Vehicle light control method and device, computer equipment, storage medium and product

Abstract

The application relates to a vehicle light control method and device, computer equipment, a computer readable storage medium, and a computer program product. The method comprises the following steps: acquiring laser point cloud data in a preset range in front of a vehicle; determining object identification information in the preset range in front of the vehicle according to the laser point cloud data; and in the case that it is determined that a target object exists in the preset range in front of the vehicle according to the object identification information, controlling the light state of the vehicle according to the relative position between the target object and the vehicle. In the case that it is determined that the target object exists in the preset range in front of the vehicle, the relative position between the target object and the vehicle is further used to control the vehicle to adopt a suitable light state, intelligent control of the light is realized, the visual range of the vehicle is ensured, and the vehicle light is prevented from having an adverse effect on the target object.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, and in particular to a vehicle lighting control method, device, computer equipment, computer-readable storage medium, and computer program product. Background Technology

[0002] With the development of vehicle technology, people are paying more and more attention to vehicle safety. When driving at night, in heavy fog, or in the rain, vehicles usually need to turn on their lights to ensure driving safety. Vehicle lights are generally divided into high beams and low beams. Low beams have a shorter beam distance and a wider beam range, while high beams have a longer beam range, but the light from high beams is more concentrated and the beam range is relatively narrower.

[0003] Currently, vehicle headlight control mainly relies on manual control by the driver. For example, when driving in a scene with poor visibility, the driver can decide whether to turn on the high beams or low beams.

[0004] However, switching between high beams and low beams can distract the driver and pose a safety hazard. Summary of the Invention

[0005] Therefore, it is necessary to provide a vehicle lighting control method, device, computer equipment, computer-readable storage medium, or computer program product that can automatically control the vehicle's lighting status and intelligently switch vehicle lights, addressing the aforementioned technical problems.

[0006] A vehicle lighting control method, comprising:

[0007] Acquire laser point cloud data within a preset range in front of the vehicle;

[0008] Based on the laser point cloud data, object recognition information within a preset range in front of the vehicle is determined;

[0009] If, based on the object recognition information, it is determined that there is a target object within a preset range in front of the vehicle, the vehicle's lighting status is controlled according to the relative position of the target object and the vehicle.

[0010] In one embodiment, determining object recognition information within a preset range in front of the vehicle based on the laser point cloud data includes: converting the laser point cloud data into two-dimensional image data; using a preprocessing neural network model to identify the two-dimensional image data and determine object feature image data in the two-dimensional image data; restoring the object feature image data into three-dimensional point cloud data; using a detection neural network model to perform feature detection on the three-dimensional point cloud data and determine object feature detection information in the three-dimensional point cloud data, wherein the object feature detection information includes the type of object; performing feature concatenation between the object feature detection information and the laser point cloud data to obtain fused feature information; and using the detection neural network model to perform feature detection again on the fused feature information to determine the object recognition information.

[0011] In one embodiment, the step of using a detection neural network model to perform feature detection on the three-dimensional point cloud data and determine object feature detection information in the three-dimensional point cloud data includes: using a detection neural network model to perform feature detection on the three-dimensional point cloud data, generating a three-dimensional bounding box, adjusting the three-dimensional bounding box to accurately frame the object features; and detecting the adjusted three-dimensional bounding box to determine the object feature detection information.

[0012] In one embodiment, the target objects include ambient vehicles and pedestrians. Controlling the vehicle's lighting state based on the relative position of the target objects and the vehicle includes: when the distance between the ambient vehicles and the vehicle in a first direction is less than a first threshold, controlling the vehicle's lighting state to low beam, wherein the first direction is the vehicle's forward direction; when the distance between the ambient vehicles and the vehicle in the first direction is greater than or equal to the first threshold, controlling the vehicle's lighting state to high beam; when the distance between the pedestrian and the vehicle in the first direction is less than a second threshold, controlling the vehicle's lighting state to low beam, wherein the second threshold is less than the first threshold; and when the distance between the pedestrian and the vehicle in the first direction is greater than or equal to the second threshold, controlling the vehicle's lighting state to high beam.

[0013] In one embodiment, the vehicle lighting control further includes: when it is determined, based on the object recognition information, that there is an object within a preset range in front of the vehicle, determining whether to issue an alarm based on the relative position of the object and the vehicle, wherein the object includes obstacles and the target object.

[0014] In one embodiment, determining whether to issue an alarm based on the relative position of the object and the vehicle includes: determining a first distance between the object and the vehicle in a first direction and a second distance in a second direction, wherein the first direction is the forward direction of the vehicle and the second direction is perpendicular to the first direction; and issuing an audible and visual alarm when the first distance is less than a third threshold or the second distance is less than a fourth threshold.

[0015] A vehicle lighting control device, comprising:

[0016] The point cloud acquisition module is used to acquire laser point cloud data within a preset range in front of the vehicle;

[0017] An object recognition module is used to determine object recognition information within a preset range in front of the vehicle based on the laser point cloud data.

[0018] The lighting control module is used to control the lighting status of the vehicle based on the relative position of the target object and the vehicle when it is determined that there is a target object within a preset range in front of the vehicle based on the object recognition information.

[0019] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the aforementioned vehicle lighting control method.

[0020] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the aforementioned vehicle lighting control method.

[0021] In one embodiment, a computer program product is provided, the computer program product including a computer program that, when executed by a processor, implements the aforementioned vehicle lighting control method.

[0022] The aforementioned vehicle lighting control method, device, computer equipment, computer-readable storage medium, and computer program product. This method first acquires laser point cloud data within a preset range in front of the vehicle, thereby obtaining environmental information within that range, facilitating subsequent analysis to determine objects in front of the vehicle. Then, based on the laser point cloud data, it determines object recognition information within the preset range in front of the vehicle, thus achieving object recognition and providing a basis for subsequent vehicle lighting control. If, based on the object recognition information, it is determined that a target object exists within the preset range in front of the vehicle, the vehicle's lighting status is controlled according to the relative position of the target object and the vehicle. Therefore, it first determines whether a target object exists in front of the vehicle based on object recognition information. It does not require controlling the vehicle lights simply because an object appears in front of the vehicle; rather, it requires determining what the object is. If it is a target object, such as a person or vehicle, then the vehicle lights need to be controlled to avoid affecting it. If a target object is determined to exist within the preset range in front of the vehicle, the vehicle further controls its lighting status based on the relative position of the target object and the vehicle, achieving intelligent lighting control, ensuring the vehicle's field of vision, and avoiding adverse effects of the vehicle's lights on the target object. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of this application or the conventional technology, the drawings used in the description of the embodiments or the conventional technology will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating a vehicle lighting control method in one embodiment;

[0025] Figure 2 This is a flowchart illustrating an object recognition method in one embodiment;

[0026] Figure 3 This is a flowchart illustrating a method for obtaining object detection information in one embodiment;

[0027] Figure 4 This is a flowchart illustrating a method for controlling lights in one embodiment;

[0028] Figure 5 This is a flowchart illustrating a method for controlling an alarm in one embodiment;

[0029] Figure 6 This is a schematic diagram of the structure of a vehicle lighting control device in one embodiment;

[0030] Figure 7This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0031] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings, which illustrate embodiments of the present application. However, the present application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that the disclosure of this application will be thorough and complete.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application.

[0033] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish one element from another.

[0034] Spatial relation terms such as “below,” “under,” “below,” “under,” “above,” “above,” etc., are used herein to describe the relationship between one element or feature shown in the figure and other elements or features. It should be understood that, in addition to the orientation shown in the figure, spatial relation terms also include different orientations of the device in use and operation. For example, if the device in the figure is flipped, the element or feature described as “below,” “under,” or “below” will be oriented “above” the other element or feature. Therefore, the exemplary terms “below” and “under” can include both above and below orientations. Furthermore, the device may also include other orientations (e.g., rotated 90 degrees or other orientations), and the spatial descriptive terms used herein will be interpreted accordingly.

[0035] It should be noted that when one element is considered to be "connected" to another element, it can be directly connected to the other element or connected to the other element through an intermediary element. Furthermore, in the following embodiments, "connection" should be understood as "electrical connection," "communication connection," etc., if there is transmission of electrical signals or data between the connected objects.

[0036] When used herein, the singular forms of “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that the terms “comprising,” “including,” or “having,” etc., specify the presence of the stated feature, whole, step, operation, component, part, or combination thereof, but do not preclude the possibility of the presence or addition of one or more other features, wholes, steps, operations, components, parts, or combinations thereof.

[0037] In one embodiment, such as Figure 1 As shown, a vehicle lighting control method is provided, which includes steps S100-120.

[0038] Step S100: Obtain laser point cloud data within a preset range in front of the vehicle.

[0039] The vehicle is equipped with a lidar system, which emits laser beams around the vehicle. These beams are reflected back after encountering obstacles, thus acquiring laser point cloud data. Specifically, lidar is a radar device that operates in the infrared and visible light bands, using laser light as its working beam. A lidar system consists of a laser transmitter, an optical receiver, a turntable, and an information processing system. The lidar emits a laser beam towards a target object and then receives the reflected beam. By comparing the emitted and reflected beams and processing the data through the information processing system, the distance and orientation of the target object can be obtained.

[0040] For example, at least three LiDARs from the same manufacturer can be installed on a vehicle, including one forward-facing LiDAR with at least 120 lines and two side-facing LiDARs with at least 32 lines each. The forward-facing LiDAR should be positioned in a suitable location at the center of the vehicle roof, and the two side-facing LiDARs should be positioned in suitable locations outside the driver's side windows on either side of the vehicle. External parameter calibration is performed on the three LiDARs. Based on the length, width, height, pitch angle, roll angle, and yaw angle of each LiDAR relative to the vehicle's center coordinate system, the point clouds of the three LiDARs are fused into a unified coordinate system to perform target detection by integrating the point cloud data from the three LiDARs. A domain controller is required to receive the point cloud data from the three LiDARs and run the target detection model, employing a point cloud target detection algorithm to detect target objects. The neural networks involved in the object detection model can be convolutional neural networks, such as point R-CNN (Region Convolutional Neural Network), Fast R-CNN, Faster R-CNN, Fast Point R-CNN, YOLO (You Only LookOnce), VoxelNet, MV3D (Multi-View 3D Object Detection network), and AVOD (Aggregate View Object Detection network). The model is trained and validated using a large amount of historical data to ensure the accuracy of object recognition.

[0041] Step S110: Based on the laser point cloud data, determine the object recognition information within a preset range in front of the vehicle.

[0042] The objects within a preset range in front of the vehicle can include information about vehicles, pedestrians, animals, as well as rocks, mounds of earth, etc. The object identification information includes the type of object and the relative distance between the object and the vehicle.

[0043] Step S120: If a target object is determined to exist within a preset range in front of the vehicle based on the object recognition information, the vehicle's lighting status is controlled according to the relative position of the target object and the vehicle.

[0044] The vehicle lights are not controlled simply because an object is in front of it. The specific object must be identified. If it is a target, such as a person or vehicle, the lights are only controlled to avoid affecting them. Therefore, the vehicle lights are only controlled based on the relative position of the target object within a predetermined range in front of the vehicle. Lighting modes typically include high beams and low beams. Low beams have a shorter beam distance and a wider range, while high beams have a longer beam distance but a more concentrated beam and a narrower range. Therefore, low beams and high beams are suitable for different scenarios. Low beams are used when there are vehicles or pedestrians ahead to avoid obstructing their vision. High beams are used when there are no vehicles or pedestrians ahead and lighting conditions are poor to improve the driver's visibility.

[0045] In this embodiment, laser point cloud data within a preset range in front of the vehicle is first acquired, thus obtaining environmental information within that range, facilitating subsequent analysis to identify objects in front of the vehicle. Then, based on the laser point cloud data, object recognition information within the preset range in front of the vehicle is determined, enabling object recognition and providing a basis for controlling the vehicle's lights. If, based on the object recognition information, a target object is determined within the preset range in front of the vehicle, the vehicle's lights are controlled according to the relative position of the target object and the vehicle. Therefore, the process first determines whether a target object exists in front of the vehicle based on object recognition information. It doesn't mean that vehicle lights need to be controlled simply because an object appears in front of the vehicle; rather, it requires determining what the object is. If it is a target object, such as a person or vehicle, then vehicle lights need to be controlled to avoid affecting it. If a target object is determined to exist within the preset range in front of the vehicle, the vehicle's lights are further controlled based on the relative position of the target object and the vehicle, achieving intelligent light control, ensuring the vehicle's field of vision, and preventing the vehicle's lights from adversely affecting the target object.

[0046] In one embodiment, such as Figure 2 As shown, step S110 involves determining object recognition information within a preset range in front of the vehicle based on laser point cloud data. This includes steps S200-250.

[0047] Step S200: Convert the laser point cloud data into two-dimensional image data.

[0048] Laser point cloud data is three-dimensional data, including three-dimensional coordinate data. Removing coordinate data in one direction (e.g., the vertical coordinate z) converts the laser point cloud data into two-dimensional image data, or two-dimensional pseudo-image data. Since two-dimensional image data contains less information than three-dimensional data, converting laser point cloud data into two-dimensional image data reduces computational load and saves computing power. Furthermore, two-dimensional pseudo-images (2D feature images) are easier to use for deep feature extraction using 2D convolutional neural networks. The 2D convolutional neural network structure generally adopts the Feature Pyramid Network (FPN) structure. It uses multi-level feature maps to predict objects of different scales. High-level features contain high-level semantic information and a larger receptive field, suitable for detecting large objects, while shallow features contain low-level detailed semantic information and a smaller receptive field, suitable for detecting small objects. The FPN gradually fuses deep and shallow features, ultimately obtaining a feature map with rich expressive power that can extract both spatial and semantic features of the target object. The detection head of the convolutional neural network is then used to detect the two-dimensional pseudo-image to obtain the physical feature information of the target object. Specifically, the two-dimensional pseudo-image is extracted using a detector head with a few convolutional layers to obtain the physical dimensions of the target object, such as length, width, height, and orientation angle, and to determine the spatial location of the target object's center. Meanwhile, the coordinate data of each point removed from the laser point cloud data in that direction is recorded and retained for subsequent reconstruction.

[0049] For example, the coordinates of a laser point cloud can be determined using the following formula:

[0050] x = rcos(ω)sin(α+δ)

[0051] y = rcos(ω)cos(α+δ)

[0052] z = rsin(ω)

[0053] Where r is the measured distance between the target measurement point and the lidar, ω is the vertical height of the received laser beam, α is the vertical offset angle of the received laser beam, δ is the horizontal offset angle of the received laser beam, and x, y, and z are the coordinate values ​​of the polar coordinates of the target measurement point projected onto the X, Y, and Z axes of the reference coordinate system, respectively.

[0054] Step S210: Use a preprocessing neural network model to identify the two-dimensional image data and determine the object feature image data in the two-dimensional image data.

[0055] The method employs a preprocessing neural network model to identify 2D image data. This simplifies the 2D point cloud data, removing sparse point cloud data representing the environment and retaining only the feature point cloud data of objects. For example, the point clouds corresponding to the sky and the ground are relatively sparse, while the point clouds corresponding to objects are relatively dense. Therefore, simplifying the point cloud removes interfering point clouds, retaining only those useful for object detection. This effectively reduces the amount of data while preserving important point cloud information features, i.e., object feature image data.

[0056] The preprocessing neural network model can be a Point-Focus structure, using a Focus layer to rearrange image information, converting it into a 1 / 2 resolution downsampled image while preserving all image information. Then, residual connections are used to fuse color and geometric features at different scales, improving the network's feature extraction capabilities. For example, the Focus layer can obtain a value for every other pixel in an image, similar to neighbor-to-neighbor downsampling, resulting in four complementary images with no information loss. This concentrates W and H information into the channel space, expanding the input channels by four times. The stitched image becomes 12 channels compared to the original RGB three-channel mode. Finally, the new image undergoes convolution to obtain a 2x downsampled feature map without information loss, resulting in more comprehensive feature extraction.

[0057] Step S220: Reconstruct the object feature image data into three-dimensional point cloud data.

[0058] By adding the coordinate data of the direction that was removed from the original point cloud to the object feature image data, the object feature image data can be restored to three-dimensional point cloud data. In other words, by adding back the coordinate data that was previously removed from the two-dimensional point cloud in the object feature image data, it can be transformed into a three-dimensional point cloud.

[0059] Step S230: Use a detection neural network model to perform feature detection on the 3D point cloud data to determine the object feature detection information in the 3D point cloud data.

[0060] Among them, object feature detection information includes the type of object.

[0061] One approach is to perform image detection on image data from a single viewpoint using a convolutional neural network to obtain 2D detection boxes labeled with target objects. Then, post-processing segmentation or clustering is performed on the 3D point cloud data to classify and regress the point cloud. Based on the point cloud processing results and the 2D detection boxes, a 3D detection box is obtained.

[0062] The detection neural network model can be the PointNet++ backbone network model (a novel neural network that directly processes point clouds and exhibits sequence invariance of the input point cloud). The PointNet++ backbone network model can perform 3D classification and segmentation of the target point cloud data, and then output the 3D bounding box of the target object. The specific neural network model can be set according to the actual situation.

[0063] Step S240: The object feature detection information and the laser point cloud data are spliced ​​together to obtain fused feature information.

[0064] In this process, object feature detection information (such as the classification results of foreground and background points in point cloud data and the regression box on each foreground point) and laser point cloud data (such as the distance of the acquisition point from the acquisition lidar, reflection intensity and other parameters) are spliced ​​together to obtain fused feature information. This information has more dimensions and a larger amount of information, which is beneficial to improving the accuracy of subsequent recognition.

[0065] One approach is to input fused feature information into the Point-CSPNet structure. The CSPNet structure allows the model to learn more features, improving the feature fusion capability and thus obtaining more accurate feature information. This can enhance the learning ability of the detection neural network model and improve its robustness.

[0066] Step S250: Use a detection neural network model to perform feature detection again on the fused feature information to determine the object recognition information.

[0067] The fused feature information is then input into the detection neural network model for detection and recognition to obtain object recognition information, namely a 3D detection box, which can accurately label objects such as vehicles, pedestrians, animals, rocks, and mounds of earth.

[0068] In this embodiment, reducing the dimensionality of the 3D laser point cloud data before computation reduces computational load and saves computing power. Then, recognizing the 2D image data simplifies the 2D point cloud data, effectively reducing the data volume while retaining important point cloud information features, i.e., object feature image data. The object feature image data is then restored to 3D point cloud data, improving the reliability and effectiveness of subsequent detection results. Feature stitching is then performed using object feature detection information and laser point cloud data to obtain fused feature information, enriching the feature information and improving the accuracy of subsequent recognition. Using the CSPNet structure allows the model to learn more features, improving feature fusion capabilities and obtaining more accurate feature information. This enhances the learning ability of the detection neural network model, thereby improving robustness. Finally, the detection neural network model performs feature detection again on the fused feature information to determine object recognition information, thus fully extracting effective information from the point cloud data and effectively integrating features and gradient changes during network computation, improving detection accuracy and accurately identifying the type of object and its relative position to the vehicle.

[0069] In one embodiment, such as Figure 3 As shown, step S230 involves using a detection neural network model to perform feature detection on the 3D point cloud data, determining the object feature detection information in the 3D point cloud data. This includes steps S300-320.

[0070] Step S300: Use a detection neural network model to perform feature detection on the 3D point cloud data and generate a 3D bounding box.

[0071] The method involves performing image detection on image data from a single viewpoint using a convolutional neural network to obtain 2D detection boxes labeled with target objects. Then, post-processing segmentation or clustering is performed on the 3D point cloud data, and the point cloud is classified and regressed. Based on the point cloud processing results and the 2D detection boxes, a 3D detection box is obtained.

[0072] Step S310: Adjust the three-dimensional border so that it accurately frames the object features.

[0073] Among these features, the range selected by the 3D border can be fine-tuned so that the 3D border accurately frames the feature vector of the object.

[0074] Step S320: Detect the adjusted 3D bounding box to determine the object feature detection information.

[0075] In this process, by detecting the 3D bounding box and using a detection neural network model to identify it, the feature detection information of the object with the 3D bounding box can be determined.

[0076] This process involves using a detection neural network model to extract features from 3D bounding boxes, obtaining feature vectors from point cloud data, and then performing classification and regression to calculate the foreground and background point classification results, as well as the regression box on each foreground point. To further improve the accuracy of the 3D detection boxes, the obtained feature vectors are concatenated with parameters such as the foreground point mask, the distance from the point to the LiDAR sensor, and the reflection intensity, and this concatenation is used as a new local feature input into the Point-CSPNet structure. The result is then fused with the previously obtained feature vectors and input into the PointNet++ structure for computation.

[0077] One approach is to incorporate 3D bounding boxes into the Point-CSPNet structure to improve the learning ability of the detection neural network model, thereby enhancing its robustness.

[0078] In this embodiment, by employing a detection neural network model to perform feature detection on the 3D point cloud data, object feature detection information in the 3D point cloud data can be determined. The Point-CSPNet structure can be used to improve the learning ability of the detection neural network model, thereby improving its robustness.

[0079] In one embodiment, such as Figure 4 As shown, the target objects include environmental vehicles and pedestrians. Step S120 controls the vehicle's lighting status based on the relative position of the target objects and the vehicle, including steps S400-430.

[0080] In step S400, when it is determined that the distance between the ambient vehicle and the vehicle in the first direction is less than a first threshold, the vehicle's headlights are controlled to be in low beam mode.

[0081] The first direction is the direction in which the vehicle is moving. The first threshold can be 80 meters.

[0082] Step S410: When it is determined that the distance between the ambient vehicle and the vehicle in the first direction is greater than or equal to the first threshold, the vehicle's headlights are controlled to be high beam.

[0083] Specifically, when a vehicle is detected traveling ahead, if the longitudinal distance between the detected vehicle and the vehicle is less than 80m, the vehicle's headlights will be switched to low beam mode. If the longitudinal distance between the detected vehicle and the vehicle is greater than or equal to 80m, or if there is no vehicle traveling in the oncoming lane ahead, the vehicle's headlights will be switched to high beam mode.

[0084] Step S420: When it is determined that the distance between the pedestrian and the vehicle in the first direction is less than the second threshold, the vehicle's headlights are controlled to be low beam.

[0085] The second threshold is less than the first threshold. The first threshold can be 50 meters.

[0086] Step S430: When it is determined that the distance between the pedestrian and the vehicle in the first direction is greater than or equal to the second threshold, the vehicle's headlights are controlled to be high beam.

[0087] Specifically, when a pedestrian is detected ahead, if the longitudinal distance between the pedestrian and the vehicle is less than 50 meters, the vehicle's headlights will be switched to low beam mode. If the longitudinal distance between the pedestrian and the vehicle is greater than 50 meters, or if there is no pedestrian in the opposite direction, the vehicle's headlights will be switched to high beam mode.

[0088] In this embodiment, based on the relative positions of the vehicle and pedestrians in the environment, the vehicle is controlled to adopt appropriate lighting conditions, realizing intelligent lighting control, ensuring the vehicle's field of vision, and avoiding adverse effects of the vehicle's lights on the target object.

[0089] In one embodiment, the vehicle lighting control method further includes: when it is determined, based on object recognition information, that there is an object within a preset range in front of the vehicle, determining whether to issue an alarm based on the relative position of the object and the vehicle.

[0090] The objects include obstacles and target objects. For example, obstacles can be animals, rocks, mounds of earth, trees, etc., while target objects can be environmental objects such as vehicles or pedestrians.

[0091] Specifically, such as Figure 5 As shown, the process of determining whether to issue an alarm based on the relative position of the object and the vehicle includes steps S500-510.

[0092] Step S500: Determine the first distance between the object and the vehicle in the first direction and the second distance in the second direction.

[0093] The first direction is the direction in which the vehicle moves forward, and the second direction is perpendicular to the first direction.

[0094] Step S510: If the first distance is less than the third threshold or the second distance is less than the fourth threshold, issue an audible and visual alarm.

[0095] The third threshold can be 50 meters, and the fourth threshold can be 30 centimeters. That is, when the detected object is less than 30 centimeters away from the vehicle laterally, or less than 50 meters away from the vehicle longitudinally, it is determined that the object has the potential to scratch or collide with the vehicle. Therefore, an audible and visual alarm is issued in the driver's cabin, and the location of the object is displayed on the central control screen.

[0096] In this embodiment, based on object recognition and detection, it is used to predict whether an object may collide with the vehicle, and then to issue an audible and visual alarm to alert the driver when danger occurs, thereby improving vehicle safety.

[0097] It should be understood that, although Figures 1-5 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figures 1-5 At least some of the steps in the process may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but may be executed at different times. The execution order of these steps or stages is not necessarily sequential, but may be executed in turn or alternately with other steps or at least some of the steps or stages in other steps.

[0098] In one embodiment, such as Figure 6 As shown, a vehicle lighting control device is provided, including: a point cloud acquisition module 601, an object recognition module 602, and a lighting control module 603, wherein:

[0099] The point cloud acquisition module 601 is used to acquire laser point cloud data within a preset range in front of the vehicle.

[0100] The object recognition module 602 is used to determine object recognition information within a preset range in front of the vehicle based on laser point cloud data.

[0101] The lighting control module 603 is used to control the lighting status of the vehicle based on the relative position of the target object and the vehicle when it is determined that there is a target object within a preset range in front of the vehicle based on object recognition information.

[0102] Specific limitations regarding the vehicle lighting control device can be found in the above-described limitations of the vehicle lighting control method, and will not be repeated here. Each module in the aforementioned vehicle lighting control device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. It should be noted that the module division in this embodiment is illustrative and only represents a logical functional division; in actual implementation, other division methods may be used.

[0103] In one embodiment, a computer device is provided, the internal structure of which can be shown in the following diagram. Figure 7As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a vehicle lighting control method.

[0104] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0105] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0106] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0107] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0108] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0109] In the description of this specification, references to terms such as "some embodiments," "other embodiments," and "ideal embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative descriptions of the above terms do not necessarily refer to the same embodiments or examples.

[0110] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0111] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A vehicle lighting control method, characterized in that, include: Acquire laser point cloud data within a preset range in front of the vehicle; Based on the laser point cloud data, object recognition information within a preset range in front of the vehicle is determined; If, based on the object recognition information, it is determined that there is a target object within a preset range in front of the vehicle, the vehicle's lighting status is controlled according to the relative position of the target object and the vehicle. The step of determining object recognition information within a preset range in front of the vehicle based on the laser point cloud data includes: converting the laser point cloud data into two-dimensional image data; using a preprocessing neural network model to identify the two-dimensional image data and determine object feature image data in the two-dimensional image data; restoring the object feature image data into three-dimensional point cloud data; using a detection neural network model to perform feature detection on the three-dimensional point cloud data and determine object feature detection information in the three-dimensional point cloud data, wherein the object feature detection information includes the type of object; performing feature concatenation between the object feature detection information and the laser point cloud data to obtain fused feature information; and using the detection neural network model to perform feature detection again on the fused feature information to determine the object recognition information. The step of using a detection neural network model to perform feature detection on the 3D point cloud data and determine object feature detection information in the 3D point cloud data includes: using a detection neural network model to perform feature detection on the 3D point cloud data and generating a 3D bounding box; adjusting the 3D bounding box to accurately frame the object features; and detecting the adjusted 3D bounding box to determine the object feature detection information. The laser point cloud data includes three-dimensional coordinate data; converting the laser point cloud data into two-dimensional image data includes removing coordinate data in one direction from the three-dimensional coordinate data; restoring the object feature image data into three-dimensional point cloud data includes supplementing the object feature image data with the coordinate data in the direction that was removed from the originally recorded point cloud.

2. The vehicle lighting control method according to claim 1, characterized in that, The target objects include environmental vehicles and pedestrians. Controlling the vehicle's lighting status based on the relative position of the target objects and the vehicle includes: When it is determined that the distance between the environmental vehicle and the autonomous vehicle in a first direction is less than a first threshold, the lighting state of the autonomous vehicle is controlled to be low beam, wherein the first direction is the forward direction of the autonomous vehicle; When it is determined that the distance between the environmental vehicle and the autonomous vehicle in the first direction is greater than or equal to the first threshold, the autonomous vehicle's headlights are controlled to be in high beam mode. When it is determined that the distance between the pedestrian and the vehicle in the first direction is less than a second threshold, the vehicle's headlights are controlled to be low beam, wherein the second threshold is less than the first threshold; When it is determined that the distance between the pedestrian and the vehicle in the first direction is greater than or equal to the second threshold, the vehicle's headlights are controlled to be high beam.

3. The vehicle lighting control method according to claim 1, characterized in that, The method further includes: If an object is determined to exist within a preset range in front of the vehicle based on the object recognition information, an alarm is issued based on the relative position of the object and the vehicle. The object includes obstacles and the target object.

4. The vehicle lighting control method according to claim 3, characterized in that, The step of determining whether to issue an alarm based on the relative position of the object and the vehicle includes: Determine a first distance between the object and the vehicle in a first direction and a second distance in a second direction, wherein the first direction is the forward direction of the vehicle and the second direction is perpendicular to the first direction; An audible and visual alarm is triggered when the first distance is less than the third threshold, or when the second distance is less than the fourth threshold.

5. The vehicle lighting control method according to claim 1, characterized in that, The vehicle is equipped with a lidar system; the formula for the coordinates of the lidar point cloud in the lidar point cloud data is: x=rcos(ω)sin(α+δ), y=rcos(ω)cos(α+δ), z=rsin(ω); Where r is the measured distance between the target measurement point and the lidar, ω is the vertical height of the received laser beam, α is the vertical offset angle of the received laser beam, δ is the horizontal offset angle of the received laser beam, and x, y, and z are the coordinate values ​​of the polar coordinates of the target measurement point projected onto the X, Y, and Z axes of the reference coordinate system, respectively.

6. A vehicle lighting control device, characterized in that, include: The point cloud acquisition module is used to acquire laser point cloud data within a preset range in front of the vehicle; An object recognition module is used to determine object recognition information within a preset range in front of the vehicle based on the laser point cloud data. The lighting control module is used to control the lighting status of the vehicle based on the relative position of the target object and the vehicle when it is determined that there is a target object within a preset range in front of the vehicle based on the object recognition information. The object recognition module is further configured to convert the laser point cloud data into two-dimensional image data; use a preprocessing neural network model to recognize the two-dimensional image data and determine the object feature image data in the two-dimensional image data; restore the object feature image data into three-dimensional point cloud data; use a detection neural network model to perform feature detection on the three-dimensional point cloud data and determine the object feature detection information in the three-dimensional point cloud data, wherein the object feature detection information includes the type of object; perform feature concatenation between the object feature detection information and the laser point cloud data to obtain fused feature information; and use the detection neural network model to perform feature detection again on the fused feature information to determine the object recognition information. The object recognition module is also used to perform feature detection on the three-dimensional point cloud data using a detection neural network model to generate a three-dimensional bounding box; and to adjust the three-dimensional bounding box so that the three-dimensional bounding box accurately frames the object features; The adjusted 3D bounding box is detected to determine the object feature detection information; The laser point cloud data includes three-dimensional coordinate data; converting the laser point cloud data into two-dimensional image data includes removing coordinate data in one direction from the three-dimensional coordinate data; restoring the object feature image data into three-dimensional point cloud data includes supplementing the object feature image data with the coordinate data in the direction that was removed from the originally recorded point cloud.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images