Method for determining target shooting position of virtual camera and vehicle

Abstract

The invention provides a method for determining a target shooting position of a virtual camera and a vehicle, and relates to the technical field of intelligent driving. When the target vehicle is in the parking scene, based on multiple pieces of environment element information around the target vehicle, rendering restoration world elements corresponding to the environment element information in a virtual space, and aligning the virtual space with a real environment; based on element position information of the plurality of restored world elements in the virtual space, camera parameters of a virtual camera and a preset observation angle, direction indication information of the plurality of restored world elements and a target shooting distance are determined so as to respond to environment changes in real time, and the suitability of a parking scene is improved; determining a target shooting position of the virtual camera in the virtual space based on the direction indication information and the target shooting distance; and the image shot at the target shooting position is displayed, so that a user is helped to perceive the spatial relationship between the target vehicle and the surrounding environment, the parking difficulty is reduced, and the parking safety and convenience are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent driving technology, and more specifically, to a method for determining the target shooting position of a virtual camera and a vehicle thereof. Background Technology

[0002] With the rapid development of intelligent driving technology, parking assistance has become one of the core application scenarios of in-vehicle systems. To enhance the user experience during the parking process, the HUT (Human-Machine Interface Terminal) needs to render a 3D representation of the real world (such as road lines and points) in real time based on the intelligent driving domain control via the Some / IP (Scalable service-oriented middleware over IP) protocol. This renders the representation to help users quickly determine the parking path and significantly reduce the difficulty of parking operations. In the 3D parking reconstruction scenario, the position of the virtual camera directly determines the field of view coverage, the degree of image distortion, and the rationality of observation. If its position is not properly designed, key environmental elements may be cropped, the image may be stretched and distorted, or the user may not be able to clearly perceive the parking space boundary, thus affecting the parking assistance effect. Therefore, accurately determining the position of the virtual camera has become the main requirement of 3D parking reconstruction technology.

[0003] In existing technologies, fixed position parameters are preset manually and directly applied to all parking scenarios to achieve basic 3D parking visualization functions, which to a certain extent meets users' basic needs for observing the parking environment. Summary of the Invention

[0004] The method and vehicle for determining the target shooting position of the virtual camera provided in this application use environmental element information to replace manually preset fixed positions. When facing different parking scenarios, the method can respond to environmental changes in real time to determine the target shooting position of the virtual camera, improve the adaptability of parking scenarios, and avoid problems such as incomplete field of view coverage or field of view redundancy.

[0005] Firstly, a method for determining the target shooting position of a virtual camera is provided. This method includes: when a target vehicle is in a parking scene, rendering multiple restored world elements corresponding to the multiple environmental elements in a virtual space based on information about multiple environmental elements surrounding the target vehicle; determining the directional indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance of the virtual camera to shoot at the multiple restored world elements based on the element position information of the multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and a preset viewing angle; the directional indication information is used to indicate the relative positional relationship between the virtual camera and the multiple restored world elements; determining the target shooting position of the virtual camera in the virtual space based on the directional indication information and the target shooting distance; and displaying the image captured by the virtual camera at the target shooting position.

[0006] In the above technical solution, when the target vehicle is in a parking scenario, based on information about multiple environmental elements around the target vehicle, multiple restored world elements corresponding to these environmental elements are rendered in virtual space to align the relative positions of the restored world elements in the virtual space with the real environment. This ensures that the target shooting position of the virtual camera and the output 2D image are consistent with the real scene. Based on the element position information of the multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset viewing angle, the directional indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot the multiple restored world elements are determined. Therefore, this application does not require any pre-setting of fixed shooting positions. The camera coordinate system relies entirely on real-time collected environmental element information to calculate the target shooting position, responding to environmental changes in real time and improving the adaptability of parking scenarios. Based on directional information and target shooting distance, it determines the target shooting position of the virtual camera in virtual space, ensuring that when the virtual camera shoots at the target shooting position, it can both cover the entire parking scene and present the optimal parking perspective, so as to realize the synchronous update of the virtual perspective when the target vehicle moves, meeting the real-time requirements of parking scenarios. It displays the image taken by the virtual camera at the target shooting position to help users quickly perceive the spatial relationship between the target vehicle and the surrounding environment, reduce parking difficulties caused by blind spots or incomplete field of vision, reduce the risk of parking collisions, and improve the safety and convenience of parking.

[0007] In a second aspect, a vehicle is provided, comprising at least: an intelligent driving domain controller and an in-vehicle human-machine interaction terminal; the intelligent driving domain controller and the in-vehicle human-machine interaction terminal are communicatively connected; and the in-vehicle human-machine interaction terminal is used to execute the method in any possible implementation of the first aspect above. Attached Figure Description

[0008] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic diagram of a vehicle provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of an in-vehicle human-machine interaction terminal provided in an embodiment of this application; Figure 3 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 1 ; Figure 4 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 2 ; Figure 5 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 3 ; Figure 6 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 4 ; Figure 7 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 5 ; Figure 8 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 6 ; Figure 9 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 7 ; Figure 10 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 8 ; Figure 11 This application provides a schematic diagram of the structure of a virtual camera target shooting position according to an embodiment of the present application; Figure 12 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 9 ; Figure 13 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 10 ; Figure 14 This is a schematic diagram of a device for determining the target shooting position of a virtual camera, provided in an embodiment of this application. Detailed Implementation

[0009] The technical solutions in this application will now be described clearly and in detail with reference to the accompanying drawings.

[0010] Before elaborating on the technical solution, the technical terms used in this application will be explained to facilitate subsequent understanding.

[0011] 3D bounding boxes are used to represent the smallest cuboid that can completely enclose a single object or multiple objects in three-dimensional space.

[0012] Virtual Camera: In the Unity engine, the virtual camera is a component used to implement the functions of observing and rendering three-dimensional scenes. It observes the scene by setting a specific (target) position and a specific (target) angle in three-dimensional space, and renders the 3D scene into a 2D image, which is finally output to the screen for display or rendered into a texture.

[0013] The horizontal physical size of the actual vehicle camera (Camera.SensorSize), in mm. The preset physical size consists of the preset physical width and the preset physical height.

[0014] The field of view (Camera.FieldOfView) of a virtual camera can be divided into the horizontal field of view and the vertical field of view.

[0015] The preset focal length of a real vehicle camera (Camera.FocalLength), in mm.

[0016] The lens shift of a virtual camera (Camera.LensShift) can be divided into horizontal (horizontal) shift and vertical (vertical) shift (unit: °). For ease of explanation, this application uses the horizontal (horizontal) shift as an example.

[0017] The following detailed description, in conjunction with the accompanying drawings, of the method for determining the target shooting position of the virtual camera and the vehicle according to various embodiments of this application.

[0018] Figure 1 This is a schematic diagram of a vehicle provided in an embodiment of this application. (As shown...) Figure 1 As shown, the vehicle 100 includes at least: intelligent driving domain control 110 and in-vehicle human-machine interaction terminal 120.

[0019] Among them, the intelligent driving domain control 110 and the vehicle human-machine interaction terminal 120 are connected.

[0020] The intelligent driving domain controller 110 is used to acquire information about multiple environmental elements around the target vehicle and send the information about multiple environmental elements to the vehicle human-machine interaction terminal 120 through a preset communication protocol to avoid data delay or loss caused by additional hardware transfer. The preset communication protocol can be selected according to the actual situation. For example, the preset communication protocol can be selected as the Some / IP communication protocol.

[0021] The environmental element information includes at least: multiple raw environmental data and environmental attribute information. The raw environmental data is collected by various preset vehicle-mounted sensors, and includes at least: raw perception data of road lines, parking lines, and obstacles. The preset vehicle-mounted sensors can be selected according to actual conditions; for example, these preset vehicle-mounted sensors can be selected as surround-view cameras, millimeter-wave radar, etc.

[0022] Environmental attribute information is an auxiliary feature corresponding to the original environmental data. That is, environmental attribute information may include at least: the solid and virtual types of road lines, the size parameters of parking space lines, the three-dimensional size and type of obstacles, etc.

[0023] The vehicle-mounted human-machine interaction terminal 120 is a carrier for user interaction and virtual scene processing. The vehicle-mounted human-machine interaction terminal 120 integrates a control device and a virtual camera, which work together to complete the entire process of virtual scene calculation, control and rendering. The vehicle-mounted human-machine interaction terminal 120 can be implemented using HUT (Human-Machine Interaction Terminal).

[0024] The control device is connected to the virtual camera to convert the target shooting position command into a virtual shooting action; the intelligent driving domain control 110 is connected to the control device to transmit real environmental data to virtual computing data; thus forming a link from data input, command output to execution feedback, ensuring that each step of the operation has a clear hardware support.

[0025] The control device receives multiple environmental element information sent by the intelligent driving domain controller 110 and executes a method for determining the target shooting position of the virtual camera. This method converts the hardware-collected data into executable control commands to obtain the target shooting position of the virtual camera and sends this position to the virtual camera. The control device can be implemented using the Unity engine integrated into the in-vehicle human-machine interface terminal 120.

[0026] After receiving the target shooting position sent by the control device, the virtual camera adjusts its posture in the virtual space based on the target shooting position (such as spatial coordinate positioning and viewing angle calibration), and shoots the rendered elements of the virtual world. Through its built-in rendering function, it renders the elements of the virtual world in the three-dimensional virtual space into a two-dimensional image that meets the parking perspective requirements. The two-dimensional image is then sent back to the vehicle-mounted human-machine interaction terminal 120, which finally displays it for the user to view the virtual perspective of the parking scene.

[0027] The vehicle provided in this application can be composed of an intelligent driving domain controller and an in-vehicle human-machine interface terminal; the intelligent driving domain controller and the in-vehicle human-machine interface terminal are communicatively connected; the in-vehicle human-machine interface terminal integrates a control device and a virtual camera; the control device and the virtual camera are connected to ensure that the control device issues commands without delay, meeting the need for real-time adjustment of the virtual perspective in parking scenarios; the control device is communicatively connected to the intelligent driving domain controller to realize the transmission of real environmental data to virtual computing data, reducing real-time performance loss. Therefore, the intelligent driving domain controller of this application is mainly used for acquiring environmental data, and the control device is mainly used for calculating the target shooting position and issuing commands, making full use of the computing power resources of the in-vehicle human-machine interface terminal, achieving dedicated use of resources, and improving the adaptability and reliability of the system.

[0028] Optionally, Figure 2 This is a schematic diagram of the structure of an in-vehicle human-machine interaction terminal 120 provided in an embodiment of this application. Figure 2 As shown, the vehicle-mounted human-machine interaction terminal 120 may include a processor 121 and a memory 122.

[0029] The memory 122 stores machine-executable instructions that can be executed by the processor 121. When the vehicle-mounted human-machine interface terminal 120 is running, these machine-readable instructions are executed. The processor 121 and the memory 122 communicate via a bus. The processor 121 can execute these machine-executable instructions to implement a method for determining the target shooting position of the virtual camera.

[0030] The memory 122, processor 121, and bus components are electrically connected directly or indirectly to enable data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The mobile storage device includes at least one software function module that can be stored in the memory 122 as software or firmware or embedded in the operating system (OS) of the control device. The processor 121 is used to execute executable modules stored in the memory 122, such as the software function modules and computer programs included in the method for determining the target shooting position of a virtual camera on a mobile storage medium.

[0031] The memory 122 may be, but is not limited to, random access memory (RAM), read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), etc.

[0032] The vehicle-mounted human-machine interaction terminal 120 has software or application programs (APPs) that can execute methods for determining the target shooting position of a virtual camera.

[0033] The method for determining the target shooting position of a virtual camera provided in this application embodiment can be executed by the processor in the vehicle-mounted human-machine interaction terminal 120. The method for determining the target shooting position of a virtual camera provided in this application embodiment will be explained further below. Figure 3 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 1 .like Figure 3 As shown, the method may include: S310. When the target vehicle is in a parking scene, based on the information of multiple environmental elements around the target vehicle, render multiple restored world elements corresponding to the multiple environmental element information in the virtual space.

[0034] In one possible implementation, when the target vehicle is in a parking scenario, the aforementioned intelligent driving domain controller 110 acquires information on multiple environmental elements of the real parking environment around the target vehicle and transmits this information to the virtual space of the control device. In this virtual space, the information on multiple environmental elements is rendered to generate multiple corresponding restored world elements, and the relative positions of these restored world elements are completely consistent with the real parking environment of the target vehicle. This transforms the real parking environment into a calculable digital model in the virtual space, providing a reliable data foundation for the subsequent calculation of the target shooting position of the virtual camera.

[0035] S320. Based on the element position information of multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset observation angle, determine the direction indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot the multiple restored world elements.

[0036] The preset observation angle can be flexibly selected according to the actual parking environment requirements. For example, the preset observation angle can be selected as a 45° overhead view, which allows users to intuitively judge the distance between the target vehicle and the parking space in all directions. This preset observation angle is used to limit the observation direction of the virtual camera, providing a basis for subsequent shooting angle calibration.

[0037] In one possible implementation, the element position information of multiple simulated world elements acquired by S310 in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset viewing angle (such as a 45° overhead view) can jointly determine two types of information: First, direction indication information, used to indicate the relative positional relationship between the virtual camera and multiple simulated world elements. For example, if the direction indication information points to the center of the simulated world element, it means that the virtual camera needs to shoot towards the center; if the direction indication information is at a 30° angle to the longitudinal axis of the simulated world element, it means that the virtual camera needs to be tilted at a 30° angle to shoot, ensuring that the shooting angle meets the usage requirements of the parking scenario. This direction indication information can be represented by the direction vector OD. Second, target shooting distance, used to indicate the distance between the virtual camera and multiple simulated world elements, ensuring that all simulated world elements can completely fall into the view frustum of the virtual camera, avoiding the omission of key environmental element information. Essentially, it limits the target length of the direction indication information (direction vector OD).

[0038] It should be noted that the camera parameters and preset viewing angles of the virtual camera can be flexibly configured to adapt to parking needs in various scenarios. For example, when parking perpendicularly, a 45° overhead view can be set to help users judge the lateral distance between the vehicle and the parking spaces on both sides; when parking parallelly, a 30° side view can be set to help users judge the longitudinal distance between the vehicle and the vehicles in front and behind; in narrow parking space scenarios, the camera parameters of the virtual camera can be adjusted to expand the field of view, allowing users to see more surrounding obstacles.

[0039] S330 determines the target shooting position of the virtual camera in virtual space based on direction indication information and target shooting distance.

[0040] In one possible implementation, the target shooting position of the virtual camera is determined by vector operations, using the center coordinates of multiple elements' positions in the virtual space as a reference, combined with directional information and the target shooting distance. For example, if the directional information points diagonally above the center coordinates and the target shooting distance is 5 meters, then the target shooting position = center coordinates + directional information × 5 meters.

[0041] It's important to note that when calculating the target shooting position, the directional information needs to be normalized first to avoid distance calculation errors. For example, if the directional information (directional vector OD) is normalized to (3, 4, 0), calculate the unit vector of the directional information. Since the magnitude of the directional information is 5, the unit vector = directional information ÷ magnitude of directional information = (3 ÷ 5, 4 ÷ 5, 0 ÷ 5) = (0.6, 0.8, 0). Then, multiply this unit vector of directional information by the target shooting distance (e.g., 5 meters). Therefore, the target shooting position of the virtual camera = center coordinates + (0.6, 0.8, 0) × 5 = center coordinates + (3, 4, 0). Taking the magnitude of (3, 4, 0), we get √(3² + 4² + 0²) = 5 meters. This means the virtual camera is precisely positioned 5 meters from the center coordinates, and the shooting view completely covers the parking space and surrounding environment, allowing the user to clearly judge the distance between the vehicle and the parking space.

[0042] S340: Displays images captured by the virtual camera at the target shooting location.

[0043] In one possible implementation, a virtual camera captures images at the target location according to predefined parameters, rendering the three-dimensional elements of the virtual world into two-dimensional images to fit the display size of the in-vehicle screen. These two-dimensional images are then transmitted to the in-vehicle human-machine interface terminal (such as the central control screen) for display. The user then sees a virtual parking perspective aligned with the real environment. For example, the in-vehicle screen clearly shows the alignment of the virtual target vehicle with the virtual parking space, with their positional relationship perfectly matching the real target vehicle and the real parking space. This helps the user determine whether adjustments to the steering wheel, accelerator / brake, etc., are necessary.

[0044] Optionally, the multiple environmental element information in the above method includes at least: multiple raw environmental data and environmental attribute information. The raw environmental data is collected by multiple preset vehicle-mounted sensors.

[0045] Figure 4 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 2 .like Figure 4 As shown, the above method renders multiple restored world elements corresponding to multiple environmental element information in virtual space based on multiple environmental element information around the target vehicle, including: S410. Based on multiple raw environmental data around the target vehicle, classify the multiple raw environmental data to obtain multiple different categories of raw environmental data.

[0046] In one possible implementation, since the raw environmental data collected by various pre-set vehicle sensors (such as surround-view cameras, millimeter-wave radar, etc.) is massive and unclassified, it is necessary to first split and classify the multiple raw environmental data so that the information of the same type of environmental element corresponds to the same type of raw environmental data, thereby obtaining raw environmental data of different categories.

[0047] Since different types of environmental element information have different importance to the parking perspective (such as obstacles and parking lines being the main basis for judgment, while road lines are auxiliary references), by classifying and processing them, we can ensure that key environmental element information is not missed in the subsequent rendering process, and avoid redundant rendering of environmental element information that is not related to parking, thus laying the foundation for completely replicating the real parking environment.

[0048] S420. Transform the initial coordinate system of multiple different categories of original environmental data to the unified world coordinate system of the virtual space to obtain the world coordinates of each environmental element information.

[0049] The unified world coordinate system uses the center of mass of the target vehicle as its origin.

[0050] In one possible implementation, because different preset vehicle-mounted sensors have different initial coordinate systems (e.g., surround-view cameras have their own image coordinate system, millimeter-wave radar has its own radar coordinate system), directly using them would lead to confusion regarding the positions of environmental elements. Therefore, all categorized raw environmental data needs to be uniformly converted to a virtual space world coordinate system with the target vehicle's centroid as the origin. This ensures that the world coordinates of all environmental elements directly reflect their distance and orientation relative to the target vehicle, guaranteeing that the environmental element information reconstructed in the virtual space is completely consistent with the relative positions of the vehicle and its environment in the real environment, providing a reference for subsequent calculations of the target's shooting distance.

[0051] For example, the coordinates of the parking line images captured by the surround-view camera and the three-dimensional coordinates of the obstacles detected by the millimeter-wave radar are both converted into coordinates in a virtual space world coordinate system with the target vehicle's center of mass as the origin. The axes of this world coordinate system are defined as follows: the X-axis is positive along the direction of the target vehicle's movement; the Y-axis is positive along the side of the target vehicle, for example, the left side of the target vehicle is positive; and the Z-axis is positive perpendicular to the ground and upwards.

[0052] S430. Based on the environmental attribute information and world coordinates of multiple environmental elements, render the corresponding multiple restored world elements in the virtual space.

[0053] In one possible implementation, after obtaining the world coordinates of each environmental element, the intelligent driving domain controller transmits these world coordinates to the virtual space of the in-vehicle human-machine interface terminal. In this virtual space, the spatial location of each environmental element is first determined based on its world coordinates, and then, based on environmental attribute information, each environmental element is assigned unique morphological characteristics. By locating the spatial position based on world coordinates and considering the environmental attribute information of each environmental element, the spatial distribution patterns and core characteristics of the real parking environment can be fully captured, achieving a complete replication of the target vehicle's real parking environment. This provides data support for subsequent calculations of the virtual camera's directional indication information and the target shooting distance.

[0054] The method for determining the target shooting position of a virtual camera provided in this application involves multiple environmental element information, which can be composed of at least multiple original environmental data and environmental attribute information. Based on multiple original environmental data around the target vehicle, the multiple original environmental data are classified to obtain multiple categories of original environmental data. The initial coordinate system of the multiple categories of original environmental data is transformed to a unified world coordinate system in the virtual space to obtain the world coordinates of each environmental element information. According to the environmental attribute information and world coordinates of the multiple environmental element information, multiple corresponding restored world elements are rendered in the virtual space. Thus, by classifying multiple original environmental data, unifying the coordinates of each environmental element information, and associating them with environmental attribute information, and then rendering the corresponding multiple restored world elements, this application ensures that the restored world elements in the virtual space fully retain the distribution differences of the real parking environment, providing a data foundation for dynamically adapting the field of view.

[0055] Optionally, Figure 5 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 3 .like Figure 5 As shown, the above method determines the directional information corresponding to multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to capture multiple restored world elements based on the element position information of multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset viewing angle, including: S510. Based on the element position information of multiple restored world elements in the virtual space, determine the minimum coordinate point, the maximum coordinate point, and the center coordinate point.

[0056] In one possible implementation, since the positional information of multiple restored world elements in the virtual space is in three-dimensional world coordinates, the minimum coordinate point, the maximum coordinate point, and the center coordinate point are determined based on the world coordinates of these multiple restored world elements. These three points are used to delineate the spatial distribution range of all restored world elements, thereby clarifying the main area that the virtual camera needs to capture.

[0057] The minimum coordinate point represents the minimum combination of world coordinate values ​​for all reconstructed world elements; it can represent the lower left boundary of the environmental element information distribution. The maximum coordinate point represents the maximum combination of world coordinate values ​​for all reconstructed world elements; it can represent the upper right boundary of the environmental element information distribution. The center coordinate point represents the midpoint between the minimum and maximum coordinate points; it can represent the geometric center of the environmental element information distribution.

[0058] S520. Based on the camera parameters of the virtual camera in the virtual space, determine the horizontal and vertical field of view of the virtual camera.

[0059] The visual field angles include the longitudinal visual field angle and the lateral visual field angle; the longitudinal visual field angle is the visible angle in the vertical direction; the lateral visual field angle is the visible angle in the horizontal direction.

[0060] In one possible implementation, the horizontal and vertical field of view angles of the virtual camera are determined based on the camera parameters of the virtual camera; that is, the horizontal and vertical viewing angles of the virtual camera. A larger horizontal and vertical field of view angle allows the virtual camera to cover a wider area at the same distance, making it suitable for shooting large areas; a smaller horizontal and vertical field of view angle allows the virtual camera to cover a narrower area at the same distance, making it suitable for shooting local details.

[0061] For example, if the horizontal field of view is determined to be 60° and the vertical field of view is 45° based on the camera parameters of the virtual camera, it means that at a certain distance, the virtual camera can cover environmental element information within a 60° range in the horizontal direction and within a 45° range in the vertical direction.

[0062] S530: Based on the minimum coordinate point, maximum coordinate point, center coordinate point, preset observation angle, and horizontal and vertical field of view, determine the directional indication information corresponding to multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot multiple restored world elements.

[0063] In one possible implementation, a direction vector pointing to the center coordinate point can be generated based on the center coordinate point and a preset viewing angle, avoiding viewpoint shift. This direction vector serves as the directional indicator information for multiple recreated world elements in the virtual space. For example, if the preset viewing angle is a 45° overhead view, the direction vector needs to point from the virtual camera's position to the center coordinate point, forming a 45° angle with the horizontal plane. This ensures that the viewing angle conforms to the user's judgment habits and aligns with the geometric center of environmental element information, preventing local omissions caused by viewing angle shift.

[0064] Then, based on the minimum coordinate point, maximum coordinate point, center coordinate point, and combined with the horizontal and vertical field of view angles and direction indication information, the target shooting distance for the virtual camera to shoot multiple elements of the restored world can be calculated through preset trigonometric functions, avoiding deviations caused by human estimation.

[0065] The method for determining the target shooting position of the virtual camera provided in this application first determines the minimum coordinate point, maximum coordinate point, and center coordinate point based on the element position information of multiple restored world elements in the virtual space. The minimum coordinate point, maximum coordinate point, and center coordinate point are used to define the boundary of multiple environmental element information, ensuring that the calculation of the target shooting distance is based on the premise of covering the boundary. This ensures that all restored world elements can fall into the view frustum of the virtual camera, avoiding user misjudgment due to the inability to capture edge restored world elements from a fixed position viewpoint. Then, based on the camera parameters of the virtual camera in the virtual space, the horizontal and vertical field of view angles of the virtual camera are determined. Finally, based on the minimum coordinate point, maximum coordinate point, center coordinate point, preset observation angle, and horizontal and vertical field of view angles, the directional indication information corresponding to multiple restored world elements in the virtual space and the target shooting distance of the virtual camera for shooting multiple restored world elements are determined. Therefore, when determining the direction indication information and target shooting distance, this application relies on the confirmation of the minimum coordinate point, maximum coordinate point, center coordinate point, preset observation angle, and horizontal and vertical field of view of environmental element information. This enables the confirmation system to automatically adapt to all parking environments, improves environmental adaptability, and thus changes the camera parameters of the virtual camera from passively accepting fixed values ​​to actively adapting to the environmental distribution, thereby improving the accuracy and practicality of the parking perspective.

[0066] Optionally, in the above method, the element position information is the world coordinates of multiple restored world elements, and the world coordinate system has been predefined, that is, the X-axis is the forward direction, the Y-axis is the lateral distance, and the Z-axis is the height above the ground.

[0067] Figure 6 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 4 .like Figure 6 As shown, the above method determines the minimum coordinate point, maximum coordinate point, and center coordinate point based on the element position information of multiple restored world elements in virtual space, including: S610. Based on the preset coordinate axes, filter the global extreme values ​​in the world coordinates corresponding to multiple restored world elements to obtain the minimum coordinate point and the maximum coordinate point.

[0068] The preset coordinate axis can be selected according to the actual situation. For example, the preset coordinate axis can be selected as a three-dimensional coordinate axis, that is, a coordinate axis composed of the X-axis, Y-axis and Z-axis.

[0069] In one possible implementation, based on the element position information (i.e., world coordinates) of multiple reconstructed world elements in virtual space, the minimum and maximum values ​​of the world coordinates of multiple reconstructed world elements corresponding to each of the preset three-dimensional coordinate axes are extracted. That is, the global maximum and global minimum values ​​of multiple reconstructed world elements are found according to each axis. Then, the minimum coordinate point Min is determined based on the combination of the minimum values ​​of each axis. , , The minimum coordinate point Min can also represent the lower bound anchor point of the virtual space distribution. And based on the combination of the maximum values ​​for each axis, the maximum coordinate point Max is determined. , , This maximum coordinate point can also represent the upper limit anchor point of the virtual space distribution.

[0070] S620. Determine the center coordinates of multiple restored world elements based on the minimum and maximum coordinate points.

[0071] The maximum horizontal span is the maximum value between the preset X-axis span and the preset Y-axis span, and the vertex corresponding to the maximum horizontal span from the center coordinate point is determined as the preset bottom plane vertex A.

[0072] In one possible implementation, the center coordinate point O is determined based on the midpoint between the minimum coordinate point Min and the maximum coordinate point Max, i.e., the center coordinate point O is ( + ) / 2, + ) / 2, + (2) The central coordinate point O represents the geometric center of the distribution of all environmental element information in the virtual space, providing an anchor point for aligning the shooting direction with the target area. Finally, the spatial distribution range of all restored world elements is delineated based on the minimum coordinate point Min, the maximum coordinate point Max, and the central coordinate point O, thereby clarifying the main areas that the virtual camera needs to shoot.

[0073] The method for determining the target shooting position of a virtual camera provided in this application uses the world coordinates of multiple reconstructed world elements as element position information. Based on preset coordinate axes, the global extrema of the world coordinates corresponding to the multiple reconstructed world elements are filtered to obtain the minimum and maximum coordinate points. The center coordinate point of the multiple reconstructed world elements is determined based on the minimum and maximum coordinate points. This center coordinate point serves as the anchor point for the shooting direction, ensuring that the virtual camera is always aligned with the geometric center of the multiple environmental element information. This avoids excessive magnification or reduction in local areas due to perspective shift, allowing users to intuitively judge the relative position of the target vehicle and the environment, ensuring accurate perspective. Therefore, this application objectively reflects the true distribution range of multiple environmental element information based on the minimum, maximum, and center coordinate points. When calculating the target shooting distance, this true distribution range will be used as a benchmark to ensure that the multiple environmental element information can be completely covered, improving scene adaptability.

[0074] Optionally, the camera parameters of the virtual camera in the above method include at least: preset physical size and preset focal length.

[0075] Here, the preset physical size `sensor_width` is the preset effective width of the image sensor simulated by the virtual camera, representing the horizontal physical size of the virtual camera's image sensor. In contrast, the horizontal physical size of the image sensor in a real vehicle camera (`Camera.SensorSize.x`) is a fixed parameter calibrated at the factory, describing only the horizontal dimension of the image sensor in the real vehicle camera, and is independent of the vertical dimension (`Camera.SensorSize.y`). Therefore, the preset physical size `sensor_width` can be directly represented by the horizontal physical size of the image sensor in the real vehicle camera (`Camera.SensorSize.x`), i.e., `sensor_width = Camera.SensorSize.x`, ensuring that the camera parameters of the virtual camera and the real hardware are consistent.

[0076] In a real vehicle camera, the horizontal physical size of the image sensor, Camera.SensorSize.x, includes two independent dimensions: the vertical physical height, Camera.ScreenHeight, and the horizontal physical width, Camera.ScreenWidth.

[0077] Wherein, the horizontal physical width Camera.ScreenWidth represents the physical size of the image sensor of the actual vehicle camera in a single dimension in the horizontal direction; the vertical physical height Camera.ScreenHeight represents the physical size of the image sensor of the actual vehicle camera in a single dimension in the vertical direction.

[0078] It should be noted that, in addition to the preset physical size sensor_width, there is another parameter: the preset effective height of the image sensor simulated by the virtual camera. , which represents the physical size of the image sensor of the virtual camera in the vertical direction.

[0079] The preset focal length (focal_length) is the simulated optical lens focal length of the virtual camera, reflecting the lens's ability to converge light. It can be directly represented by the focal length of a real vehicle camera's optical lens (Camera.FocalLength), ensuring that the virtual camera's parameters are consistent with the real hardware. Furthermore, the preset focal length (focal_length) is inversely proportional to the horizontal and vertical field of view; that is, the longer the preset focal length (focal_length), the narrower the horizontal and vertical field of view; conversely, the shorter the preset focal length (focal_length), the wider the horizontal and vertical field of view.

[0080] It should be noted that in this application, the preset physical size and preset focal length of the virtual camera are directly obtained from the horizontal physical size and focal length of the image sensor of the real vehicle-mounted camera. This is to ensure that the aspect ratio (sensor_width / width) of the image sensor simulated by the virtual camera is consistent. The aspect ratio (screenWidth / screenHeight) of the in-vehicle human-machine interface terminal is precisely matched to ensure that the rendered parking image is free of stretching and distortion, allowing users to accurately judge the spatial relationship between the vehicle and its environment. Here, screenWidth is the screen width of the in-vehicle human-machine interface terminal; screenHeight is the screen height of the in-vehicle human-machine interface terminal.

[0081] Figure 7 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 5 .like Figure 7 As shown, the above method determines the horizontal and vertical field of view angles of the virtual camera based on the camera parameters of the virtual camera in the virtual space; including: S710: Calculate the longitudinal field of view of the virtual camera based on the preset physical dimensions and preset focal length in the camera parameters.

[0082] In one possible implementation, since the calculation of the longitudinal field of view angle α follows standard optical imaging principles, the core of which is to precisely match the aspect ratio of the image sensor simulated by the virtual camera with the screen resolution aspect ratio of the in-vehicle human-machine interaction terminal. Based on the preset physical size sensor_width, the preset effective height can be determined. Then preset the effective height In conjunction with the preset focal length (focal_length), the vertical field of view is locked to provide crucial environmental element information (such as obstacle height and ground markings) for parking scenarios through precise vertical coverage, ensuring that no critical details in the vertical space are missed. Simultaneously, hardware performance constraints are provided for the lateral field of view (θ). Because the horizontal physical width and vertical physical height of the image sensor have a fixed aspect ratio, the size of the vertical field of view (α) limits the reasonable range of the lateral field of view (θ), preventing an imbalance in the ratio of the horizontal and vertical field of view (θ). This ensures that adjustments to the lateral field of view (θ) do not exceed the optical limits of the image sensor, fundamentally avoiding problems such as image stretching and edge distortion. Here, hardware performance refers to the limits of the physical dimensions of the actual vehicle-mounted camera and the light-gathering ability of the optical lens.

[0083] S720. Determine the minimum horizontal field of view of the virtual camera based on the vertical field of view angle.

[0084] Among them, the minimum horizontal field of view β is the minimum horizontal field of view angle that covers multiple elements of the restored world, thus avoiding field of view redundancy.

[0085] In one possible implementation, based on the vertical field of view angle α, a preset reasonable range for the horizontal field of view angle θ is first set. This preset reasonable range is a range derived from both the vertical field of view angle α and the aspect ratio of the image sensor (aspect ratio = horizontal physical width / vertical physical height) to avoid image distortion caused by an imbalance in the ratio of the horizontal and vertical field of view angles. The basis for determining the preset reasonable range for the horizontal field of view angle θ is: according to the principles of optical imaging, the ratio of the horizontal and vertical field of view angles must match the aspect ratio of the image sensor. For example, when the aspect ratio of the image sensor is 18:9, the horizontal field of view angle θ should be approximately twice the vertical field of view angle α; otherwise, image stretching will occur (e.g., the horizontal width is too narrow, causing the image to be stretched vertically, or the vertical width is too narrow, causing the image to be compressed horizontally).

[0086] For example, if the vertical field of view α is 45° and the aspect ratio of the image sensor is 18:9, then the preset reasonable range of the horizontal field of view θ should be set around 45°×(18 / 9)≈90°. It should not be much less than 80° to avoid vertical stretching, nor much greater than 80° to avoid horizontal stretching.

[0087] Then, the minimum lateral visual field angle β is determined based on the intersection of the lateral visual field angle θ and the lower limit of a preset reasonable range. Specifically, this can be divided into two cases: If the horizontal field of view θ is greater than or equal to the lower limit of the preset reasonable range: this means that the horizontal field of view θ can cover all environmental element information and also meets the ratio constraints of the horizontal and vertical field of view. At this time, the minimum horizontal field of view β = the horizontal field of view θ, that is, it just covers the area and there is no distortion. If the horizontal field of view θ is less than the lower limit of the preset reasonable range: this means that although the horizontal field of view θ can cover all environmental element information, the imbalance of the ratio of the horizontal and vertical field of view will cause image distortion. At this time, the minimum horizontal field of view β = the lower limit of the preset reasonable range to ensure that the ratio of the horizontal and vertical field of view is normal. Since the lower limit of the preset reasonable range is greater than or equal to the horizontal field of view θ, the coverage of the minimum horizontal field of view β can include all environmental element information.

[0088] The lower limit of the preset reasonable range is the minimum value of the preset reasonable range. Its function is to ensure that the horizontal field of view θ is not lower than the minimum value that matches the vertical field of view α and the width-to-height ratio of the image sensor, so as to avoid image distortion caused by exceeding the hardware capabilities from the root.

[0089] The method for determining the target shooting position of a virtual camera provided in this application includes at least the following camera parameters: preset physical size and preset focal length. Based on the preset physical size and preset focal length in the camera parameters, the longitudinal field of view of the virtual camera is calculated to ensure complete coverage of key environmental element information in the vertical direction, thereby assisting users in judging the vertical space margin when parking. Then, based on the longitudinal field of view, combined with the sensor aspect ratio constraint in the preset physical size of the camera parameters, the ratio of the longitudinal and horizontal field of view is ensured to match, avoiding setting an excessively large horizontal field of view due to simply pursuing coverage of all environmental element information, which would cause image edge stretching and distortion. This ensures that when the image captured by the virtual camera is displayed on the in-vehicle screen, the horizontal and vertical ratios are coordinated, consistent with the real observation experience, reducing the user's spatial judgment error and improving parking safety. Furthermore, the minimum horizontal field of view is designed to just cover all environmental element information with the sole objective of eliminating any extra field of view. Compared with a fixed horizontal field of view, this application allows the field of view to focus on key environmental element information, improving the user's judgment efficiency.

[0090] Figure 8 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 6 .like Figure 8 As shown, the above method calculates the longitudinal field of view of the virtual camera based on the preset physical size and preset focal length in the camera parameters, including: S810. Determine the screen resolution aspect ratio of the vehicle-mounted human-machine interaction terminal based on the preset physical dimensions.

[0091] In one possible implementation, to ensure the compatibility of the screen display specifications of the real vehicle camera and the vehicle human-machine interaction terminal, the screen resolution width screenWidth of the vehicle human-machine interaction terminal is the same as the horizontal physical width Camera.ScreenWidth, i.e., screenWidth=Camera.ScreenWidth; the screen resolution height screenHeight of the vehicle human-machine interaction terminal is the same as the vertical physical height Camera.ScreenHeight, i.e., screenHeight=Camera.ScreenHeight; then the screen resolution aspect ratio of the vehicle human-machine interaction terminal is screenWidth / screenHeight (equivalent to Camera.ScreenWidth / Camera.ScreenHeight), and since there is a one-to-one matching relationship between the aspect ratio of the image sensor simulated by the virtual camera and the screen resolution aspect ratio of the vehicle human-machine interaction terminal, it can be expressed by the following formula (1): = Formula (1) S820 determines the effective imaging height of the virtual camera based on the screen resolution aspect ratio and preset physical dimensions.

[0092] Among them, the effective imaging height is the vertical physical dimension of the screen display image generated by the virtual camera during the imaging process to adapt to the screen display image of the vehicle-mounted human-machine interaction terminal. It is the core parameter (unit: mm) to ensure that the imaging image can adapt to the vertical display of the screen of the vehicle-mounted human-machine interaction terminal.

[0093] In one possible implementation, the preset physical size `sensor_width` determines the horizontal light capture range, and together with the preset focal length `focal_length`, defines the width of the horizontal field of view angle `θ`, ensuring that the horizontal field of view coverage is consistent with optical principles. Meanwhile, the preset effective height `sensor_height` determines the vertical light capture range, and together with the preset focal length `focal_length`, determines the width of the vertical field of view angle `α`, ensuring that the vertical field of view coverage is consistent with optical principles. Therefore, the effective imaging height of the virtual camera can be calculated using the above formula (1). It can be represented by the following formula (2).

[0094] Formula (2) For example, if the preset physical size sensor_width is 8mm and the screen resolution aspect ratio of the in-vehicle human-machine interaction terminal is 16:9, then the effective imaging height is... =8÷(16 / 9)=4.5mm, ensuring that the vertical physical size of the image matches the vertical resolution of the screen of the vehicle-mounted human-machine interaction terminal, avoiding vertical stretching or cropping of the image.

[0095] S830: Determine the longitudinal field of view of the virtual camera based on the preset focal length and effective imaging height of the virtual camera.

[0096] In one possible implementation, since the calculation of the field of view strictly follows standard optical imaging principles, it is based on the virtual camera's preset focal length (focal_length) and effective imaging height. The longitudinal field of view α of the virtual camera is determined according to the following formula (3).

[0097] α = 2 × arctan( ÷focal_length) formula (3) Among them, effective imaging height The vertical light capture range is determined, and together with the preset focal length (focal_length), it locks the width of the vertical field of view α, i.e., the effective imaging height. The larger the focal length and the shorter the preset focal length, the larger the vertical field of view α and the wider the vertical coverage; conversely, the smaller the effective imaging height, the shorter the effective imaging height. The smaller the value, the longer the preset focal length (focal_length), the smaller the vertical field of view (α), and the more focused the vertical local area.

[0098] The method for determining the target shooting position of the virtual camera provided in this application determines the resolution aspect ratio of the screen of the in-vehicle human-machine interaction terminal based on preset physical dimensions; then, based on the screen's resolution aspect ratio and preset physical dimensions, it determines the effective imaging height of the virtual camera, ensuring that the vertical physical dimensions of the image accurately match the vertical resolution of the screen. Subsequent parking images rendered based on this method are displayed on the in-vehicle screen without vertical stretching or cropping, conforming to the user's real visual judgment habits and reducing spatial perception errors. Furthermore, based on the virtual camera's preset focal length and effective imaging height, the vertical field of view of the virtual camera is determined to accurately cover vertical shooting requirements and prevent the omission of environmental element information.

[0099] Figure 9 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 7 .like Figure 9 As shown, the method described above determines the minimum lateral field of view of the virtual camera based on the longitudinal field of view, including: S910: Based on a preset field-of-view conversion method, the longitudinal field of view of the virtual camera is converted into the initial lateral field of view.

[0100] The preset view conversion method can be selected according to the actual situation. For example, the view conversion method can be implemented through the API component in the Unity engine, which is configured with a way to convert the vertical view angle to the initial horizontal view angle.

[0101] In one possible implementation, the essence of this preset field-of-view conversion method is based on the aspect ratio conversion method, and the conversion basis is the aspect ratio of the screen resolution of the vehicle human-machine interaction terminal (or the aspect ratio of the analog image sensor). That is, the screen resolution aspect ratio is known to be screenWidth / screenHeight (equivalent to Camera.ScreenWidth / Camera.ScreenHeight), and the longitudinal field of view angle obtained according to the above formula (3) is α. Then, the initial lateral field of view angle is determined by the following formula (4). .

[0102] =α×(screenWidth / screenHeigh) formula (4) For example, if the vertical viewing angle α is 45° and the screen resolution aspect ratio is 16:9 (≈1.78), then the initial horizontal viewing angle is... =45°×1.78≈80°, ensuring the horizontal and vertical viewing angle ratio matches the screen resolution's aspect ratio to avoid image stretching. S920. Based on the initial horizontal field of view and the preset lens offset of the virtual camera, determine the minimum horizontal field of view of the virtual camera using the preset offset calculation method.

[0103] The preset lens shift, lensShift.x, refers to a slight angular offset in the horizontal direction (XY plane) of the virtual camera to adapt to specific parking observation needs (such as focusing on the right-hand parking space or avoiding areas obstructed by the vehicle body). Its core function is to optimize the focusing direction of the observation viewpoint. The preset lens shift, lensShift.x, can be configured according to actual conditions, and its value ranges from [-0.5, 0.5]. It should be noted that, for ease of calculation, the preset lens shift, lensShift.x, can be selected as 0.

[0104] Due to the virtual camera lens shift, the original initial horizontal field of view... It will tilt in the offset direction, which may cause elements of the restored world on the other side edge (such as the left side line of the parking space, distant obstacles) to be out of the field of view. Therefore, it is necessary to expand the horizontal field of view by using the preset lens offset amount lensShift.x to compensate for this gap and ensure that all environmental element information is fully covered.

[0105] In one possible implementation, based on the initial lateral field of view angle and the preset lens offset of the virtual camera Based on the preset offset calculation method, the minimum horizontal field of view β of the virtual camera is determined by the following formula (5).

[0106] β= / 2-( -( ×(1-Mathf.Abs(lensShift.x)))) Formula (5) Among these requirements, the minimum lateral field of view β must not exceed the hardware performance limit of the virtual camera in the vehicle-mounted human-machine interaction terminal to avoid image distortion.

[0107] The method for determining the target shooting position of a virtual camera provided in this application is based on a preset field-of-view conversion method. This method converts the vertical field-of-view angle of the virtual camera into an initial horizontal field-of-view angle, ensuring that the ratio of the vertical to horizontal field-of-view angles matches the aspect ratio of the screen and the analog image sensor, eliminating display distortion and conforming to the user's visual judgment habits. Based on the initial horizontal field-of-view angle and the preset lens offset of the virtual camera, and using a preset offset calculation method, the minimum horizontal field-of-view angle of the virtual camera is determined. Since the preset lens offset is intended to optimize the observation focusing direction, the offset can cause the original field of view to tilt. By compensating for the field-of-view gap through the preset offset calculation method, the minimum horizontal field-of-view angle can cover edge elements that may exceed the field of view after the offset, preserving the focusing advantage of the offset perspective while ensuring that all elements of the reconstructed world are not missed.

[0108] Figure 10 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 8 .like Figure 10 As shown, the above method determines the directional information of multiple reconstructed world elements in the virtual space and the target shooting distance of the virtual camera for shooting multiple reconstructed world elements based on the minimum coordinate point, maximum coordinate point, center coordinate point, preset observation angle, and horizontal and vertical field of view angles, including: S1010. Based on the center coordinate point, minimum coordinate point, and maximum coordinate point, construct an axis-aligned 3D bounding box and determine the bottom vector of the 3D bounding box.

[0109] Wherein, the bottom vector OA is a unit vector pointing from the center coordinate point O to the preset bottom plane vertex A, and the length of the bottom vector is the straight-line distance from the center coordinate point O to the preset bottom plane vertex A. The preset bottom plane can be selected according to the actual situation. For example, the preset bottom plane can be the point with the minimum preset Z-axis coordinate min.z (such as 0).

[0110] The default bottom plane vertex A is the vertex with the smallest default Z-axis coordinate in the 3D bounding box, that is, the vertex closest to the real ground, such as Min( , , ) or Max ( , , ).

[0111] In one possible implementation, based on the center coordinate point O and the minimum coordinate point Min ( , , ), Maximum coordinates Max ( , , Based on this, a 3D cube is constructed with edges parallel to preset coordinate axes (such as the X / Y / Z axes). This 3D cube is an axis-aligned 3D bounding box. This 3D bounding box is used to enclose all elements of the reconstructed world.

[0112] The bottom vector OA is a unit direction vector pointing from the center coordinate point O to the preset bottom plane vertex A, representing only the direction. The length of the bottom vector OA is the length of the straight line in space from the center coordinate point O to the preset bottom plane vertex A, representing the actual distance between the two points. Based on the bottom vector, the direction and distance from the center coordinate point O to the real ground element information are simultaneously locked, providing a geometric reference for subsequent viewpoint adaptation and target shooting distance calculation.

[0113] S1020. Rotate the bottom vector around any axis of the preset coordinate system by a preset viewing angle to obtain the directional indication information of multiple restored world elements in the virtual space.

[0114] In one possible implementation, the bottom vector OA is rotated around any axis (such as the preset X-axis, the preset Y-axis, the lateral distance to the target vehicle, and the preset Z-axis, the height above the ground) by a preset viewing angle (such as a 45° overhead view, i.e., the virtual viewing angle desired by the user). The resulting vector is the direction indication information (direction vector OD). This direction indication information (direction vector OD) indicates the spatial orientation of multiple restored world elements (enclosed by a 3D bounding box) relative to the virtual camera at the preset viewing angle (such as a 45° overhead view). For example, if the rotated direction vector OD points to the lower left of the 3D bounding box, it means that the restored world elements are in the lower left area of ​​the virtual camera.

[0115] It should be noted that the orientation vector OD can be calculated using the Unity engine's API component (Quaternion.AngleAxis(b, Vector3.right)); OD = OA.normalized * Quaternion.AngleAxis(b, Vector3.right). Here, the orientation vector OD is the orientation indicator information, which is the vector obtained by rotating the bottom vector OA around the X-axis (Vector3.right) by an angle b, and is used to define the viewing direction of the virtual camera.

[0116] S1030. Based on the horizontal and vertical field of view angles, bottom vector, direction indication information, and preset observation angle, and based on the preset sine theorem, determine the target shooting distance for the virtual camera to shoot multiple elements of the restored world.

[0117] The preset observation angle is the angle between the bottom vector OA and the direction vector OD (direction indication information), which is b = Vector3.Angle(OD, OA).

[0118] In one possible implementation method Figure 11 This is a schematic diagram illustrating the structure of a virtual camera target shooting position provided in an embodiment of this application. Figure 11 As shown, the length and angle b of the bottom vector OA are known, and the minimum horizontal field of view β of the virtual camera is determined from the horizontal and vertical field of view angles according to the above formula (5). Then, based on the preset sine theorem, the target shooting distance length for the virtual camera to shoot multiple restored world elements is determined according to the following formula (6).

[0119] Length=|OA|× Formula (6) Among them, the target shooting distance length of the direction vector OD (direction indication information) is the straight-line distance from the virtual camera to the center coordinate point O of the 3D bounding box. This target shooting distance length means that when the virtual camera is placed at the minimum distance in the direction of the direction vector OD (direction indication information), its view frustum can just completely surround the entire 3D bounding box, so that the field of view is not wasted due to excessive distance, nor are world elements omitted due to excessive distance.

[0120] The method for determining the target shooting position of the virtual camera provided in this application constructs an axis-aligned 3D bounding box based on the center coordinate point, minimum coordinate point, and maximum coordinate point to encompass all elements of the virtual world, and determines the bottom vector of the 3D bounding box. The bottom vector is rotated around any axis of the preset coordinate system by a preset viewing angle to obtain directional indication information corresponding to multiple elements of the virtual world in the virtual space, so that the directional indication information adapts to the needs of the preset viewing angle and conforms to the user's actual observation habits. Based on the horizontal and vertical field of view angles, the bottom vector, the directional indication information, and the preset viewing angle, and using a preset sine theorem, the target shooting distance for the virtual camera to shoot multiple elements of the virtual world is determined. This target shooting distance ensures that the virtual image focuses on the core parking scene, allowing the user to quickly capture key information and improving the stability and reliability of the confirmation system.

[0121] Figure 12 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 9 .like Figure 12 As shown, the image captured by the virtual camera at the target shooting location in the above method includes: S1110. Based on the target shooting position of the virtual camera, set the orientation of the virtual camera to be from the target shooting position toward the center coordinate point.

[0122] In one possible implementation, the orientation of the virtual camera is set according to the target shooting position of the virtual camera, that is, the shooting direction of the simulated optical lens. It is explicitly set to a straight line from the target shooting position to the center coordinate point O, so as to ensure that the parking scene being shot is always in the center of the frame and to avoid the restored world elements being shifted to the edge or outside of the frame.

[0123] S1120: Render multiple restored world elements within the corresponding field of view cone range of the target shooting position, and output the rendered image to the screen of the vehicle-mounted human-machine interaction terminal.

[0124] The view frustum is the visible space of the virtual camera, a square pyramid with the virtual camera as its vertex, defined by the horizontal and vertical field of view angles and the target shooting distance.

[0125] In one possible implementation, the system performs geometric modeling, texture mapping, and other rendering processes only on multiple real-world elements falling within the view frustum. During this rendering process, the proportional relationships of the real-world elements are maintained to ensure that the virtual image matches the proportions of the real environment. The rendered image is then output according to the screen resolution and aspect ratio of the in-vehicle human-machine interface terminal, ensuring that the output image size perfectly matches the screen without stretching or cropping.

[0126] The method for determining the target shooting position of the virtual camera provided in this application sets the orientation of the virtual camera to the direction from the target shooting position toward the center coordinate point, ensuring that the core area of ​​the parking scene is in the center of the screen; it renders multiple restored world elements within the field of view cone corresponding to the target shooting position, and adapts and outputs the rendered image to the screen of the in-vehicle human-machine interaction terminal without manual adjustment, ensuring a consistent visual experience for all users and improving the safety and convenience of parking operations.

[0127] To facilitate understanding of the method for determining the target shooting position of the virtual camera described above, this application embodiment also provides a flowchart example of the method for determining the target shooting position of the virtual camera, which will be further described below with reference to the accompanying drawings. Figure 13 A flowchart illustrating a method for determining the target shooting position of a virtual camera provided in this application embodiment. Figure 10 .like Figure 13 As shown in the illustration, the embodiments provided in this application provide... Figure 7 It may include: S1201. Render multiple restored world elements corresponding to the multiple environmental element information around the target vehicle in virtual space.

[0128] Specifically, when the target vehicle is in a parking scenario, the intelligent driving domain control mentioned above acquires information on multiple environmental elements of the real parking environment around the target vehicle and transmits this information to the virtual space of the control device. In this virtual space, the mesh renderer renders the information on multiple environmental elements in real time, generating multiple corresponding restored world elements. The relative positions of these restored world elements are completely consistent with the real parking environment of the target vehicle, thereby transforming the real parking environment into a calculable digital model in the virtual space, providing a reliable data foundation for the subsequent calculation of the target shooting position of the virtual camera.

[0129] S1202. Based on the element position information of multiple restored world elements, select the minimum and maximum coordinate points and construct an axis-aligned 3D bounding box.

[0130] Specifically, based on the element position information (i.e., world coordinates) of multiple restored world elements in virtual space, the minimum and maximum values ​​of the world coordinates of multiple restored world elements corresponding to each of the preset three-dimensional coordinate axes are extracted. That is, the global maximum and global minimum values ​​of multiple restored world elements are found according to each axis. Then, the minimum coordinate point Min is determined based on the combination of the minimum values ​​of each axis. , , ), and the maximum coordinate point Max ( , , Then, based on the minimum coordinate point Min ( , , ), and the maximum coordinate point Max ( , , Determine the center coordinate point O. Based on the center coordinate point O and the minimum coordinate point Min... , , ), Maximum coordinates Max ( , , Based on this, a 3D cube is constructed with edges parallel to preset coordinate axes (such as the X / Y / Z axes). This 3D cube is an axis-aligned 3D bounding box. This 3D bounding box is used to enclose all elements of the reconstructed world.

[0131] S1203. Determine the effective imaging height of the virtual camera based on the preset physical dimensions.

[0132] Specifically, since the aspect ratio of the screen resolution matches the aspect ratio of the image sensor simulated by the virtual camera, the effective imaging height of the virtual camera is calculated according to the above formula (2). , where sensor_width is the preset physical size.

[0133] S1204. Based on the effective imaging height and preset focal length, confirm the longitudinal field of view angle.

[0134] Specifically, since the calculation of the field of view strictly follows the standard optical imaging principles, it is based on the virtual camera's preset focal length and effective imaging height. Then, the longitudinal field of view α of the virtual camera is determined according to the above formula (3).

[0135] S1205. Based on the longitudinal field of vision angle, confirm the initial lateral field of vision angle.

[0136] Specifically, this is achieved through a Unity engine API component, which is configured with a method to convert the vertical field of view to the initial horizontal field of view (i.e., horizontalFov(initial horizontal field of view) = Camera.VerticalToHorizontalFieldOfView).

[0137] S1206. Determine the minimum horizontal field of view of the virtual camera based on the initial horizontal field of view and the preset lens offset of the virtual camera.

[0138] Specifically, based on the initial lateral field of view angle and the preset lens offset of the virtual camera Based on the preset offset calculation method, the minimum horizontal field of view β of the virtual camera is determined by the above formula (5).

[0139] S1207. Calculate the bottom vector and orientation information of the 3D bounding box.

[0140] Specifically, based on the center coordinates O and the minimum coordinates Min of the 3D bounding box... , , ), Maximum coordinates Max ( , , ), confirm the bottom vector OA, which is a unit direction vector from the center coordinate point O to the preset bottom plane vertex A, representing only the direction. The length of the bottom vector OA is the length of the spatial straight line from the center coordinate point O to the preset bottom plane vertex A, representing the actual distance between the two points.

[0141] Then, the orientation vector OD is calculated using the Unity engine's API component (Quaternion.AngleAxis(b, Vector3.right)); OD = OA.normalized * Quaternion.AngleAxis(b, Vector3.right). Here, the orientation vector OD is the orientation indicator information, which is the vector obtained by rotating the bottom vector OA around the X-axis (Vector3.right) by an angle b, and is used to define the viewing direction of the virtual camera.

[0142] S1208. Determine the target shooting distance for the virtual camera to shoot multiple elements of the restored world.

[0143] Specifically, refer to Figure 11 Given the length and angle b of the bottom vector OA, and according to the above formula (5), the minimum horizontal field of view β of the virtual camera is determined from the horizontal and vertical field of view angles. Then, based on the preset sine theorem, the target shooting distance length for the virtual camera to shoot multiple restored world elements is determined according to the above formula (6).

[0144] S1209. Determine the target shooting position of the virtual camera in the virtual space.

[0145] Specifically, using the center coordinates of multiple elements' positions in the virtual space as a reference, and combining directional information with the target shooting distance, the target shooting position of the virtual camera is determined through vector operations. Then, a shot is taken based on this target shooting position, and the image is rendered and displayed on the screen of the in-vehicle human-machine interface terminal.

[0146] The method for determining the target shooting position of a virtual camera provided in this application renders multiple restored world elements corresponding to multiple environmental element information around the target vehicle in virtual space; selects the minimum and maximum coordinate points based on the element position information of the multiple restored world elements, and constructs an axis-aligned 3D bounding box; determines the effective imaging height of the virtual camera based on preset physical dimensions; confirms the longitudinal field of view angle based on the effective imaging height and preset focal length; confirms the initial lateral field of view angle based on the longitudinal field of view angle; determines the minimum lateral field of view angle of the virtual camera based on the initial lateral field of view angle and the preset lens offset of the virtual camera; calculates the bottom vector and direction indication information of the 3D bounding box to determine the target shooting distance of the virtual camera to shoot multiple restored world elements, and the target shooting position of the virtual camera in virtual space. Therefore, this application calculates a direction vector with an angle of the horizontal cross-section of the 3D bounding box as the preset viewing angle, and determines the target shooting distance where a camera's viewing cone just surrounds the 3D bounding box. By multiplying the direction vector by the target shooting distance, the world coordinates of the virtual camera are obtained. These world coordinates are the world coordinates of the parking world previewed from the preset viewing angle. This allows the cone camera to see a 3D bounding box at a preset viewing angle with minimal cost, enabling synchronous updates of the virtual perspective as the target vehicle moves. This meets the real-time requirements of parking scenarios, helping users quickly perceive the spatial relationship between the target vehicle and its surroundings, reducing parking difficulties caused by blind spots or incomplete field of view, reducing the risk of parking collisions, and improving the safety and convenience of parking.

[0147] Based on the same inventive concept, this application also provides a device for determining the target shooting position of a virtual camera. Since the principle of the device in this application is similar to the method for determining the target shooting position of a virtual camera described above, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0148] Figure 14 This is a schematic diagram of a device for determining the target shooting position of a virtual camera, provided in an embodiment of this application. Figure 14 As shown, the virtual camera target shooting position determination device 1300 is applied to a parking scenario. The virtual camera target shooting position determination device 1300 may include: The rendering module 1301 is used to render multiple restored world elements corresponding to multiple environmental element information in the virtual space when the target vehicle is in a parking scene, based on multiple environmental element information around the target vehicle. The first determining module 1302 is used to determine the direction indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance of the virtual camera to shoot the multiple restored world elements based on the element position information of multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space and the preset observation angle; the direction indication information is used to indicate the relative positional relationship between the virtual camera and the multiple restored world elements. The second determining module 1303 is used to determine the target shooting position of the virtual camera in the virtual space based on the direction indication information and the target shooting distance; Display module 1304 is used to display images captured by a virtual camera at the target shooting location.

[0149] In one optional implementation, the multiple environmental element information includes at least: multiple original environmental data and environmental attribute information; the original environmental data is collected by a preset vehicle-mounted sensor; the rendering module 1301 is specifically used for: classifying the original environmental data based on the multiple original environmental data around the target vehicle to obtain multiple different categories of original environmental data; transforming the initial coordinate system of the multiple different categories of original environmental data to a unified world coordinate system in the virtual space to obtain the world coordinates of each environmental element information, the unified world coordinate system taking the centroid of the target vehicle as the origin; and rendering the corresponding multiple restored world elements in the virtual space according to the environmental attribute information and world coordinates of the multiple environmental element information.

[0150] In one optional implementation, the first determining module 1302 is specifically used to: determine the minimum coordinate point, the maximum coordinate point, and the center coordinate point based on the element position information of multiple restored world elements in the virtual space; determine the horizontal and vertical field of view angles of the virtual camera based on the camera parameters of the virtual camera in the virtual space; and determine the directional indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot the multiple restored world elements according to the minimum coordinate point, the maximum coordinate point, the center coordinate point, the preset observation angle, and the horizontal and vertical field of view angles.

[0151] In one optional implementation, the element position information is the world coordinates of multiple restored world elements; the first determining module 1302 is specifically used to: based on a preset coordinate axis, filter the global extreme values ​​among the world coordinates corresponding to multiple restored world elements to obtain the minimum coordinate point and the maximum coordinate point; determine the center coordinate point and the maximum horizontal span of multiple restored world elements according to the minimum coordinate point and the maximum coordinate point; wherein, the maximum horizontal span is the maximum value between the preset X-axis span and the preset Y-axis span, and determine the vertex corresponding to the maximum horizontal span as the vertex of the preset bottom plane.

[0152] In one optional implementation, the camera parameters of the virtual camera include at least: a preset physical size and a preset focal length; the first determining module 1302 is specifically used to: calculate the longitudinal field of view of the virtual camera based on the preset physical size and preset focal length in the camera parameters; determine the minimum lateral field of view of the virtual camera based on the longitudinal field of view; the minimum lateral field of view is the minimum field of view angle that covers multiple restored world elements.

[0153] In one optional implementation, the first determining module 1302 is specifically used to: determine the resolution aspect ratio of the screen of the vehicle-mounted human-machine interaction terminal according to the preset physical dimensions; determine the effective imaging height of the virtual camera according to the resolution aspect ratio of the screen and the preset physical dimensions; the effective imaging height is the vertical physical dimension of the screen display image adapted to the vehicle-mounted human-machine interaction terminal generated by the virtual camera during the imaging process; and determine the vertical field of view of the virtual camera according to the preset focal length of the virtual camera and the effective imaging height.

[0154] In one optional implementation, the first determining module 1302 is specifically used to: convert the longitudinal field of view of the virtual camera into an initial lateral field of view based on a preset field of view conversion method; and determine the minimum lateral field of view of the virtual camera based on the initial lateral field of view and the preset lens offset of the virtual camera, using a preset offset calculation method.

[0155] In one optional implementation, the first determining module 1302 is specifically used to: construct an axis-aligned 3D bounding box based on the center coordinate point, the minimum coordinate point, and the maximum coordinate point, and determine the bottom vector of the 3D bounding box; the bottom vector is a unit vector pointing from the center coordinate point to a preset bottom plane vertex, and the length of the bottom vector is the straight-line distance from the center coordinate point to the preset bottom plane vertex; rotate the bottom vector around any axis of the preset coordinate axis by a preset viewing angle to obtain the directional indication information corresponding to multiple restored world elements in the virtual space; and determine the target shooting distance for the virtual camera to shoot multiple restored world elements based on the horizontal and vertical field of view angles, the bottom vector, the directional indication information, and the preset viewing angle, and based on the preset sine theorem.

[0156] In one optional implementation, the display module 1304 is specifically used to: set the orientation of the virtual camera to the direction from the target shooting position toward the center coordinate point according to the target shooting position of the virtual camera; render multiple restored world elements within the field of view frustum corresponding to the target shooting position, and output the rendered image to the screen of the vehicle-mounted human-machine interaction terminal.

[0157] It should be noted that for details not disclosed in the virtual camera target shooting position determination device of the embodiments of this application, please refer to the details disclosed in the virtual camera target shooting position determination method of the embodiments of this application, which will not be repeated here.

[0158] Optionally, embodiments of this application also provide a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor performs the steps of the method for determining the target shooting position of the virtual camera on the mobile storage medium described in the above embodiments. The specific implementation and technical effects are similar and will not be repeated here.

[0159] Optionally, this embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned related steps to implement the method for determining the target shooting position of a virtual camera provided in the above embodiment.

[0160] In this embodiment, the device, computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0161] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for determining the target shooting position of a virtual camera, characterized in that, include: When the target vehicle is in a parking scenario, based on multiple environmental element information around the target vehicle, multiple restored world elements corresponding to the multiple environmental element information are rendered in the virtual space; Based on the element position information of the multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset observation angle, the direction indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance of the virtual camera to shoot the multiple restored world elements are determined. The direction indication information is used to indicate the relative positional relationship between the virtual camera and the plurality of restored world elements; Based on the direction indication information and the target shooting distance, the target shooting position of the virtual camera in the virtual space is determined; Displays an image taken by the virtual camera at the target shooting location.

2. The method according to claim 1, characterized in that, The multiple environmental element information includes at least: multiple raw environmental data and environmental attribute information; the raw environmental data is collected by multiple preset vehicle-mounted sensors. The step of rendering multiple restored world elements corresponding to the multiple environmental element information around the target vehicle in virtual space includes: Based on multiple raw environmental data around the target vehicle, the raw environmental data is classified to obtain multiple different categories of raw environmental data; The initial coordinate system of the multiple different categories of raw environmental data is transformed to the unified world coordinate system of the virtual space to obtain the world coordinates of each environmental element information. The unified world coordinate system takes the centroid of the target vehicle as its origin. Based on the environmental attribute information of the multiple environmental elements and the world coordinates, the corresponding multiple restored world elements are rendered in the virtual space.

3. The method according to claim 1 or 2, characterized in that, The step of determining the directional indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to capture the multiple restored world elements, based on the element position information of the multiple restored world elements in the virtual space, the camera parameters of the virtual camera in the virtual space, and the preset viewing angle, includes: The minimum coordinate point, the maximum coordinate point, and the center coordinate point are determined based on the element position information of the multiple restored world elements in the virtual space. The horizontal and vertical field of view of the virtual camera are determined based on the camera parameters of the virtual camera in the virtual space; the horizontal and vertical field of view includes: vertical field of view and horizontal field of view. Based on the minimum coordinate point, the maximum coordinate point, the center coordinate point, the preset observation angle, and the horizontal and vertical field of view angles, the directional indication information corresponding to the multiple restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot the multiple restored world elements are determined.

4. The method according to claim 3, characterized in that, The element location information is the world coordinates of the plurality of restored world elements; The process of determining the minimum coordinate point, maximum coordinate point, and center coordinate point based on the element position information of the multiple restored world elements in the virtual space includes: Based on the preset coordinate axes, the global extreme values ​​in the world coordinates corresponding to the multiple restored world elements are filtered to obtain the minimum coordinate point and the maximum coordinate point; Based on the minimum and maximum coordinate points, determine the center coordinate points of the plurality of restored world elements.

5. The method according to claim 3, characterized in that, The camera parameters of the virtual camera include at least: preset physical size and preset focal length; The determination of the horizontal and vertical field of view angles of the virtual camera based on the camera parameters of the virtual camera in the virtual space includes: Based on the preset physical dimensions and preset focal length in the camera parameters, calculate the longitudinal field of view of the virtual camera; Based on the longitudinal field of view, the minimum lateral field of view of the virtual camera is determined; the minimum lateral field of view is the minimum field of view angle that covers the plurality of restored world elements.

6. The method according to claim 5, characterized in that, The step of calculating the longitudinal field of view of the virtual camera based on the preset physical size and preset focal length in the camera parameters includes: Based on the preset physical dimensions, determine the screen resolution aspect ratio of the in-vehicle human-machine interaction terminal; The effective imaging height of the virtual camera is determined based on the screen resolution aspect ratio and the preset physical dimensions; the effective imaging height is the vertical physical dimension of the screen display image generated by the virtual camera during the imaging process to adapt to the in-vehicle human-machine interaction terminal. The longitudinal field of view of the virtual camera is determined based on the preset focal length of the virtual camera and the effective imaging height.

7. The method according to claim 5, characterized in that, Determining the minimum lateral field of view of the virtual camera based on the longitudinal field of view angle includes: Based on a preset field-of-view conversion method, the longitudinal field-of-view angle of the virtual camera is converted into an initial lateral field-of-view angle; Based on the initial lateral field of view and the preset lens offset of the virtual camera, the minimum lateral field of view of the virtual camera is determined using a preset offset calculation method.

8. The method according to claim 3, characterized in that, The step of determining the directional indication information corresponding to the plurality of restored world elements in the virtual space and the target shooting distance for the virtual camera to shoot the plurality of restored world elements based on the minimum coordinate point, the maximum coordinate point, the center coordinate point, the preset observation angle, and the horizontal and vertical field of view angles includes: Based on the center coordinate point, the minimum coordinate point, and the maximum coordinate point, an axis-aligned 3D bounding box is constructed, and the bottom vector of the 3D bounding box is determined; the bottom vector is a unit vector pointing from the center coordinate point to a preset bottom plane vertex, and the length of the bottom vector is the straight-line distance from the center coordinate point to the preset bottom plane vertex; The bottom vector is rotated around any axis of the preset coordinate system by the preset viewing angle to obtain the directional indication information corresponding to the multiple restored world elements in the virtual space; Based on the horizontal and vertical field of view angles, the bottom vector, the direction indication information, and the preset observation angle, and using the preset sine theorem, the target shooting distance for the virtual camera to capture the multiple restored world elements is determined.

9. The method according to claim 1, characterized in that, The display of the image captured by the virtual camera at the target shooting location includes: Based on the target shooting position of the virtual camera, the orientation of the virtual camera is set to be from the target shooting position toward the center coordinate point; The multiple restored world elements within the corresponding field of view cone range of the target shooting position are rendered, and the rendered image is output to the screen of the vehicle-mounted human-machine interaction terminal.

10. A vehicle, characterized in that, At least including: Intelligent driving domain control and in-vehicle human-machine interaction terminal; The intelligent driving domain controller and the in-vehicle human-machine interaction terminal are communicatively connected; The vehicle-mounted human-machine interaction terminal is used to perform the method as described in any one of claims 1 to 9.

Images