Vehicle visual mode switching method, device, equipment and computer program product

Abstract

The invention discloses a vehicle visual mode switching method, device and equipment and a computer program product. The method comprises the steps that the real-time vehicle distance, the threat fuzzy coefficient, the real-time vehicle speed, the traffic jam condition and the dynamic balance level of a current vehicle and the driving habit style of a driver are obtained; inputting the real-time vehicle distance, the threat fuzzy coefficient, the real-time vehicle speed, the traffic jam condition, the dynamic balance level and the driving habit style into a pre-trained visual mode decision model to obtain a target visual mode; switching the visual mode of the current vehicle to a target visual mode, and controlling a corresponding target visual sensor to operate; wherein the target visual mode is one of a vector-level visual mode, a contour-level visual mode and a high-precision visual mode. On the premise of ensuring the driving safety of the vehicle, the computing power resource consumption of the vehicle is optimized, so that the performance of vehicle hardware is improved, the service life of the vehicle hardware is prolonged, and the method can be applied to the technical field of intelligent driving.

Description

Technical Field

[0001] This invention relates to the field of intelligent driving technology, and in particular to a vehicle vision mode switching method, device, equipment, and computer program product. Background Technology

[0002] In the current technology, in order to pursue the intelligent driving experience, vehicles are generally equipped with high-precision maps, high-specification visual sensors (cameras, radar) and high-performance chips. However, the intelligent driving data model of the intelligent driving system will be continuously upgraded via OTA over time, but the vehicle hardware will only gradually age and cannot be automatically expanded or upgraded. Therefore, the long-term high-precision data processing of vehicle hardware will lead to hardware failure and thus create safety risks.

[0003] The above problems urgently need to be addressed. Summary of the Invention

[0004] The purpose of this invention is to at least partially solve one of the technical problems existing in the prior art.

[0005] Therefore, one objective of this invention is to provide a vehicle vision mode switching method. This method can determine a target vision mode that meets safety and computing power optimization requirements based on the real-time driving scenario of the vehicle, and control the corresponding vision sensors to operate according to the target vision mode. Under the premise of ensuring vehicle driving safety, this method optimizes the consumption of vehicle computing power resources, thereby improving the performance and service life of vehicle hardware.

[0006] Another objective of this invention is to provide a vehicle vision mode switching device.

[0007] To achieve the above-mentioned technical objectives, the technical solutions adopted in the embodiments of the present invention include: On one hand, embodiments of the present invention provide a method for switching vehicle vision modes, including the following steps: Obtain the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style; The real-time vehicle distance, the threat ambiguity coefficient, the real-time vehicle speed, the traffic congestion status, the dynamic balance level, and the driving habit style are input into a pre-trained visual pattern decision model to obtain the target visual pattern. Switch the current vehicle's vision mode to the target vision mode and control the corresponding target vision sensor to operate; The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

[0008] Furthermore, in one embodiment of the present invention, the acquisition of the current vehicle's real-time distance, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driver's driving habits specifically includes: The real-time vehicle distance and the threat ambiguity coefficient are determined based on radar data and camera data; The real-time vehicle speed is determined based on wheel speed sensor data and GPS positioning data; The traffic congestion situation is determined based on real-time traffic conditions from the navigation system and V2X data. The dynamic balance level is determined based on the vehicle body's IMU (Inertial Measurement Unit). The driving habit style is determined based on the driver's historical driving behavior data.

[0009] Furthermore, in one embodiment of the present invention, the threat ambiguity coefficient is determined through the following steps: The real-time risk value of the current vehicle is determined based on the radar data and the camera data; The safety threshold is determined based on the current vehicle model and the current driving scenario; Determine the environmental correction factor based on the current vehicle's environmental information; The initial threat coefficient is obtained by comparing the real-time risk value with the security threshold, and the initial threat coefficient is then corrected by the environmental correction coefficient to obtain the threat ambiguity coefficient.

[0010] Furthermore, in one embodiment of the present invention, the visual pattern decision model is obtained through the following steps: The sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habits and styles of the test drivers are obtained. Training samples are constructed based on the sample vehicle distance, the sample threat ambiguity coefficient, the sample vehicle speed, the sample traffic congestion status, the sample dynamic balance level, and the sample driving habit style. The corresponding visual pattern labels are determined by manual annotation to obtain the training dataset. The training samples are input into a pre-constructed deep neural network to obtain visual pattern prediction values; The loss value is determined based on the visual pattern prediction value and the visual pattern label; The parameters of the deep neural network are updated based on the loss value to obtain the trained visual pattern decision model.

[0011] Furthermore, in one embodiment of the present invention, the visual pattern decision model is obtained through the following steps: The sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habits and styles of the test drivers are obtained. Training samples are constructed based on the sample vehicle distance, the sample threat ambiguity coefficient, the sample vehicle speed, the sample traffic congestion status, the sample dynamic balance level, and the sample driving habit style. The corresponding visual pattern labels are determined by manual annotation to obtain the training dataset. Initialize the gradient boosting decision tree; The training samples are input into the gradient boosting decision tree to obtain visual pattern prediction values. The loss value is determined based on the visual pattern prediction value and the visual pattern label, and the negative gradient of the loss value is determined. The negative gradient is used as the prediction target to train a first decision tree based on the training samples. The gradient boosting decision tree is updated based on the first decision tree, and the process returns to the step of inputting the training samples into the gradient boosting decision tree. When the number of decision trees in the gradient boosting decision tree reaches a preset first threshold, or the loss value is less than or equal to a preset second threshold, training stops, and the trained visual pattern decision model is obtained.

[0012] Furthermore, in one embodiment of the present invention, controlling the operation of the corresponding target visual sensor specifically includes: When the target visual mode is a vector-level visual mode, a conventional camera and a millimeter-wave radar are identified as the target visual sensors, and the conventional camera and the millimeter-wave radar are controlled to operate. When the target visual mode is a contour-level visual mode, the high-definition camera and the LiDAR are identified as the target visual sensors, and the high-definition camera and the LiDAR are controlled to operate according to a preset low-frequency sampling mode; When the target vision mode is a high-precision vision mode, the high-definition camera, the lidar, and the infrared camera are identified as the target vision sensors, and the high-definition camera, the lidar, and the infrared camera are controlled to operate according to a preset high-frequency sampling mode.

[0013] Furthermore, in one embodiment of the present invention, when the current vehicle is in the vector-level vision mode, the visual data collected by the target vision sensor is used to perform lightweight feature extraction on the target object, thereby obtaining the motion velocity vector and coarse position information of the target object; When the current vehicle is in the contour-level vision mode, the visual data collected by the target vision sensor is used to extract structured features of the target object, thereby obtaining the edge contour, semantic category and motion trajectory information of the target object; When the current vehicle is in the high-precision vision mode, the visual data collected by the target vision sensor is used to perform full-element perception of the target object, thereby obtaining the three-dimensional structure, detailed semantics and high-precision pose information of the target object.

[0014] On the other hand, embodiments of the present invention provide a vehicle vision mode switching device, comprising: The data acquisition module is used to acquire the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style. The visual pattern decision module is used to input the real-time vehicle distance, the threat ambiguity coefficient, the real-time vehicle speed, the traffic congestion status, the dynamic balance level, and the driving habit style into the pre-trained visual pattern decision model to obtain the target visual pattern. The vision mode switching module is used to switch the current vehicle's vision mode to the target vision mode and control the operation of the corresponding target vision sensor. The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

[0015] On the other hand, embodiments of the present invention provide an electronic device, including: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the vehicle vision mode switching method described above.

[0016] On the other hand, embodiments of the present invention also provide a computer-readable storage medium storing a processor-executable computer program that, when executed by a processor, implements the above-described vehicle vision mode switching method.

[0017] On the other hand, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the above-described vehicle vision mode switching method.

[0018] The advantages and beneficial effects of the present invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention: This invention acquires the real-time distance to other vehicles, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits of the current vehicle. These data are then input into a pre-trained visual pattern decision model to obtain a target visual pattern. The current vehicle's visual pattern is switched to the target visual pattern, and the corresponding target visual sensor is controlled to operate. The target visual pattern can be one of a vector-level visual pattern, a contour-level visual pattern, or a high-precision visual pattern. This invention can determine a target visual pattern that meets both safety and computational optimization requirements based on the vehicle's real-time driving scenario, and control the corresponding visual sensors to operate according to the target visual pattern. While ensuring vehicle driving safety, it optimizes the consumption of vehicle computing resources, thereby improving the performance and lifespan of the vehicle hardware. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the embodiments of the present invention are described below. It should be understood that the drawings described below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 A flowchart illustrating the steps of a vehicle vision mode switching method provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a vehicle vision mode switching device provided in an embodiment of the present invention; Figure 3 This is a structural block diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of this invention; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this invention as detailed in the appended claims.

[0022] 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 invention pertains. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to limit the invention.

[0023] The vehicle vision mode switching method provided in this invention can be applied to a terminal, a server, or software running on a terminal or server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or in-vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application that implements the vehicle vision mode switching method, but is not limited to the above forms.

[0024] This invention can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This invention can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This invention can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0025] It should be noted that in various specific embodiments of the present invention, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of the present invention require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to a confirmation page. Only after obtaining the user's separate permission or consent is the necessary user-related data for the normal operation of the embodiments of the present invention acquired.

[0026] Reference Figure 1 This invention provides a method for switching vehicle vision modes, specifically including the following steps: S101. Obtain the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style. S102. Input the real-time vehicle distance, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driving habit style into the pre-trained visual pattern decision model to obtain the target visual pattern. S103. Switch the current vehicle's vision mode to the target vision mode and control the corresponding target vision sensor to operate; The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

[0027] The embodiments of the present invention can determine the target visual mode that meets the requirements of safety and computing power optimization based on the real-time driving scenario of the vehicle, and control the corresponding visual sensors to operate according to the target visual mode. Under the premise of ensuring vehicle driving safety, the consumption of vehicle computing power resources is optimized, thereby improving the performance and service life of vehicle hardware.

[0028] As an optional implementation, the real-time distance to the current vehicle, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driver's driving habits and style are obtained, specifically including: S201. Determine the real-time vehicle distance and threat ambiguity coefficient based on radar and camera data; S202. Determine the real-time vehicle speed based on wheel speed sensor data and GPS positioning data; S203. Determine traffic congestion status based on real-time traffic conditions from the navigation system and V2X data. S204. Determine the dynamic balance level based on the vehicle body IMU inertial measurement unit; S205. Determine driving habits and styles based on the driver's historical driving behavior data.

[0029] Specifically, the system calculates the real-time distance and threat ambiguity coefficient between the current vehicle and the vehicle ahead based on radar and camera data; it calculates the instantaneous vehicle speed based on wheel speed sensor data and tire circumference, and the average vehicle speed based on the rate of change of position using GPS positioning data, then fuses these data using a Kalman filter algorithm to obtain the real-time vehicle speed; it classifies traffic congestion into levels 0, 1, 2, and 3 based on navigation coefficients, real-time road conditions, and V2X data, corresponding to smooth traffic, slow traffic, congestion, and severe congestion, respectively; it calculates the vehicle's dynamic balance level by measuring the vehicle's 3-axis acceleration and 3-axis angular velocity using the vehicle's IMU (Inertial Measurement Unit), with a high dynamic balance level indicating better grip and stability, better handling at high speeds or in emergencies, and a reduced risk of skidding and loss of control, while a low dynamic balance level means the vehicle may be more prone to instability at high speeds or in emergencies, increasing the risk of skidding and loss of control; and it analyzes the driver's historical driving behavior data, including accelerator pedal depth, braking frequency, and steering angle, to classify the driver's driving style as aggressive, steady, or cautious.

[0030] As an optional implementation, the threat ambiguity coefficient is determined through the following steps: S301. Determine the real-time risk value of the current vehicle based on radar data and camera data; S302. Determine the safety threshold based on the current vehicle model and current driving scenario; S303. Determine the environmental correction factor based on the current vehicle environmental information; S304. The initial threat coefficient is obtained based on the ratio of the real-time risk value to the safety threshold, and the initial threat coefficient is corrected based on the environmental correction coefficient to obtain the threat fuzziness coefficient.

[0031] Specifically, the threat fuzziness coefficient is a comprehensive indicator used in intelligent driving systems to quantify the degree of danger in the surrounding environment. Its value typically ranges from 0 to 1 (0 represents no risk, and 1 represents extremely high risk). By fusing data from multiple sensors (cameras, radar, etc.), it performs a fuzzy assessment of potential threats such as vehicles, pedestrians, and obstacles ahead, ultimately providing a basis for visual mode switching, braking decisions, and other functions.

[0032] The instantaneous hazard score (such as the confidence level of millimeter-wave radar detecting obstacles ahead) is calculated using radar and camera data to obtain a real-time risk value. A safety threshold is determined based on the vehicle model and pre-set safety benchmarks for the scenario (such as a higher threshold for highway scenarios than for urban roads). An environmental correction coefficient is determined based on the current vehicle's environmental information (such as amplification of the coefficient in rainy / nighttime scenarios). An initial threat coefficient is obtained based on the ratio of the real-time risk value to the safety threshold. The initial threat coefficient is then corrected based on the environmental correction coefficient to obtain a threat ambiguity coefficient. The calculation formula is: Threat ambiguity coefficient = (safety threshold / real-time risk value) × environmental correction coefficient.

[0033] The visual pattern decision model of this invention is used to take into account real-time vehicle distance, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driving habit style, and outputs the target visual pattern. It can be trained by a neural network model and a gradient boosting decision tree model. The two training methods are described below.

[0034] As an optional implementation, the visual pattern decision model is obtained through the following steps: S401. Obtain the sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habit style of the test driver. S402. Construct training samples based on sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habit style. Determine the corresponding visual pattern labels through manual annotation to obtain the training dataset. S403. Input the training samples into a pre-built deep neural network to obtain visual pattern prediction values; S404. Determine the loss value based on the visual pattern prediction value and the visual pattern label; S405. Update the parameters of the deep neural network based on the loss value to obtain the trained visual pattern decision model.

[0035] Specifically, the process involves acquiring sample vehicle distances, threat ambiguity coefficients, vehicle speeds, traffic congestion conditions, dynamic balance levels, and driving habits of test drivers. Training samples are then constructed based on these data, and corresponding visual pattern labels are manually assigned. After collecting a sufficient number of training samples, a training dataset is obtained. These training samples are then input into a pre-built deep neural network to obtain visual pattern predictions. A loss value is determined based on the visual pattern predictions and labels, and the parameters of the deep neural network are updated according to the loss value, completing one iteration of training. Training stops when the number of iterations reaches a preset threshold or the loss value is less than or equal to the preset threshold, resulting in a well-trained visual pattern decision model.

[0036] As an optional implementation, the visual pattern decision model is obtained through the following steps: S501. Obtain the sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habit style of the test driver. S502. Construct training samples based on sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habit style. Determine the corresponding visual pattern labels through manual annotation to obtain the training dataset. S503, Initialize the gradient boosting decision tree; S504. Input the training samples into the gradient boosting decision tree to obtain the visual pattern prediction value; S505. Determine the loss value based on the visual pattern prediction value and the visual pattern label, and determine the negative gradient of the loss value; S506. Using the negative gradient as the prediction target, train the first decision tree based on the training samples, update the gradient boosting decision tree based on the first decision tree, and return to the step of inputting the training samples into the gradient boosting decision tree. S507. When the number of decision trees in the gradient boosting decision tree reaches the preset first threshold, or the loss value is less than or equal to the preset second threshold, training is stopped, and the trained visual pattern decision model is obtained.

[0037] Specifically, Gradient Boosting Decision Tree (GBDT) is a serial ensemble learning algorithm that iteratively trains multiple decision trees to gradually correct prediction errors, ultimately forming a strong classifier. The process involves acquiring sample vehicle distances, threat ambiguity coefficients, vehicle speeds, traffic congestion conditions, dynamic balance levels, and driver driving habits from test vehicles. Training samples are constructed based on these data, and corresponding visual pattern labels are determined through manual annotation. After collecting a sufficient number of training samples, a training dataset is obtained, ensuring coverage of all extreme scenarios. The gradient boosting decision tree is initialized; the initial model can use a simple constant predictor, aiming to minimize the initial loss. The training samples are then input into the gradient boosting decision tree to obtain visual pattern predictions. Finally, the visual pattern predictions and labels are compared. First, determine the loss value and its negative gradient. The negative gradient is the opposite of the derivative of the loss value with respect to the visual pattern prediction value, representing the direction in which the model needs to be corrected. Then, use the negative gradient as the prediction target to train the first decision tree based on the training samples. To prevent overfitting, a learning rate is assigned to the first decision tree to control its contribution to the final prediction. The prediction results of the first decision tree are accumulated into the current gradient boosting decision tree, thus updating the gradient boosting decision tree based on the first decision tree. Then, the process returns to the step of inputting training samples into the gradient boosting decision tree for iterative training. When the number of decision trees in the gradient boosting decision tree reaches a preset first threshold, or the loss value is less than or equal to a preset second threshold, training stops, and the trained visual pattern decision model is obtained.

[0038] As a further optional implementation, controlling the operation of the corresponding target vision sensor specifically includes: S601. When the target vision mode is vector level vision mode, determine the conventional camera and millimeter-wave radar as the target vision sensors, and control the operation of the conventional camera and millimeter-wave radar. S602. When the target visual mode is a contour-level visual mode, determine the high-definition camera and the LiDAR as the target visual sensors, and control the high-definition camera and the LiDAR to operate according to the preset low-frequency sampling mode. S603. When the target vision mode is high-precision vision mode, determine the high-definition camera, LiDAR and infrared camera as target vision sensors, and control the high-definition camera, LiDAR and infrared camera to operate according to the preset high-frequency sampling mode.

[0039] Specifically, vector-level vision mode quickly identifies the approximate shape and trajectory of obstacles based on simple geometric shape and orientation information. The hardware requirements are conventional cameras and millimeter-wave radar, which can be applied to low-speed, low-threat environments such as parking lots and residential areas. Contour-level vision mode identifies the contour information of obstacles, providing more detailed shape and position data than vector vision. The hardware requirements are high-definition cameras and LiDAR (partial functionality), which can be applied to medium-speed, medium-threat environments such as urban roads and lightly congested sections. High-precision vision mode provides high-precision environmental maps, including detailed information on obstacles, road markings, pedestrians, etc. The hardware requirements are high-definition cameras, LiDAR (full functionality), and specific sensors (such as infrared cameras), which can be applied to high-speed, high-threat environments such as highways, complex traffic intersections, and severe weather conditions.

[0040] As an optional implementation, when the vehicle is in vector-level vision mode, the visual data collected by the target vision sensor is used to perform lightweight feature extraction on the target object, thereby obtaining the target object's motion velocity vector and coarse position information. When the vehicle is in contour-level vision mode, the visual data collected by the target vision sensor is used to extract the structured features of the target object, thereby obtaining the edge contour, semantic category and motion trajectory information of the target object; When the vehicle is in high-precision vision mode, the visual data collected by the target vision sensor is used to perceive all elements of the target object, thereby obtaining the three-dimensional structure, detailed semantics, and high-precision pose information of the target object.

[0041] Specifically, visual data in Vector-level Mode is used for lightweight feature extraction, retaining only the target's motion vectors and coarse position information. The input resolution is low (e.g., 640×480), the frame rate is 30fps, and the algorithm employs optical flow and target tracking algorithms (e.g., KLT feature point tracking). The output is the target's bounding box (x, y, w, h) and motion velocity vector (v). x , v yThe first method, **Contour-level Visual Mode**, is applicable to highway cruising scenarios, requiring only the monitoring of the relative displacement of vehicles ahead. The second method uses visual data in Contour-level Mode for structured feature extraction, adding edge contours and semantic category information of the target. It has a medium input resolution (1080P) and a frame rate of 60fps. The algorithm employs edge detection (Canny) and lightweight semantic segmentation (such as MobileNet-SSD), outputting a target contour mask and category labels (such as "car," "pedestrian," and "lane line"). It can be applied to urban expressways where vehicle and pedestrian distinction is required. The third method uses visual data in High-precision Mode for full-element perception, outputting the target's 3D structure and detailed semantic information. It has a high input resolution (4K) and a frame rate of 120fps. The algorithm employs stereo matching, instance segmentation (such as MaskR-CNN), and depth estimation, outputting the target's 3D coordinates (x, y, z), a pixel-level semantic mask, and surface normal vectors. It can be applied to complex intersections where pedestrian gestures and traffic light colors need to be recognized.

[0042] The method steps of the embodiments of the present invention have been described above. It can be understood that the embodiments of the present invention can determine the target visual mode that meets the requirements of safety and computing power optimization based on the real-time driving scenario of the vehicle, and control the corresponding visual sensors to operate according to the target visual mode. Under the premise of ensuring vehicle driving safety, the consumption of vehicle computing power resources is optimized, thereby improving the performance and service life of vehicle hardware.

[0043] Reference Figure 2 This invention provides a vehicle vision mode switching device, comprising: The data acquisition module is used to acquire the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style. The visual pattern decision module is used to input real-time vehicle distance, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driving habit style into the pre-trained visual pattern decision model to obtain the target visual pattern. The vision mode switching module is used to switch the current vehicle's vision mode to the target vision mode and control the operation of the corresponding target vision sensor. The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

[0044] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0045] Reference Figure 3 This invention provides an electronic device, comprising: At least one processor; At least one memory for storing at least one program; When the above-mentioned at least one program is executed by the above-mentioned at least one processor, the above-mentioned at least one processor implements the above-mentioned vehicle vision mode switching method.

[0046] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0047] This invention also provides a computer-readable storage medium storing a processor-executable computer program that, when executed by a processor, implements the above-described vehicle vision mode switching method.

[0048] This invention provides a computer-readable storage medium that can execute a vehicle vision mode switching method provided in the method embodiment of this invention. It can execute any combination of the implementation steps of the method embodiment and has the corresponding functions and beneficial effects of the method.

[0049] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described vehicle vision mode switching method.

[0050] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0051] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0052] The embodiments described in this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.

[0053] The terms "first," "second," "third," "fourth," etc. (if present) in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0054] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the aforementioned blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0055] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the aforementioned functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0056] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0057] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0058] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the aforementioned program can be printed, because the aforementioned program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0059] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0060] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0061] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0062] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A method for switching vehicle vision modes, characterized in that, Includes the following steps: Obtain the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style; The real-time vehicle distance, the threat ambiguity coefficient, the real-time vehicle speed, the traffic congestion status, the dynamic balance level, and the driving habit style are input into a pre-trained visual pattern decision model to obtain the target visual pattern. Switch the current vehicle's vision mode to the target vision mode and control the corresponding target vision sensor to operate; The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

2. The vehicle vision mode switching method according to claim 1, characterized in that, The acquisition of real-time vehicle distance, threat ambiguity coefficient, real-time vehicle speed, traffic congestion status, dynamic balance level, and driver's driving habits and style specifically includes: The real-time vehicle distance and the threat ambiguity coefficient are determined based on radar data and camera data; The real-time vehicle speed is determined based on wheel speed sensor data and GPS positioning data; The traffic congestion situation is determined based on real-time traffic conditions from the navigation system and V2X data. The dynamic balance level is determined based on the vehicle body's IMU (Inertial Measurement Unit). The driving habit style is determined based on the driver's historical driving behavior data.

3. The vehicle vision mode switching method according to claim 2, characterized in that, The threat ambiguity coefficient is determined through the following steps: The real-time risk value of the current vehicle is determined based on the radar data and the camera data; The safety threshold is determined based on the current vehicle model and the current driving scenario; Determine the environmental correction factor based on the current vehicle's environmental information; The initial threat coefficient is obtained by comparing the real-time risk value with the security threshold, and the initial threat coefficient is then corrected by the environmental correction coefficient to obtain the threat ambiguity coefficient.

4. The vehicle vision mode switching method according to claim 1, characterized in that, The visual pattern decision model is obtained through the following steps: The sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habits and styles of the test drivers are obtained. Training samples are constructed based on the sample vehicle distance, the sample threat ambiguity coefficient, the sample vehicle speed, the sample traffic congestion status, the sample dynamic balance level, and the sample driving habit style. The corresponding visual pattern labels are determined by manual annotation to obtain the training dataset. The training samples are input into a pre-constructed deep neural network to obtain visual pattern prediction values; The loss value is determined based on the visual pattern prediction value and the visual pattern label; The parameters of the deep neural network are updated based on the loss value to obtain the trained visual pattern decision model.

5. The vehicle vision mode switching method according to claim 1, characterized in that, The visual pattern decision model is obtained through the following steps: The sample vehicle distance, sample threat ambiguity coefficient, sample vehicle speed, sample traffic congestion status, sample dynamic balance level, and sample driving habits and styles of the test drivers are obtained. Training samples are constructed based on the sample vehicle distance, the sample threat ambiguity coefficient, the sample vehicle speed, the sample traffic congestion status, the sample dynamic balance level, and the sample driving habit style. The corresponding visual pattern labels are determined by manual annotation to obtain the training dataset. Initialize the gradient boosting decision tree; The training samples are input into the gradient boosting decision tree to obtain visual pattern prediction values. The loss value is determined based on the visual pattern prediction value and the visual pattern label, and the negative gradient of the loss value is determined. The negative gradient is used as the prediction target to train a first decision tree based on the training samples. The gradient boosting decision tree is updated based on the first decision tree, and the process returns to the step of inputting the training samples into the gradient boosting decision tree. When the number of decision trees in the gradient boosting decision tree reaches a preset first threshold, or the loss value is less than or equal to a preset second threshold, training stops, and the trained visual pattern decision model is obtained.

6. The vehicle vision mode switching method according to claim 1, characterized in that, The control of the corresponding target vision sensor operation specifically includes: When the target visual mode is a vector-level visual mode, a conventional camera and a millimeter-wave radar are identified as the target visual sensors, and the conventional camera and the millimeter-wave radar are controlled to operate. When the target visual mode is a contour-level visual mode, the high-definition camera and the LiDAR are identified as the target visual sensors, and the high-definition camera and the LiDAR are controlled to operate according to a preset low-frequency sampling mode; When the target vision mode is a high-precision vision mode, the high-definition camera, the lidar, and the infrared camera are identified as the target vision sensors, and the high-definition camera, the lidar, and the infrared camera are controlled to operate according to a preset high-frequency sampling mode.

7. A vehicle vision mode switching method according to any one of claims 1 to 6, characterized in that: When the current vehicle is in the vector-level vision mode, the visual data collected by the target vision sensor is used to perform lightweight feature extraction on the target object, thereby obtaining the motion velocity vector and coarse position information of the target object; When the current vehicle is in the contour-level vision mode, the visual data collected by the target vision sensor is used to extract structured features of the target object, thereby obtaining the edge contour, semantic category and motion trajectory information of the target object; When the current vehicle is in the high-precision vision mode, the visual data collected by the target vision sensor is used to perform full-element perception of the target object, thereby obtaining the three-dimensional structure, detailed semantics and high-precision pose information of the target object.

8. A vehicle vision mode switching device, characterized in that, include: The data acquisition module is used to acquire the current vehicle's real-time distance, threat ambiguity coefficient, real-time speed, traffic congestion status, dynamic balance level, and driver's driving habits and style. The visual pattern decision module is used to input the real-time vehicle distance, the threat ambiguity coefficient, the real-time vehicle speed, the traffic congestion status, the dynamic balance level, and the driving habit style into the pre-trained visual pattern decision model to obtain the target visual pattern. The vision mode switching module is used to switch the current vehicle's vision mode to the target vision mode and control the operation of the corresponding target vision sensor. The target visual mode is one of the following: vector-level visual mode, contour-level visual mode, and high-precision visual mode.

9. An electronic device, characterized in that, include: At least one processor; At least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements a vehicle vision mode switching method as described in any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements a vehicle vision mode switching method as described in any one of claims 1 to 7.

Images