a plurality of voxel blocks are created according to the size corresponding to the distance region, which can specifically include: determining a distance calculation result between the target distance region and the target vehicle; and creating a plurality of voxel blocks according to the size corresponding to the distance calculation result.

[0074] It should be noted that the target distance region can be any one of each distance region; for any target distance region, distance calculation is performed to obtain a distance calculation result, which can be understood as calculating the average of the distances between each position in the target distance region and the target vehicle or comparing the distance sizes, so that the corresponding distance calculation result can include the average of each distance, or the minimum value, or the maximum value, so as to create a plurality of voxel blocks according to the size corresponding to the average of each distance; or create a plurality of voxel blocks according to the size corresponding to the minimum value of each distance; or create a plurality of voxel blocks according to the size corresponding to the maximum value of each distance; wherein each voxel block has the same size.

[0075] 505、aggregate a plurality of voxel blocks in each distance region to obtain a preset voxel grid.

[0076] In the embodiment of the present application, after a plurality of voxel blocks are created in each distance region, each distance region is combined to obtain a preset voxel grid, which can specifically include the size of each voxel block, the distance and orientation between each voxel block and the target vehicle, etc.

[0077] In some embodiments, the voxel grid can be adaptively constructed according to different distances between the vehicle and the object, so that the distances between different objects and the vehicle can be distinguished, and the voxel blocks can be used to indicate the position of the object for subsequent calculation, so that the position, distance, angle, etc. between the object and the vehicle can be more intuitive, and the voxel grid and the coordinate system are fused, so that the amount of calculation for analyzing the voxel is greatly reduced, easy to understand, and the accuracy of vehicle control is effectively improved.

[0078] In the embodiment of the present application, when the relative position information between the at least one target object and the target vehicle is determined, the distance between the target object and the target vehicle can be determined, specifically, as shown in Figure 6 The step 102 can specifically include the following steps:

[0079] 1021, determining the straight-line distance between each position point on the target object and the target vehicle.

[0080] In the embodiment of the present application, the target object has a certain volume, and a plurality of different position points distributed on the surface of the target object can be arranged on the target object, and the straight-line distance between each position point and the target vehicle is calculated.

[0081] 1022, determining the minimum value of the straight-line distance as the relative distance.

[0082] In the embodiment of the present application, since the target object has a certain volume, that is, the distance between each position on the target object and the target vehicle can be different, and considering that preventing the ego vehicle from colliding with other objects is extremely important in vehicle driving, the closest straight-line distance between the outer edge of the ego vehicle and the outer edge of the other object can be defined as the distance between the two vehicles, that is, the minimum value of the straight-line distance between each position point on the target object and the target vehicle is determined as the relative distance between the target object and the target vehicle.

[0083] 1023、The position point corresponding to the minimum value of the straight-line distance is determined as the relative position of the target vehicle.

[0084] In the embodiment of the present application, the target object has a certain volume, and the distance between each position on the target object and the target vehicle can be different, and for the same reason, the relative position of each position on the target object to the target vehicle can also be different, so after determining the minimum value of the straight-line distance between each position point on the target object and the target vehicle, the position point corresponding to the minimum value of the straight-line distance can be determined as the closest position point, and the relative position of the closest position point to the target vehicle is determined as the relative position.

[0085] In some embodiments, after obtaining the original image, the original image can be converted by a perception model, that is, the real scene of the vehicle is converted into a two-dimensional planar graph, as shown in Figure 7 It can be seen from the graph that the positional relationship between each target object and the target vehicle is intuitive.

[0086] In the embodiment of the present application, when mapping the target object into the preset voxel coordinate system, the mapping is performed according to the relative position and the relative distance between the target object and the target vehicle, and specifically, as shown in Figure 8 The step 103 can specifically include the following steps:

[0087] 1031, according to the relative position between the target object and the target vehicle, determining the target quadrant of the target object in the preset voxel coordinate system.

[0088] In the embodiment of the present application, in the process of mapping the target object into the preset voxel coordinate system, the target quadrant corresponding to the target object in the preset voxel coordinate system can be first determined according to the relative position between the target object and the target vehicle, for example, the positive direction of the x-axis of the preset voxel coordinate system is the positive east direction, and the positive direction of the y-axis is the positive north direction, assuming that the target object is 30 degrees west and south of the target vehicle, then taking the target vehicle as the origin of the preset voxel coordinate system, the target object can be located in the third quadrant.

[0089] 1032. Determine the target voxel block corresponding to the target object in the target quadrant according to the relative orientation and relative distance between the target object and the target vehicle.

[0090] In the embodiments of the present application, after determining the target quadrant where the target object is located, the unique position of the target object in the target quadrant can be determined according to the relative orientation and relative distance, so as to determine the target voxel block corresponding to the position.

[0091] In some embodiments, when mapping the target object, the two-dimensional planar map obtained after converting the original image shown in FIG. 8 can also be directly mapped into the preset voxel coordinate system, so as to obtain the target voxel block where each target object is located. Figure 7

[0092] 1033. Map the target object into the target voxel block and update the object identifier corresponding to the target voxel block to obtain target voxel information.

[0093] In the embodiments of the present application, after determining the target voxel block corresponding to the target object in the target quadrant, the target object can be mapped into the target voxel block, and at this time, the target object exists in the target voxel block, so the object identifier corresponding to the target voxel block can be updated, for example, assuming that there is no target object in the target voxel block before, the object identifier corresponding to the target voxel block is 0, after mapping the target object into the target voxel block, the object identifier corresponding to the target voxel block can be updated to 1.

[0094] In some embodiments, because the sizes of the voxel blocks in different distance regions are different, a target object can be located in multiple voxel blocks, and multiple target objects can also be located in one voxel block.

[0095] ​For the case that one target object is located in multiple voxel blocks, the target object can be split according to the voxel blocks where the target object is located, that is, each part of the target object located in different voxel blocks is split into multiple target objects, and each part is calculated according to the voxel block where the part is located, for example: a vehicle is very close to the target vehicle, and the size of the voxel block where the vehicle is located is very small, so the vehicle may occupy 4 voxel blocks, and the vehicle can be split into 4 parts, each part is located in a voxel block, and is re-determined as 4 target objects, and subsequent voxel block analysis is performed; or the distance between the multiple voxel blocks where the target object is located and the target vehicle can be determined first, and the target vehicle is moved to the voxel block closest to the target vehicle, here, since the voxel blocks are located in the preset voxel coordinate system, the distance between the multiple voxel blocks and the target vehicle can be determined according to the coordinates corresponding to the voxel blocks, for example: the vehicle may occupy 4 voxel blocks, which are (3, 2), (3, 3), (3, 4) and (4, 2) respectively, it can be seen through comparison that the (3, 2) voxel block is closest to the origin, and therefore the vehicle can be moved to the (3, 2) voxel block, and the object identifier of the (3, 2) voxel block is updated to 1, and the object identifiers of the other three voxel blocks remain 0.

[0096] For the case that multiple target objects are located in one voxel block, it can be considered that the positions of the multiple target objects are relatively close, and therefore the multiple target objects can be determined as one whole to reduce the calculation complexity.

[0097] In the embodiments of the present application, when mapping the target object to the preset voxel coordinate system, the mapping is performed according to the relative position and relative distance between the target object and the target vehicle, specifically, as shown in Figure 9 The step 104 can specifically include the following steps:

[0098] 1041. Perform semantic segmentation on the original image to determine the semantic label of the at least one target object.

[0099] In the embodiments of the present application, when analyzing each voxel block, in addition to the position, the object category is also needed to be considered, for example: the moving vehicle or pedestrian needs to be focused on, and the fixed road sign or traffic light does not need to be focused on, therefore, the original image can be subjected to semantic segmentation to determine the semantic label of each target object.

[0100] 1042. Control the target vehicle according to the target voxel information and the semantic label of the at least one target object.

[0101] In the embodiment of the present application, the semantic label can be information carried by the target object all the time. After the target voxel information is determined, the semantic label of the corresponding target object in each voxel block can be extracted, and then the voxel block can be analyzed according to the target voxel information and the semantic label, so as to control the target vehicle.

[0102] In some embodiments, the target voxel information can further include an image of the target object in the voxel block and a state of the target object. The image of the target object can be obtained by cutting from the original image, and the state of the target object can be determined according to the position difference of the target object in continuous multiple frames of original images, which can be a moving state or a stationary state.

[0103] In some embodiments, the vehicle can be controlled in combination with the object existing in the voxel block indicated by the target voxel information and the object category obtained by analysis, so that the object that needs to be focused on or avoided can be quickly captured, and the accuracy of vehicle control is effectively improved.

[0104] As shown in Figure 10 The vehicle control device provided by the embodiment of the present application can include:

[0105] The acquisition module 1001 is configured to acquire an original image, and the original image includes at least one target object.

[0106] The processing module 1002 is configured to determine relative position information between the at least one target object and a target vehicle, and the relative position information includes a relative distance and a relative direction.

[0107] The processing module 1002 is further configured to map the at least one target object to a preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle, to obtain target voxel information. The preset voxel coordinate system is a coordinate system constructed with the target vehicle as the origin. The preset voxel coordinate system includes a plurality of voxel blocks. The size of the voxel block and the coordinates of the voxel block have a corresponding relationship. The target voxel information at least includes an object identifier, and the object identifier is used to indicate whether the target object exists in the voxel block.

[0108] The processing module 1002 is further configured to control the target vehicle according to the target voxel information.

[0109] In some embodiments, the processing module 1002 is further configured to create an original coordinate system with the target vehicle as the origin.

[0110] The processing module 1002 is further configured to map the preset voxel grid to the original coordinate system to obtain the preset voxel coordinate system.

[0111] In some embodiments, the processing module 1002 is further configured to divide a space where the target vehicle is located into a plurality of distance regions, each distance region corresponding to a different distance between the target vehicle;

[0112] The processing module 1002 is further configured to create a plurality of voxel blocks in each distance region according to a size corresponding to the distance region.

[0113] The processing module 1002 is further configured to aggregate the plurality of voxel blocks in each distance region to obtain a preset voxel grid.

[0114] In some embodiments, the processing module 1002 is specifically configured to determine, for a target distance region, a distance calculation result between the target distance region and the target vehicle, the distance calculation result being an average value, a minimum value, or a maximum value of distances between each position in the target distance region and the target vehicle, the target distance region being any one of the distance regions.

[0115] The processing module 1002 is specifically configured to create a plurality of voxel blocks according to a size corresponding to the distance calculation result, the plurality of voxel blocks having the same size.

[0116] In some embodiments, the processing module 1002 is specifically configured to determine a straight-line distance between each position point on the target object and the target vehicle.

[0117] The processing module 1002 is specifically configured to determine a minimum value of the straight-line distance as a relative distance.

[0118] The processing module 1002 is specifically configured to determine a direction of the position point corresponding to the minimum value of the straight-line distance relative to the target vehicle as a relative direction.

[0119] In some embodiments, the processing module 1002 is specifically configured to determine, according to the relative direction between the target object and the target vehicle, a target quadrant of the target object in a preset voxel coordinate system.

[0120] The processing module 1002 is specifically configured to determine, according to the relative direction and the relative distance between the target object and the target vehicle, a target voxel block of the target object in the target quadrant.

[0121] The processing module 1002 is specifically configured to map the target object into the target voxel block and update an object identifier corresponding to the target voxel block to obtain target voxel information.

[0122] In some embodiments, the processing module 1002 is specifically configured to perform semantic segmentation on the original image to determine a semantic label of the at least one target object.

[0123] The processing module 1002 is specifically configured to control the target vehicle according to the target voxel information and the semantic label of the at least one target object.

[0124] In the embodiments of the present application, each module can implement the vehicle control method provided by the above method embodiments, and achieve the same technical effects. To avoid repetition, it will not be repeated here.

[0125] As shown in Figure 11 The embodiments of the present application also provide an electronic device, which can include:

[0126] The memory 1101 stores executable program codes;

[0127] The processor 1102 is coupled with the memory 1101;

[0128] The processor 1102 calls the executable program codes stored in the memory 1101 to execute the vehicle control method performed by the electronic device in each method embodiment.

[0129] The embodiments of the present application provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement each process of the vehicle control method in the above method embodiments, and can achieve the same technical effects. To avoid repetition, it will not be repeated here.

[0130] The embodiments of the present application also provide a computer program product, which stores a computer program. The computer program is executed by a processor to implement each process of the vehicle control method in the above method embodiments, and can achieve the same technical effects. To avoid repetition, it will not be repeated here.

[0131] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media containing computer usable program codes.

[0132] It should be understood that all the devices and methods disclosed in the embodiments of the present application can be implemented by other ways. The device embodiments described above are only schematic, and for instance, the flowcharts and the block diagrams in the figures illustrate the possible implementation modes of the devices, methods and computer program products according to the embodiments of the present application. In this regard, each block in the flowcharts or the block diagrams can represent a module, a program segment or a part of code, which contains one or more executable instructions for implementing the specified logic function. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur in a different order from that noted in the figures. For example, two consecutive blocks can actually be executed in parallel or in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or the flowcharts, and the combination of blocks in the block diagrams and / or the flowcharts, can be implemented by a dedicated hardware-based system for implementing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0133] In the present application, the processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0134] In the present application, the memory can include a non-persistent memory in a computer readable medium, random access memory (RAM) and / or non-volatile memory, etc., such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of the computer readable medium.

[0135] In the present application, those of ordinary skill in the art can understand that all or part of the steps of various methods in the above embodiments can be completed by relevant hardware instructed by a program, and the program can be stored in a computer readable storage medium, including permanent and non-permanent, removable and non-removable storage medium. The storage medium can realize information storage by any method or technology, and the information can be computer readable instructions, data structure, program module or other data. Examples of computer storage medium include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), other types of random access memory (RAM), read-only memory (ROM), one-time programmable read-only memory (OTPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, disk storage or other magnetic storage device or any other non-transmission medium that can be used to store information accessible by a computing device. According to the definition herein, the computer readable medium does not include transitory computer readable media such as modulated data signals and carriers.

[0136] It should be noted that, in the specification, the terms "one embodiment" or "an embodiment" merely indicate a possible embodiment and do not necessarily specify a single or particular embodiment. Furthermore, the above-described features and characteristics can be combined in any suitable manner in various embodiments. One of ordinary skill in the art will recognize that the embodiments described herein can be practiced with various modifications and alterations, and equivalents thereof. Accordingly, the application should not be considered as limited to the specific recitations. Therefore, the above disclosure is not to be considered as limiting the scope of the application, but rather, is provided for illustrative purposes.

[0137] It should be understood that, throughout the specification, "one embodiment" or "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Therefore, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout the specification are not necessarily referring to the same embodiment. Further, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It will be appreciated by those skilled in the art that the embodiments described herein that can be implemented with or without employing any hardware and that the software that can be implemented operates on any platform capable of running a messaging application.

[0138] In various embodiments of the present application, it should be understood that the magnitude of the serial number of the above processes does not mean the inevitable sequence of execution order, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0139] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e. may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0140] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0141] The integrated units described above, if implemented in the form of software function units and sold or used as independent products, can be stored in a computer accessible memory. Based on such understanding, the technical solutions of the present application essentially or the part of the prior art that contributes to the present application or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of parts or all steps of the above-mentioned method for enabling a computer device (which can be a personal computer, a server or a network device, etc., and specifically can be a processor in the computer device) to execute the embodiments of the present application.

[0142] The above is only a specific implementation of the present application, enabling those skilled in the art to understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A vehicle control method characterized by, The method comprises: acquiring an original image, the original image comprising at least one target object; determining relative position information between the at least one target object and a target vehicle, the relative position information comprising: a relative distance and a relative direction; mapping the at least one target object into a preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed with the target vehicle as the origin, the preset voxel coordinate system comprising a plurality of voxel blocks, the size of the voxel block and the coordinates of the voxel block having a corresponding relationship; the target voxel information at least comprises an object identifier, the object identifier being used to indicate whether the target object exists in the voxel block; controlling the target vehicle according to the target voxel information.

2. The method of claim 1, wherein, Before the mapping of the at least one target object into the preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle to obtain the target voxel information, the method further comprises: creating an original coordinate system with the target vehicle as the origin; mapping a preset voxel grid into the original coordinate system to obtain the preset voxel coordinate system.

3. The method of claim 2, wherein, Before the mapping of the preset voxel grid into the original coordinate system to obtain the preset voxel coordinate system, the method further comprises: dividing a space where the target vehicle is located into a plurality of distance regions, the distance between each distance region and the target vehicle being different; in each distance region, creating a plurality of voxel blocks according to the size corresponding to the distance region; summarizing the plurality of voxel blocks in each distance region to obtain the preset voxel grid.

4. The method of claim 3, wherein, The creating of the plurality of voxel blocks in each distance region according to the size corresponding to the distance region comprises: for a target distance region, determining a distance calculation result between the target distance region and the target vehicle, the distance calculation result being an average value, or a minimum value, or a maximum value of the distance between each position in the target distance region and the target vehicle; the target distance region being any one of the distance regions; creating a plurality of voxel blocks according to the size corresponding to the distance calculation result, the size of the plurality of voxel blocks being the same.

5. The method of claim 1, wherein, The determination of the relative position information between the at least one target object and the target vehicle comprises: determining a straight-line distance between each position point on the target object and the target vehicle; determining the minimum value of the straight-line distance as the relative distance; determining the direction of the position point corresponding to the minimum value of the straight-line distance relative to the target vehicle as the relative direction.

6. The method of claim 1, wherein, The mapping of the at least one target object into the preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle to obtain the target voxel information comprises: determining a target quadrant of the target object in the preset voxel coordinate system according to the relative direction between the target object and the target vehicle; According to the relative position and the relative distance between the target object and the target vehicle, a corresponding target voxel block of the target object in the target quadrant is determined; The target object is mapped into the target voxel block, and a corresponding object identifier of the target voxel block is updated to obtain target voxel information.

7. The method of claim 1, wherein, According to the target voxel information, the target vehicle is controlled, including: The original image is subjected to semantic segmentation to determine a semantic label of the at least one target object; According to the target voxel information and the semantic label of the at least one target object, the target vehicle is controlled.

8. A vehicle control device characterized by comprising: The device includes: An acquisition module is configured to acquire an original image, the original image including at least one target object; A processing module is configured to determine relative position information between the at least one target object and a target vehicle, the relative position information including a relative distance and a relative position; The processing module is further configured to map the at least one target object into a preset voxel coordinate system according to the relative position information between the at least one target object and the target vehicle to obtain target voxel information; the preset voxel coordinate system is a coordinate system constructed with the target vehicle as an origin, the preset voxel coordinate system including a plurality of voxel blocks, a size of the voxel block and a coordinate of the voxel block having a corresponding relationship; the target voxel information at least including an object identifier, the object identifier being used to indicate whether the target object exists in the voxel block; The processing module is further configured to control the target vehicle according to the target voxel information.

9. An electronic device, comprising: It includes: A memory storing executable program codes; and a processor coupled with the memory; The processor invokes the executable program codes stored in the memory to execute the vehicle control method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It includes: Computer instructions are stored on the computer readable storage medium, and the computer instructions are executed by the processor to implement the vehicle control method according to any one of claims 1 to 7.

Images