Obstacle avoidance processing method and apparatus for vehicle, vehicle, and storage medium

Abstract

An obstacle avoidance processing method and apparatus for a vehicle, a vehicle, and a storage medium. The method comprises: receiving a driving scenario image of a vehicle in a driving direction, wherein the driving scenario image at least comprises an ego lane; and classifying road condition information on the ego lane on the basis of a preset classification model to obtain corresponding driving decision-making information, wherein the road condition information includes: an unconventional obstacle is present in the ego lane, a preset obstacle is present in the ego lane or no obstacle is present in the ego lane. According to the solution provided by the present application, driving decision-making information can be efficiently generated, thereby shortening the system response time, enhancing the reliability of obstacle avoidance strategies, and reducing the probabilities of missed triggering and false triggering.

Description

Obstacle avoidance processing method and device of vehicle, vehicle and storage medium

[0001] The present application claims priority from the Chinese patent application No. 2024104251958, filed on April 9, 2024, and entitled "Obstacle avoidance processing method and device of vehicle, vehicle and storage medium", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0002] The present application relates to the technical field of automatic driving, in particular to an obstacle avoidance processing method and device of vehicle, vehicle and storage medium. BACKGROUND

[0003] In the process of driving, the vehicle with automatic driving function needs to identify the obstacle and execute the driving decision to avoid the obstacle if the obstacle appears in front of the vehicle. In the related art, especially for static obstacles, such as barriers placed on the lane, the generation of the driving decision of the vehicle needs the collaborative processing of the perception module, the fusion module and the planning module. The perception module identifies the obstacle and the lane line, the fusion module confirms the position of the obstacle in the lane by performing three-dimensional map reconstruction, and the planning module selects the corresponding obstacle avoidance strategy according to the position and the danger level of the obstacle.

[0004] However, the technical solution of the collaborative operation of the "perception-fusion-planning" multi-module needs to process information from different sources, so that the implementation mode of the whole technical solution is complex and the system delay is high. In addition, the three-dimensional reconstruction of the obstacle at a long distance has the problem of inaccuracy, which is difficult to guarantee the reliability and completeness of the obstacle avoidance strategy. TECHNICAL SOLUTION

[0005] The first aspect of the present application provides an obstacle avoidance processing method of vehicle, comprising: receiving a driving scene image of a vehicle in a driving direction, the driving scene image at least comprising a self-lane; classifying road condition information on the self-lane according to a preset classification model to obtain corresponding driving decision information; wherein the road condition information comprises: the self-lane exists an irregular obstacle, the self-lane exists a preset obstacle or the self-lane has no obstacle.

[0006] In the obstacle avoidance processing method of the vehicle, the driving decision information comprises lane changing, braking or keeping driving, wherein: if the road condition information is that the self-lane exists an irregular obstacle, the driving decision information is braking; if the road condition information is that the self-lane exists a preset obstacle, the driving decision information is lane changing; and if the road condition information is that the self-lane has no obstacle, the driving decision information is keeping driving.

[0007] In the obstacle avoidance processing method of the vehicle, the preset obstacle is a static obstacle; if the road condition information is that the lane of the ego vehicle has the preset obstacle, the driving decision information is lane changing, comprising: if the road condition information is that the lane of the ego vehicle has the preset obstacle, according to the placement position of the preset obstacle, the driving decision information is left lane changing or right lane changing.

[0008] In the obstacle avoidance processing method of the vehicle, the non-conventional obstacle includes a vehicle rolled over, a wheel, a box or garbage; and the preset obstacle includes a roadblock.

[0009] In the obstacle avoidance processing method of the vehicle, the training method of the preset classification model comprises: constructing training data according to collected training images and corresponding decision labels; wherein the corresponding decision label is labeled according to the road condition information in each frame of the training image; the decision label includes lane changing, braking or keeping driving; and the preset classification model is trained by using the training data to obtain a trained preset classification model.

[0010] In the obstacle avoidance processing method of the vehicle, the training image includes a real scene image and / or a simulation image; wherein at least part of the real scene image is a scene image containing a preset obstacle or a conventional obstacle; and the simulation image is a scene image containing a non-conventional obstacle.

[0011] In the obstacle avoidance processing method of the vehicle, part of the real scene image contains a scene image in which there is no obstacle in front of the lane of the ego vehicle.

[0012] In the obstacle avoidance processing method of the vehicle, the driving scene image further includes an adjacent lane; if the road condition information is that the adjacent lane has an obstacle, the driving decision information includes lane left intrusion or lane right intrusion.

[0013] In the obstacle avoidance processing method of the vehicle, the method further comprises: sending the driving decision information to a preset planning module for processing to output path planning information.

[0014] The second aspect of the present application provides an obstacle avoidance processing device for a vehicle, comprising an image receiving module and a classification decision module. Wherein: the image receiving module is used for receiving a driving scene image of the vehicle in the driving direction, and the driving scene image at least includes a lane of the ego vehicle; the classification decision module is used for classifying road condition information on the lane of the ego vehicle according to a preset classification model to obtain corresponding driving decision information; wherein the road condition information includes: a non-conventional obstacle exists in the lane of the ego vehicle, a preset obstacle exists in the lane of the ego vehicle, or no obstacle exists in the lane of the ego vehicle.

[0015] The third aspect of the present application provides a vehicle comprising a processor and a memory. The memory stores executable code. When the executable code is executed by the processor, the processor performs the obstacle avoidance processing method of the vehicle as described above.

[0016] The fourth aspect of the present application provides a computer readable storage medium, which stores executable code. When the executable code is executed by the processor of the vehicle, the processor performs the obstacle avoidance processing method of the vehicle as described above.

[0017] The fifth aspect of the present application provides a computer program product comprising a computer program, which is used to execute computer program code instructions for performing part or all steps of the obstacle avoidance processing method of the vehicle as described above. Advantages

[0018] The obstacle avoidance processing method of the vehicle of the present application classifies the road condition information in the driving direction of the vehicle lane based on the real-time collected driving scene image, quickly and accurately outputs the corresponding driving decision information for different types of road condition information, especially for the presence of preset obstacles, so that the vehicle of automatic driving can avoid obstacles based on shorter response time, improve processing efficiency, and ensure safe driving.

[0019] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0020] The above and other objects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings, in which like reference characters designate like elements throughout the several views.

[0021] FIG. 1 is a flowchart of the obstacle avoidance processing method of the vehicle according to an embodiment of the present application;

[0022] FIG. 2 is a training image in which a preset obstacle facing the left front is placed in the vehicle lane according to an embodiment of the present application;

[0023] FIG. 3 is a training image in which a preset obstacle facing the right front is placed in the vehicle lane according to an embodiment of the present application;

[0024] FIG. 4 is another flowchart of the obstacle avoidance processing method of the vehicle according to an embodiment of the present application;

[0025] FIG. 5 is a training image in which an irregular obstacle exists in the vehicle lane according to an embodiment of the present application;

[0026] FIG. 6 is a structural diagram of the obstacle avoidance processing device of the vehicle according to an embodiment of the present application;

[0027] FIG. 7 is a structural schematic diagram of a vehicle according to an embodiment of the present application. Embodiments of the present application

[0028] Embodiments of the present application will be described in more detail with reference to the drawings. Although embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0029] The terminology used in the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used in the present application and the appended claims, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0030] It should be understood that although the terms "first", "second", "third" and the like can be used herein to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one piece of information from another piece of information of the same type. For example, the first information can also be referred to as the second information, and similarly, the second information can also be referred to as the first information without departing from the scope of the present application. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "a plurality of" is two or more, unless otherwise specifically limited.

[0031] The technical solutions of the present application will be described in detail below in conjunction with the drawings.

[0032] FIG. 1 is a flowchart of a method for obstacle avoidance processing of a vehicle according to an embodiment of the present application.

[0033] Referring to FIG. 1, the method for obstacle avoidance processing of a vehicle according to an embodiment of the present application includes:

[0034] S110, receiving a driving scene image in a driving direction of the vehicle, the driving scene image at least including a lane of the ego vehicle.

[0035] During normal driving of the vehicle, a camera installed on the vehicle can capture images in front of the driving direction in real time, i.e., obtain real-time driving scene images. It can be understood that the lane in which the vehicle is driving is the lane of the ego vehicle, and the driving scene image can include the lane of the ego vehicle in the driving direction.

[0036] For example, the vehicle travels along the self-lane towards the front, and the travel scene image includes a real scene image of the front of the self-lane, so that the road condition information of the front of the self-lane can be displayed at the corresponding position in the image.

[0037] In this application, the simulation of what the human eye sees can be realized according to the travel scene image, so that compared with the radar sensor which is limited by the sensing distance, the obstacle farther away from the vehicle can be quickly identified based on the travel scene image, which is not limited by the distance, so that the system can quickly plan the obstacle avoidance.

[0038] In S120, the road condition information on the self-lane is classified according to the preset classification model to obtain corresponding driving decision information; wherein the road condition information includes: the self-lane has a non-conventional obstacle, the self-lane has a preset obstacle, or the self-lane has no obstacle.

[0039] It can be understood that there can be an obstacle in front of the self-lane of the vehicle, or there can be no obstacle. Further, the road condition information includes that the self-lane has a non-conventional obstacle, the self-lane has a preset obstacle, or the self-lane has no obstacle. The non-conventional obstacle may, for example, be a vehicle that has rolled over, a wheel, a box (such as a paper box, a cargo basket, etc.), garbage, and other abnormal, rare or random obstacles that hinder the vehicle from driving. The preset obstacle refers to a static obstacle, which may, for example, be a roadblock such as a popsicle tube, a fence, etc., which are only used as examples and are not limited.

[0040] In some specific embodiments, if the road condition information is that the self-lane has a non-conventional obstacle, the driving decision information is braking; if the road condition information is that the self-lane has a preset obstacle, the driving decision information is lane changing; and if the road condition information is that the self-lane has no obstacle, the driving decision information is maintaining driving. It can be understood that for non-conventional obstacles, i.e. abnormal and rare obstacles, in order to ensure the safety of the vehicle driving, the driving decision information of braking can be output. For the preset obstacle, the preset classification model has been sufficiently trained in the training process and can accurately identify, so that the driving decision information of lane changing can be used for obstacle avoidance. When the self-lane has no obstacle, the vehicle can normally drive without obstacle avoidance.

[0041] In this step, based on whether there is an obstacle in front of the self-lane and the type of the obstacle, the corresponding driving decision information is directly generated without introducing other data processing steps of different functional modules, improving the data processing efficiency and the error rate of triggering, so that the system can quickly and accurately control the vehicle according to the obtained driving decision information.

[0042] In some embodiments, the preset classification model can be a deep learning model based on a CNN (Convolutional Neural Network) network, a model that can classify the input driving scene image by being pre-trained with training data. The preset classification model can also be other models capable of classifying images in related technologies, such as the Transformer model, which is only an example and is not limited.

[0043] As can be seen from the example, the obstacle avoidance processing method of the vehicle of the present application classifies the road condition information in the driving direction of the vehicle's lane based on the real-time collected driving scene image, quickly and accurately outputs the corresponding driving decision information for different types of road condition information, so that the vehicle can avoid obstacles based on a shorter response time, improve processing efficiency, and ensure safe driving.

[0044] An embodiment of the present application also provides a training method of a preset classification model, comprising:

[0045] S210, constructing training data according to the collected training images and corresponding decision labels; wherein the corresponding decision labels are labeled according to the road condition information in each frame of training image.

[0046] In some embodiments, the training images include real scene images and / or simulation images; wherein the real scene images are scene images containing preset obstacles or conventional obstacles; and the simulation images are scene images containing unconventional obstacles. That is, the real scene images are images captured by real shooting, and at least part of the real scene images contain scenes where there are preset obstacles or conventional obstacles in front of the vehicle lane. That is, some real scene images contain scenes where there are preset obstacles in front of the vehicle lane, and some real scene images contain scenes where there are conventional obstacles in front of the vehicle lane. A small part of the real scene images contain scenes where there are no obstacles in front of the vehicle lane. By enriching the various scenes of the real scene images, the model can be more comprehensively trained to recognize different scenes and trigger in time and accurately. The conventional obstacles can be, for example, other normally driving vehicles in front of the current vehicle, and the types of other vehicles are not limited, such as sedans, trucks, bicycles, and motorcycles.

[0047] Further, the data amount of the real scene image with the unconventional obstacle is less, so there is not enough training data for training the model. Based on this, the simulation image is designed to make up for the lack of rare scenes. In some embodiments, the simulation image can be a completely virtually designed image, or an image obtained by combining part of the real scene and part of the virtual elements, or an image obtained by combining different real scene images. For example, a real scene image containing the ego lane can be taken as the background, and a virtually designed unconventional obstacle can be placed on the ego lane of the real scene image to form a frame of simulation image. For another example, the AIGC (Artificial Intelligence Generated Content) can also be used to generate a series of simulation images meeting the requirements, which are only examples and are not limited. The training method of the present application can improve the triggering rate of the preset classification model after training by using the real scene image containing the conventional obstacle and the simulation image containing the unconventional obstacle as the training data, and can avoid the situation of missed triggering, so as to generate driving decision information in time and improve the safety of autonomous driving.

[0048] In some embodiments, according to the road condition information in the training image, i.e., the obstacle position and obstacle type of the ego lane, the corresponding decision label is set in advance. The decision label can include lane changing, braking, or keeping driving. That is, each decision label corresponds to a driving decision information. By associating the decision label with the road condition information of the training image, the model can learn the mapping relationship between the road condition information and the driving decision information during the training process.

[0049] In some specific embodiments, the decision label of lane changing can also include left lane changing and right lane changing. When the training image contains a preset obstacle with a preset placement position, the label of the training image is set as left lane changing or right lane changing. As shown in FIG. 2, the preset obstacle is a plurality of static obstacles on the ego lane, such as a plurality of ice cream tubes. The plurality of static obstacles are regularly placed towards the left front, indicating that the vehicle needs to change lanes to the left, and the decision label of the training image is left lane changing. Similarly, as shown in FIG. 3, when the plurality of static obstacles are regularly placed towards the right front, it indicates that the vehicle needs to change lanes to the right, and the decision label of the training image corresponds to right lane changing. By introducing the training image with the preset obstacle, the obstacle avoidance performance of the model for the static obstacle can be efficiently improved.

[0050] In some embodiments, the training image can also include a scenario in which there is an obstacle in the adjacent lane. If the adjacent lane is located on the left side of the lane of the ego vehicle and there is an obstacle, the decision label of the training image is increased by left intrusion; if the adjacent lane is located on the right side of the lane of the ego vehicle and there is an obstacle, the decision label of the training image is increased by right intrusion; if the left side and the right side of the lane of the ego vehicle both have adjacent lanes and there are obstacles, the decision label of the training image is increased by both left intrusion and right intrusion. That is, the adjacent vehicle is first determined to be located in the corresponding direction of the lane of the ego vehicle, and then the road condition information on the adjacent lane is further determined. By marking the corresponding decision label according to the road condition information of the lane of the ego vehicle while also increasing the corresponding decision label according to the road condition information of the adjacent lane, the subsequent trained preset classification model can output more rich driving decision information for reference.

[0051] In this step, a plurality of training images containing different road condition information and corresponding decision labels are combined to form training data for subsequent training of the preset classification model.

[0052] S220, training the preset classification model using the training data to obtain a trained preset classification model.

[0053] In this step, the preset classification model can be a deep learning model based on a convolutional neural network. Based on the above training data as input data, a preset classification model that can output corresponding driving decision information according to different driving scene images can be trained through model iteration.

[0054] The preset classification model of the present application can be trained by various real scene images and simulation images with different scenes to form a classification model that can output corresponding driving decision information according to input driving scene images in an end-to-end manner, efficiently improve model performance and system performance, avoid situations of missed triggering and false triggering, and improve the accuracy and timeliness of vehicle obstacle avoidance.

[0055] Referring to FIG. 4, an embodiment of the present application also provides a vehicle obstacle avoidance processing method, comprising:

[0056] S310, receiving a driving scene image of a vehicle in a driving direction, the driving scene image including a lane of the ego vehicle and / or an adjacent lane.

[0057] In the driving scene image captured in real time during the driving of the vehicle, the driving scene image can include not only the lane of the ego vehicle, but also the adjacent lane on the left and / or right side of the lane of the ego vehicle. That is, in addition to the existence of an obstacle in front of the vehicle in the lane of the ego vehicle, there can also be an obstacle in front of the adjacent lane.

[0058] S320, classify the road condition information on the self-vehicle lane and the adjacent lane according to the preset classification model, and obtain the corresponding classification result.

[0059] In this step, in addition to classifying the road condition information on the self-vehicle lane, the road condition information on the adjacent lane can also be classified to obtain the classification result of the road condition information corresponding to different lanes. In some embodiments, the main driving decision information is generated based on the road condition information corresponding to the self-vehicle lane, and the auxiliary driving decision information is generated based on the road condition information corresponding to the adjacent lane. The main driving decision information includes lane changing, braking, or keeping driving, and the auxiliary driving decision information includes intrusion on the left side of the lane and / or intrusion on the right side of the lane.

[0060] S330, if the road condition information is that there is an irregular obstacle on the self-vehicle lane, the output driving decision information is braking.

[0061] Referring to FIG. 5, in this step, if the preset classification model identifies that the road condition information on the self-vehicle lane is that there is an irregular obstacle on the self-vehicle lane, such as an abnormal, rare, or randomly appearing obstacle that hinders the vehicle from driving, for example, a vehicle that has rolled over, a wheel, a carton, garbage, etc., the corresponding driving decision information braking is output.

[0062] S340, if the road condition information is that there is a preset obstacle on the self-vehicle lane, according to the placement position of the preset obstacle, the output driving decision information is left lane changing or right lane changing.

[0063] Referring to FIGS. 2 and 3, in this step, the preset classification model identifies that the road condition information on the self-vehicle lane is that there is a preset obstacle on the self-vehicle lane, for example, a plurality of roadblocks placed in sequence toward the left front in front of the self-vehicle lane, and the output driving decision information is left lane changing; for example, a plurality of roadblocks placed in sequence toward the right front in front of the self-vehicle lane, and the output driving decision information is right lane changing. The preset classification model of the present application can efficiently output obstacle avoidance strategies for static obstacles.

[0064] It can be understood that, in this step, if the road condition information is that there is a regular obstacle on the self-vehicle lane, the output driving decision information is lane changing.

[0065] S350, if the road condition information is that there is no obstacle on the self-vehicle lane, the output driving decision information is keeping driving.

[0066] Obviously, if the preset classification model identifies that the road condition information on the self-vehicle lane is that there is no obstacle, the driving decision information is keeping driving, that is, the vehicle does not need to avoid obstacles and can continue to drive normally along the self-vehicle lane.

[0067] The steps S330 to S350 are executed according to the real road condition information corresponding to the driving scene image. It can be understood that if there is an obstacle on the lane of the ego vehicle in the driving scene image, the interval distance between the obstacle and the current vehicle can be ignored, and no matter how far the actual interval distance is, the corresponding driving decision information can be executed.

[0068] In S360, if the road condition information is that there is an obstacle in the adjacent lane, the output driving decision information includes that there is an intrusion on the left side of the lane or there is an intrusion on the right side of the lane.

[0069] While identifying the road condition information of the lane of the ego vehicle, the preset classification model can also identify the road condition information of the adjacent lane at the same time. For example, if there is an adjacent lane on the left side of the lane of the ego vehicle and there is an obstacle in the adjacent lane on the left side, the preset classification model not only outputs the driving decision information corresponding to the lane of the ego vehicle, but also synchronously outputs the driving decision information corresponding to the adjacent lane, such as that there is an intrusion on the left side of the lane; or the preset classification model only outputs the driving decision information corresponding to the lane of the ego vehicle.

[0070] The step S360 can be executed at the same time as one of the steps S330 to S350 or after the steps S330 to S350, or S360 is selectively executed.

[0071] In some embodiments, after obtaining the driving decision information, the driving decision information is sent to a preset planning module for processing to output path planning information.

[0072] It can be understood that according to the preset classification model, the driving decision information can be output faster, and then the preset planning module can be sent earlier for path planning, so that the vehicle can automatically drive according to the latest path planning information.

[0073] Compared with the conventional driving decision information obtained by using multiple functional modules for step-by-step processing, the present application can directly generate the driving decision information through the preset classification model in an end-to-end manner, shorten the processing flow of the entire system, and then shorten the response time of the system, and can cope with a longer detection distance, obtain an obstacle avoidance strategy faster, and reserve more sufficient time and distance for the vehicle to complete path planning and lane changing distance, thereby improving the safety performance of automatic driving.

[0074] Corresponding to the foregoing application function implementation method embodiments, the present application also provides an obstacle avoidance processing device of a vehicle, a vehicle, and corresponding embodiments.

[0075] FIG. 6 is a structural schematic diagram of an obstacle avoidance processing device of a vehicle according to an embodiment of the present application.

[0076] Referring to FIG. 6, the obstacle avoidance processing device of a vehicle according to an embodiment of the present application includes an image receiving module 610 and a classification decision module 620. Wherein:

[0077] The image receiving module 610 is configured to receive a driving scene image of the vehicle in a driving direction, and the driving scene image at least includes a lane of the ego vehicle.

[0078] The classification decision module 620 is configured to classify road condition information on the lane of the ego vehicle according to a preset classification model, and obtain corresponding driving decision information; wherein the road condition information includes that there is an irregular obstacle on the lane of the ego vehicle, there is a preset obstacle on the lane of the ego vehicle, or there is no obstacle on the lane of the ego vehicle.

[0079] In a specific embodiment, the classification decision module 620 is configured to, if the road condition information is that there is an irregular obstacle on the lane of the ego vehicle, the driving decision information is braking; if the road condition information is that there is a preset obstacle on the lane of the ego vehicle, the driving decision information is lane changing; and if the road condition information is that there is no obstacle on the lane of the ego vehicle, the driving decision information is maintaining driving.

[0080] In a specific embodiment, the classification decision module 620 is configured to, if the road condition information is that there is a preset obstacle on the lane of the ego vehicle, according to the placement position of the preset obstacle, the driving decision information is left lane changing or right lane changing.

[0081] In a specific embodiment, the classification decision module 620 is configured to, if the road condition information is that there is an obstacle on the adjacent lane, the driving decision information includes that there is an intrusion on the left side of the lane or there is an intrusion on the right side of the lane.

[0082] As can be seen from the example, the obstacle avoidance processing device of the vehicle of the present application can directly output driving decision information based on a preset classification model, obtain accurate and timely obstacle avoidance strategies, shorten the processing flow, reduce the response time of the system, and be suitable for a longer detection distance, thereby improving the safety factor when the vehicle is automatically driven.

[0083] As to the device in the above embodiments, the specific manner in which each module performs operations has been described in detail in the embodiments of the method, and will not be described in detail here.

[0084] FIG. 7 is a structural schematic diagram of a vehicle according to an embodiment of the present application.

[0085] Referring to FIG. 7, the vehicle 1000 includes a memory 1010 and a processor 1020.

[0086] The processor 1020 can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like. The general-purpose processor can be a microprocessor, or the processor can be any conventional processor.

[0087] The memory 1010 can include various types of storage units, such as a system memory, a read-only memory (ROM), and a permanent storage device. Among them, the ROM can store static data or instructions required by the processor 1020 or other modules of the computer. The permanent storage device can be a readable and writable storage device. The permanent storage device can be a non-volatile storage device that does not lose stored instructions and data even after the computer is powered off. In some embodiments, the permanent storage device uses a mass storage device (such as a magnetic or optical disk, a flash memory) as a permanent storage device. In some other embodiments, the permanent storage device can be a removable storage device (such as a floppy disk, an optical drive). The system memory can be a readable and writable storage device or a volatile readable and writable storage device, such as a dynamic random access memory. The system memory can store some or all instructions and data required by the processor during runtime. In addition, the memory 1010 can include a combination of any computer readable storage media, including various types of semiconductor memory chips (such as DRAM, SRAM, SDRAM, flash memory, programmable read-only memory), magnetic disks and / or optical disks. In some embodiments, the memory 1010 can include a readable and / or writable removable storage device, such as a compact disc (CD), a read-only digital versatile disc (such as DVD-ROM, double-layer DVD-ROM), a read-only Blu-ray disc, an ultra-density optical disc, a flash memory card (such as an SD card, a min SD card, a Micro-SD card, etc.), a magnetic floppy disk, etc. The computer readable storage medium does not include a carrier wave and a transient electronic signal transmitted through wireless or wired transmission.

[0088] The memory 1010 stores executable code, which, when processed by the processor 1020, can cause the processor 1020 to perform some or all of the steps in the above-mentioned methods.

[0089] Alternatively, the present application can also be implemented as a server, which can include a memory and a processor; similarly to the vehicle described above, the memory of the server stores executable code, which, when processed by the processor, can cause the processor to perform some or all of the steps of the obstacle avoidance processing method of the vehicle described above.

[0090] In addition, the method according to the present application can also be implemented as a computer program or computer program product, which includes computer program code instructions for performing some or all of the steps of the obstacle avoidance processing method of the vehicle described above.

[0091] Alternatively, the present application can also be implemented as a computer readable storage medium (or non-transitory machine readable storage medium or machine readable storage medium) having stored executable code (or computer program or computer instruction code) which, when executed by a processor of a vehicle (or a server, etc.), causes the processor to perform some or all of the steps of the method described above according to the present application.

[0092] The embodiments of the present application have been described above, the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical application or improvement of the technology in the market, or to enable other ordinary skilled in the art to understand the embodiments disclosed herein.

Claims

1. A vehicle obstacle avoidance method, characterized in that: include: Receiving a driving scene image of the vehicle in a driving direction, the driving scene image at least including a lane of the vehicle; The road condition information on the own vehicle lane is classified according to a preset classification model to obtain corresponding driving decision information; wherein the road condition information includes: the presence of an unconventional obstacle in the own vehicle lane, the presence of a preset obstacle in the own vehicle lane, or the absence of an obstacle in the own vehicle lane.

2. The method according to claim 1, characterized in that The driving decision information includes changing lanes, braking, or maintaining driving, wherein: If the road condition information indicates that there is an unconventional obstacle in the vehicle's lane, the driving decision information is braking; If the road condition information indicates that there is a preset obstacle in the vehicle's lane, the driving decision information is to change lanes; If the road condition information indicates that there are no obstacles in the vehicle's lane, the driving decision information is to keep driving.

3. The method according to claim 2, characterized in that The preset obstacle is a static obstacle; If the road condition information indicates that there is a preset obstacle in the vehicle's lane, the driving decision information is lane change, including: If the road condition information indicates that there is a preset obstacle in the vehicle's lane, the driving decision information is to change lanes to the left or to the right based on the placement of the preset obstacle.

4. The method according to claim 2, wherein: The unconventional obstacles include overturned vehicles, wheels, boxes or garbage; The preset obstacles include roadblocks.

5. The method according to claim 1, wherein The training method of the preset classification model includes: Constructing training data based on the collected training images and corresponding decision labels; wherein the corresponding decision label is annotated according to the road condition information in each frame of the training image; the decision label includes changing lanes, braking, or maintaining driving; The training data is used to train the preset classification model to obtain a trained preset classification model.

6. The method according to claim 5, characterized in that The training images include: real scene images and / or simulation images; Among them, at least part of the real scene images are scene images containing preset obstacles or conventional obstacles; and the simulation images are scene images containing unconventional obstacles.

7. The method according to claim 6, characterized in that: Some of the real scene images include scene images in which there are no obstacles in front of the vehicle lane.

8. The method according to claim 1, characterized in that The driving scene image also includes adjacent lanes; The method further includes: if the road condition information indicates that there is an obstacle in an adjacent lane, the driving decision information includes intrusion on the left side of the lane or intrusion on the right side of the lane.

9. The method according to claim 1, characterized in that The method further comprises: The driving decision information is sent to the preset planning module for processing to output path planning information.

10. A vehicle obstacle avoidance device, characterized in that: include: An image receiving module is used to receive a driving scene image of the vehicle in the driving direction, wherein the driving scene image at least includes the vehicle lane; A classification decision module is used to classify the road condition information on the own vehicle lane according to a preset classification model to obtain corresponding driving decision information; wherein the road condition information includes: the presence of unconventional obstacles in the own vehicle lane, the presence of preset obstacles in the own vehicle lane, or the absence of obstacles in the own vehicle lane.

11. A vehicle, characterized in that: include: processor; as well as A memory having executable code stored thereon, which, when executed by the processor, causes the processor to execute the vehicle obstacle avoidance processing method according to any one of claims 1 to 9.

12. A computer-readable storage medium having executable code stored thereon, which, when executed by a processor of a vehicle, causes the processor to execute the vehicle obstacle avoidance method according to any one of claims 1 to 9.

13. A computer program product comprising a computer program, characterized in that The computer program is used to execute computer program code instructions of some or all steps in the vehicle obstacle avoidance processing method described in any one of claims 1-9.

Images