Assistant method and system for automatic driving in severe weather based on artificial intelligence

Abstract

The invention provides a severe weather automatic driving assistance method and system based on artificial intelligence, and the method comprises the steps: collecting a vehicle advancing road image and a surrounding environment image in severe weather in real time, and carrying out the preprocessing of the vehicle advancing road image and the surrounding environment image; determining driving characteristics of the vehicle based on the preprocessing result, and determining a current automatic driving assistance strategy of the vehicle based on the driving characteristics; and performing auxiliary driving on the vehicle based on the automatic driving auxiliary strategy, and dynamically updating the automatic driving auxiliary strategy based on the driving state of the vehicle. The reliability, accuracy and stability of automatic driving assistance in severe weather are guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an artificial intelligence-based severe weather automatic driving assistance method and system. Background Art

[0002] At present, with the continuous development of autonomous driving technology, the autonomous driving function of vehicles under normal weather conditions is gradually becoming mature;

[0003] However, severe weather conditions (such as heavy rain, snow, and dense fog) remain a significant challenge for autonomous driving. Reduced visibility, altered road friction, and interference with sensor performance severely impact the safety and reliability of autonomous driving. Traditional approaches struggle to adapt to complex and changing inclement weather conditions, leading to inaccurate detection of the road environment and other vehicles on the road and significant errors. This significantly reduces the stability and reliability of autonomous driving assistance in inclement weather.

[0004] Therefore, in order to overcome the above-mentioned defects, the present invention provides an artificial intelligence-based severe weather automatic driving assistance method and system. Summary of the Invention

[0005] The present invention provides an artificial intelligence-based method and system for automatic driving assistance in severe weather, which is used to improve the preprocessing effect of the vehicle's forward road image and surrounding environment image in severe weather, thereby achieving accurate and effective determination of the vehicle's driving characteristics based on the preprocessing effect, and then achieving effective construction of automatic driving assistance strategies based on the driving characteristics, and finally achieving assisted driving of the vehicle according to the automatic driving assistance strategies, and dynamically updating the automatic driving assistance strategies during the driving process, thereby ensuring the reliability, accuracy and stability of automatic driving assistance in severe weather.

[0006] The present invention provides an artificial intelligence-based automatic driving assistance method for severe weather conditions. Figure 1 Shown, including:

[0007] Step 1: Real-time acquisition of vehicle road images and surrounding environment images in severe weather conditions, and pre-processing of the vehicle road images and surrounding environment images;

[0008] Step 2: Determine the vehicle's driving characteristics based on the preprocessing results, and determine the vehicle's current automatic driving assistance strategy based on the driving characteristics;

[0009] Step 3: Assist the vehicle in driving based on the automatic driving assistance strategy, and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

[0010] Preferably, an artificial intelligence-based inclement weather autonomous driving assistance method includes, in step 1, before real-time acquisition of vehicle forward road images and surrounding environment images in inclement weather, the following steps:

[0011] Based on the sensors deployed on the vehicle, the corresponding working area is pre-collected, and the spatial range of the pre-collected images is identified to obtain the image acquisition angle of each sensor;

[0012] Based on the spatial range recognition results, the image capturing range of each sensor under the image acquisition angle is determined. When the image capturing range does not meet the vehicle's automatic driving assistance requirements, the image capturing angle of each sensor is adaptively corrected multiple times until the image capturing range meets the vehicle's automatic driving assistance requirements.

[0013] Preferably, an artificial intelligence-based inclement weather autonomous driving assistance method includes, in step 1, collecting images of the vehicle's forward road and surrounding environment in real time under inclement weather conditions, including:

[0014] Receive automatic driving assistance requests from users, perform self-checks on the vehicle's driving status, and enters the automatic driving state after passing the self-checks;

[0015] After the self-test passes, each sensor is synchronously controlled in parallel, and based on the parallel control results, each sensor is controlled to collect images of the vehicle's forward road and surrounding environment in real time under severe weather conditions;

[0016] The collected images of the vehicle's forward road and surrounding environment are marked and transmitted to the image processing center.

[0017] Preferably, an artificial intelligence-based method for assisting autonomous driving in severe weather conditions marks collected images of the vehicle's path and surrounding environment and transmits them to an image processing center, including:

[0018] Establish independent communication links between each sensor and the image processing center, and transmit the vehicle's road ahead image and surrounding environment image collected by different sensors to the image processing center based on the independent communication links. The independent communication links mean that the communication links between each sensor and the image processing center are independent and do not affect each other;

[0019] Based on the image processing center, the vehicle's forward road images and surrounding environment images transmitted by different sensors are distinguished and cached.

[0020] Preferably, in an artificial intelligence-based severe weather autonomous driving assistance method, in step 1, preprocessing the vehicle's forward road image and surrounding environment image includes:

[0021] Obtaining the obtained vehicle forward road image and surrounding environment image, and inputting the vehicle forward road image and surrounding environment image into the filter for denoising;

[0022] Collect image datasets under different severe weather scenarios and generate labels corresponding to different severe weather conditions;

[0023] Annotate the corresponding image dataset based on the marked labels to obtain an annotated image dataset, and train a trainable model framework based on the annotated image dataset to obtain an inclement weather recognition model;

[0024] The denoised images of the vehicle's forward road and surrounding environment are input into the severe weather recognition model for identification to obtain the severe weather type the vehicle is currently in.

[0025] Determining interference objects in the vehicle's forward road image and the surrounding environment image based on the type of severe weather, and determining first morphological features of the interference objects;

[0026] Locking a target pixel of an interference object in an image of the vehicle's forward road and an image of the surrounding environment based on the first morphological feature, and determining a second morphological feature of the main object based on neighboring pixels of the target pixel;

[0027] Determining depth information of the vehicle's forward road image and the vehicle's surrounding environment image based on the first morphological feature and the second morphological feature, and performing layer separation on the vehicle's forward road image and the vehicle's surrounding environment image based on the depth information;

[0028] Based on the layer splitting result, morphological processing is performed on the target pixel of the interference object corresponding to the first morphological feature, where the morphological processing includes corrosion, expansion, opening operation and closing operation.

[0029] Preferably, in an artificial intelligence-based severe weather autonomous driving assistance method, in step 2, determining the driving characteristics of the vehicle based on the preprocessing results, and determining the vehicle's current autonomous driving assistance strategy based on the driving characteristics, including:

[0030] Object extraction is performed on the preprocessed images of the vehicle's forward road and surrounding environment to obtain the distribution of road obstacles in the vehicle's current forward direction and the road parameters of the vehicle's current travel. The road parameters include road width and road flatness, and road obstacles include pedestrians, road vehicles, and fixed, immovable objects.

[0031] Determine the relative position and relative distance between the current vehicle and the road obstacle based on the distribution of road obstacles, and use the relative position and relative distance as the first consideration parameters, while using the road parameters as the second consideration parameters;

[0032] Based on the first consideration parameter and the second consideration parameter, available automatic driving assistance strategies are matched from a predefined series of driving strategies. At the same time, the influence coefficient of severe weather on automatic driving is determined, and local parameters in the available automatic driving assistance strategies are adaptively corrected based on the influence coefficient to obtain the final automatic driving assistance strategy.

[0033] Preferably, an artificial intelligence-based severe weather autonomous driving assistance method obtains a final autonomous driving assistance strategy, including:

[0034] The automated driving assistance strategy is split into task nodes to obtain a set of task nodes, and the specific data content corresponding to each task node is verified based on safe driving standards.

[0035] After the parameter value verification is passed, the automatic driving assistance strategy is authorized to take effect.

[0036] Preferably, in an artificial intelligence-based severe weather automatic driving assistance method, in step 3, the vehicle is assisted in driving based on the automatic driving assistance strategy, and the automatic driving assistance strategy is dynamically updated based on the driving state of the vehicle, including:

[0037] The automated driving assistance strategy is decomposed based on the vehicle operating components to obtain the control parameters of each vehicle operating component. The control parameters are then divided into stages based on the vehicle state change, which includes the improvement stage and the fully automated assisted driving stage. The vehicle state change is the degree of change in vehicle speed and steering amplitude.

[0038] Adjust the control parameters at different stages, and coordinately control the vehicle's operating components based on the parameter adjustment results at each stage to change the vehicle's state to assisted driving;

[0039] Based on the change results, the vehicle is controlled to perform trial assisted driving in the perfect driving stage, and the automatic driving assistance strategy is improved based on the trial assisted driving results. After a fixed mileage without accidents, the vehicle enters the fully automatic assisted driving stage;

[0040] At the same time, during the fully automatic assisted driving stage, the vehicle's driving status, the vehicle's forward road status and the surrounding environment status are monitored in real time, and the automatic driving assistance strategy is adjusted locally and dynamically in real time according to the vehicle's driving status, the vehicle's forward road status and the surrounding environment status until the assisted driving is completed.

[0041] The present invention provides an artificial intelligence-based severe weather automatic driving assistance system, comprising:

[0042] An image acquisition module is used to collect images of the vehicle's road ahead and surrounding environment in real time under severe weather conditions, and to pre-process the images of the vehicle's road ahead and surrounding environment;

[0043] A strategy formulation module, configured to determine the driving characteristics of the vehicle based on the preprocessing results, and determine the current automatic driving assistance strategy of the vehicle based on the driving characteristics;

[0044] The control module is used to assist the vehicle in driving based on the automatic driving assistance strategy and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

[0045] Preferably, an artificial intelligence-based severe weather automatic driving assistance system, an image acquisition module, comprises:

[0046] An image pre-capture unit is used to pre-capture images of the corresponding working area based on sensors deployed on the vehicle, and to perform spatial range recognition on the pre-captured images to obtain the image acquisition angles of each sensor;

[0047] The device configuration unit is used to determine the image capture range of each sensor under the image acquisition angle based on the spatial range recognition result, and when the image capture range does not meet the vehicle's automatic driving assistance requirements, perform multiple adaptive corrections on the image capture angle of each sensor until the image capture range meets the vehicle's automatic driving assistance requirements.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The preprocessing effect of the vehicle's forward road image and surrounding environment image in severe weather conditions is improved, so that the vehicle's driving characteristics can be accurately and effectively determined based on the preprocessing effect, and then the automatic driving assistance strategy can be effectively constructed based on the driving characteristics. Finally, the vehicle can be assisted in driving according to the automatic driving assistance strategy, and the automatic driving assistance strategy can be dynamically updated during driving, ensuring the reliability, accuracy and stability of automatic driving assistance in severe weather conditions.

[0050] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0051] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0053] Figure 1 This is a flow chart of an artificial intelligence-based severe weather autonomous driving assistance method according to an embodiment of the present invention;

[0054] Figure 2 This is a flowchart of step 1 in an artificial intelligence-based severe weather autonomous driving assistance method according to an embodiment of the present invention;

[0055] Figure 3 This is a structural diagram of an artificial intelligence-based severe weather automatic driving assistance system in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0057] In one embodiment, an artificial intelligence-based severe weather autonomous driving assistance method, such as Figure 1 Shown, including:

[0058] Step 1: Real-time acquisition of vehicle road images and surrounding environment images in severe weather conditions, and pre-processing of the vehicle road images and surrounding environment images;

[0059] Step 2: Determine the vehicle's driving characteristics based on the preprocessing results, and determine the vehicle's current automatic driving assistance strategy based on the driving characteristics;

[0060] Step 3: Assist the vehicle in driving based on the automatic driving assistance strategy, and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

[0061] In this embodiment, severe weather includes rainy weather, snowy weather, strong winds, sandstorms and other severe weather.

[0062] In this embodiment, the vehicle forward road image refers to a road image in the vehicle's driving direction when the vehicle is driving on the road, including vehicles and pedestrians on the road.

[0063] In this embodiment, the surrounding environment image refers to a spatial image around the vehicle, including information such as the width of the road.

[0064] In this embodiment, the preprocessing refers to removing noise from the vehicle's forward road image and the surrounding environment image and removing the influence of climate factors in a harsh environment.

[0065] In this embodiment, the driving characteristics refer to the environment in which the vehicle is currently located and the relative direction and distance between the vehicle and other vehicles.

[0066] In this embodiment, the driving state refers to information such as the current driving speed and driving trajectory of the vehicle.

[0067] The beneficial effects of the above technical solution are: improving the preprocessing effect of the vehicle's forward road image and surrounding environment image in severe weather, thereby realizing accurate and effective determination of the vehicle's driving characteristics based on the preprocessing effect, and then realizing effective construction of the automatic driving assistance strategy based on the driving characteristics, and finally realizing assisted driving of the vehicle according to the automatic driving assistance strategy, and dynamically updating the automatic driving assistance strategy during the driving process, ensuring the reliability, accuracy and stability of automatic driving assistance in severe weather.

[0068] In one embodiment, an artificial intelligence-based severe weather autonomous driving assistance method, such as Figure 2 As shown, in step 1, before collecting the vehicle's forward road image and surrounding environment image in real time under severe weather conditions, the following steps are included:

[0069] Step 101: Pre-collect images of the corresponding working area based on sensors deployed on the vehicle, and perform spatial range recognition on the pre-collected images to obtain image collection angles of each sensor;

[0070] Step 102: Determine the image capture range of each sensor under the image acquisition angle based on the spatial range recognition result, and when the image capture range does not meet the vehicle automatic driving assistance requirements, perform multiple adaptive corrections on the image capture angle of each sensor until the image capture range meets the vehicle automatic driving assistance requirements.

[0071] In this embodiment, image pre-capture refers to controlling the sensors deployed on the vehicle to perform simulated image capture, in order to ensure the accuracy and reliability of the angles and positions of the sensors deployed on the vehicle.

[0072] In this embodiment, spatial range recognition refers to determining the wide angle and range of the image captured by the sensor, in order to ensure the suitability of the current capture angle of the sensor.

[0073] In this embodiment, the vehicle automatic driving assistance requirements are set in advance, including specific requirements for the recognition range of road conditions and specific requirements for image clarity.

[0074] The beneficial effect of the above technical solution is: it ensures the reliability of the deployment of sensors on the vehicle, thereby ensuring the reliability of the sensor's acquisition of images of the vehicle's forward road and surrounding environment, and provides a guarantee for automatic driving assistance in severe weather.

[0075] In one embodiment, an artificial intelligence-based method for assisting autonomous driving in severe weather conditions includes, in step 1, collecting images of the vehicle's forward path and surrounding environment in real time in severe weather conditions, including:

[0076] Receive the user's request for automatic driving assistance, perform a self-check on the vehicle's driving status, and enter the automatic driving state after the self-check passes. Specifically:

[0077] Upon receiving an autonomous driving assistance request from a user, the system collects functional status data of each component related to autonomous driving and performs a status assessment on the functional status data of each component based on a vehicle status assessment index. The functional status data includes the current operating status data and corresponding performance data of each component. The vehicle status assessment index is known in advance and serves as a metric for evaluating whether the vehicle meets the autonomous driving criteria.

[0078] Based on the status assessment results, the system performs self-inspections on the performance status of each component. Once the autonomous driving conditions are met, the system responds to the user's autonomous driving assistance request, receives permission to drive the vehicle, and enters the autonomous driving state.

[0079] After the self-test passes, each sensor is synchronously controlled in parallel, and based on the parallel control results, each sensor is controlled to collect images of the vehicle's forward road and surrounding environment in real time under severe weather conditions. Specifically:

[0080] Based on the response results, each sensor is synchronously started in parallel, and based on the parallel start results, each sensor is controlled to collect multiple images of the vehicle's forward road image and surrounding environment image at the same point in severe weather;

[0081] The collected images of the vehicle's forward road and surrounding environment are marked and transmitted to the image processing center. Specifically:

[0082] The collected images of the vehicle's road ahead and surrounding environment are traced to the device terminal to obtain the corresponding relationship between each vehicle's road ahead and surrounding environment image and each sensor. The device terminal traceability is to determine the sensor corresponding to each vehicle's road ahead and surrounding environment image;

[0083] Based on the deployment orientation and corresponding relationship of each sensor on the vehicle, the target orientation of each vehicle's forward road image and surrounding environment image relative to the vehicle is determined, and each vehicle's forward road image and surrounding environment image is marked based on the target orientation, and the marked vehicle's forward road image and surrounding environment image are transmitted to the image processing center based on a preset transmission link.

[0084] The beneficial effect of the above technical solution is that it ensures the comprehensiveness and reliability of the vehicle's collection of images of the vehicle's forward road and surrounding environment in the autonomous driving state, and provides reliable image data for autonomous driving assistance.

[0085] In one embodiment, an artificial intelligence-based method for assisting autonomous driving in severe weather conditions labels collected images of a vehicle's path and surrounding environment and transmits them to an image processing center, including:

[0086] Establish independent communication links between each sensor and the image processing center, and transmit the vehicle's road ahead image and surrounding environment image collected by different sensors to the image processing center based on the independent communication links. The independent communication links mean that the communication links between each sensor and the image processing center are independent and do not affect each other;

[0087] Based on the image processing center, the vehicle's forward road images and surrounding environment images transmitted by different sensors are distinguished and cached.

[0088] The beneficial effect of the above technical solution is: ensuring that the vehicle's forward road images and surrounding environment images collected by different sensors are transmitted to the image processing center respectively, ensuring the stability and reliability of image transmission, and providing reliable image information for automatic driving assistance in severe weather.

[0089] In one embodiment, an artificial intelligence-based method for assisting an automated driving in severe weather conditions includes preprocessing an image of the vehicle's forward road and surrounding environment in step 1, including:

[0090] Obtaining the obtained vehicle forward road image and surrounding environment image, and inputting the vehicle forward road image and surrounding environment image into the filter for denoising;

[0091] Collect image datasets under different severe weather scenarios and generate labels corresponding to different severe weather conditions;

[0092] Annotate the corresponding image dataset based on the marked labels to obtain an annotated image dataset, and train a trainable model framework based on the annotated image dataset to obtain an inclement weather recognition model;

[0093] The denoised images of the vehicle's forward road and surrounding environment are input into the severe weather recognition model for identification to obtain the severe weather type the vehicle is currently in.

[0094] Determining interference objects in the vehicle's forward road image and the surrounding environment image based on the type of severe weather, and determining first morphological features of the interference objects;

[0095] Locking a target pixel of an interference object in an image of the vehicle's forward road and an image of the surrounding environment based on the first morphological feature, and determining a second morphological feature of the main object based on neighboring pixels of the target pixel;

[0096] Determining depth information of the vehicle's forward road image and the vehicle's surrounding environment image based on the first morphological feature and the second morphological feature, and performing layer separation on the vehicle's forward road image and the vehicle's surrounding environment image based on the depth information;

[0097] Based on the layer splitting result, morphological processing is performed on the target pixel of the interference object corresponding to the first morphological feature, where the morphological processing includes corrosion, expansion, opening operation and closing operation.

[0098] In this embodiment, the marking labels refer to marking symbols corresponding to different severe weather conditions.

[0099] In this embodiment, the labeled image dataset refers to a result obtained by labeling image data of a corresponding type using labeling tags.

[0100] In this embodiment, the trainable model framework is known in advance, for example, it can be a neural network model.

[0101] In this embodiment, the interference object refers to the impact of bad weather on the vehicle's driving condition, such as raindrops or wind and sand.

[0102] In this embodiment, the first morphological feature refers to the position and outline of the interference object.

[0103] In this embodiment, the target pixel refers to the specific area where the interference object is located in the vehicle's forward road image and the surrounding environment image, that is, the corresponding specific pixel point.

[0104] In this embodiment, the main object refers to obstacles and the like in the vehicle's forward road image and the surrounding environment image.

[0105] In this embodiment, the second morphological feature refers to information such as the outline and position of the main object.

[0106] In this embodiment, layer splitting refers to splitting the interference object and the main object, thereby removing the influence of the interference factors.

[0107] In this embodiment, the purpose of performing morphological processing on the target pixels of the interference object corresponding to the first morphological feature based on the layer segmentation result is to eliminate the interference object.

[0108] The beneficial effects of the above technical solution are: ensuring the accuracy and reliability of the preprocessing of the vehicle's forward road image and the surrounding environment image, and also providing convenience and protection for the vehicle's automatic driving assistance in bad weather.

[0109] In one embodiment, an artificial intelligence-based method for assisting automated driving in severe weather conditions includes determining driving characteristics of the vehicle based on preprocessing results in step 2, and determining a current automated driving assistance strategy for the vehicle based on the driving characteristics, including:

[0110] Object extraction is performed on the preprocessed images of the vehicle's forward road and surrounding environment to obtain the distribution of road obstacles in the vehicle's current forward direction and the road parameters of the vehicle's current travel. The road parameters include road width and road flatness, and road obstacles include pedestrians, road vehicles, and fixed, immovable objects.

[0111] Determine the relative position and relative distance between the current vehicle and the road obstacle based on the distribution of road obstacles, and use the relative position and relative distance as the first consideration parameters, while using the road parameters as the second consideration parameters;

[0112] Based on the first consideration parameter and the second consideration parameter, available automatic driving assistance strategies are matched from a predefined series of driving strategies. At the same time, the influence coefficient of severe weather on automatic driving is determined, and local parameters in the available automatic driving assistance strategies are adaptively corrected based on the influence coefficient to obtain the final automatic driving assistance strategy.

[0113] In this embodiment, a predefined series of driving strategies is known in advance, including corresponding automatic driving strategies in different situations.

[0114] In this embodiment, the impact coefficient refers to the impact of bad weather on autonomous driving, such as the impact of rainy days on anti-skid.

[0115] In this embodiment, adaptive correction refers to the final executable autonomous driving strategy obtained by adjusting some parameters of the selected available autonomous driving assistance strategy according to the impact of the current bad weather on autonomous driving. For example, if rainy days affect the friction between the vehicle and the ground, the vehicle's driving speed during autonomous driving will be adjusted.

[0116] The beneficial effect of the above technical solution is: ensuring the reliability of the final automatic driving assistance strategy, thereby ensuring the reliability and safety of the vehicle's automatic driving assistance.

[0117] In one embodiment, an artificial intelligence-based method for assisting an autonomous driving system in severe weather conditions, which obtains a final autonomous driving assistance strategy, includes:

[0118] The automated driving assistance strategy is split into task nodes to obtain a set of task nodes, and the specific data content corresponding to each task node is verified based on safe driving standards.

[0119] After the parameter value verification is passed, the automatic driving assistance strategy is authorized to take effect.

[0120] In this embodiment, task node splitting refers to splitting the automatic driving assistance strategy in order to obtain control parameters corresponding to different components in the vehicle, such as the power of the transmitter, the vehicle's driving speed, and the temperature inside the vehicle.

[0121] In this embodiment, the safe driving standard is known in advance and is used to characterize the speed and steering angle corresponding to safe driving of the vehicle under different conditions.

[0122] In this embodiment, parameter value verification refers to verifying the value of the specific data content corresponding to each task node according to the safe driving standard, in order to ensure that the safe driving requirements are met.

[0123] In this embodiment, the effective authorization refers to applying the obtained automatic driving assistance strategy after the verification is passed.

[0124] The beneficial effect of the above technical solution is: to verify the value of each task node of the automatic driving assistance strategy according to the safe driving standards, thereby ensuring the safety and reliability of the final automatic driving assistance strategy.

[0125] In one embodiment, an artificial intelligence-based method for assisting in severe weather autonomous driving, in step 3, assisting the vehicle in driving based on an autonomous driving assistance strategy, and dynamically updating the autonomous driving assistance strategy based on the vehicle's driving state, includes:

[0126] The automated driving assistance strategy is decomposed based on the vehicle operating components to obtain the control parameters of each vehicle operating component. The control parameters are then divided into stages based on the vehicle state change, which includes the improvement stage and the fully automated assisted driving stage. The vehicle state change is the degree of change in vehicle speed and steering amplitude.

[0127] Adjust the control parameters at different stages, and coordinately control the vehicle's operating components based on the parameter adjustment results at each stage to change the vehicle's state to assisted driving;

[0128] Based on the change results, the vehicle is controlled to perform trial assisted driving in the perfect driving stage, and the automatic driving assistance strategy is improved based on the trial assisted driving results. After a fixed mileage without accidents, the vehicle enters the fully automatic assisted driving stage;

[0129] At the same time, during the fully automatic assisted driving stage, the vehicle's driving status, the vehicle's forward road status and the surrounding environment status are monitored in real time, and the automatic driving assistance strategy is adjusted locally and dynamically in real time according to the vehicle's driving status, the vehicle's forward road status and the surrounding environment status until the assisted driving is completed.

[0130] In this embodiment, splitting the automatic driving assistance strategy based on the vehicle operation components refers to splitting the automatic driving assistance strategy according to the components of the vehicle, that is, the purpose is to obtain the specific control scheme corresponding to each vehicle operation component, that is, the control parameters, such as the specific execution parameters of the speed or the specific size of the traction, etc.

[0131] In this embodiment, trial assisted driving refers to preliminary autonomous driving control of the vehicle in the improvement stage, that is, the control parameters are effective within the minimum range, for example, the speed can be reduced to the minimum for autonomous driving.

[0132] In this embodiment, improving the automatic driving assistance strategy based on the trial assisted driving results refers to improving the imperfect parameters of the automatic driving assistance strategy during the trial assisted driving, such as adjusting the steering amplitude, etc.

[0133] The beneficial effects of the above technical solution are: ensuring the reliability of the vehicle's automatic driving assistance control in severe weather, and improving the safety factor of the vehicle's automatic driving assistance in severe weather.

[0134] In one embodiment, an artificial intelligence-based severe weather autonomous driving assistance system, such as Figure 3 Shown, including:

[0135] An image acquisition module is used to collect images of the vehicle's road ahead and surrounding environment in real time under severe weather conditions, and to pre-process the images of the vehicle's road ahead and surrounding environment;

[0136] A strategy formulation module, configured to determine the driving characteristics of the vehicle based on the preprocessing results, and determine the current automatic driving assistance strategy of the vehicle based on the driving characteristics;

[0137] The control module is used to assist the vehicle in driving based on the automatic driving assistance strategy and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

[0138] The beneficial effects of the above technical solution are: improving the preprocessing effect of the vehicle's forward road image and surrounding environment image in severe weather, thereby realizing accurate and effective determination of the vehicle's driving characteristics based on the preprocessing effect, and then realizing effective construction of the automatic driving assistance strategy based on the driving characteristics, and finally realizing assisted driving of the vehicle according to the automatic driving assistance strategy, and dynamically updating the automatic driving assistance strategy during the driving process, ensuring the reliability, accuracy and stability of automatic driving assistance in severe weather.

[0139] In one embodiment, an artificial intelligence-based severe weather autonomous driving assistance system, an image acquisition module, includes:

[0140] An image pre-capture unit is used to pre-capture images of the corresponding working area based on sensors deployed on the vehicle, and to perform spatial range recognition on the pre-captured images to obtain the image acquisition angles of each sensor;

[0141] The device configuration unit is used to determine the image capture range of each sensor under the image acquisition angle based on the spatial range recognition result, and when the image capture range does not meet the vehicle's automatic driving assistance requirements, perform multiple adaptive corrections on the image capture angle of each sensor until the image capture range meets the vehicle's automatic driving assistance requirements.

[0142] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An artificial intelligence-based automatic driving assistance method for severe weather conditions, characterized in that: include: Step 1: Real-time acquisition of vehicle road images and surrounding environment images in severe weather conditions, and pre-processing of the vehicle road images and surrounding environment images; Step 2: Determine the vehicle's driving characteristics based on the preprocessing results, and determine the vehicle's current automatic driving assistance strategy based on the driving characteristics; Step 3: Assist the vehicle in driving based on the automatic driving assistance strategy, and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

2. The artificial intelligence-based severe weather automatic driving assistance method according to claim 1, characterized in that: In step 1, before collecting images of the vehicle's forward road and surrounding environment in real time under severe weather conditions, the following steps are performed: Based on the sensors deployed on the vehicle, the corresponding working area is pre-collected, and the spatial range of the pre-collected images is identified to obtain the image acquisition angle of each sensor; Based on the spatial range recognition results, the image capturing range of each sensor under the image acquisition angle is determined. When the image capturing range does not meet the vehicle's automatic driving assistance requirements, the image capturing angle of each sensor is adaptively corrected multiple times until the image capturing range meets the vehicle's automatic driving assistance requirements.

3. The artificial intelligence-based severe weather automatic driving assistance method according to claim 1, characterized in that: In step 1, real-time images of the vehicle's forward path and surrounding environment in severe weather conditions are collected, including: Receive automatic driving assistance requests from users, perform self-checks on the vehicle's driving status, and enters the automatic driving state after passing the self-checks; After the self-test passes, each sensor is synchronously controlled in parallel, and based on the parallel control results, each sensor is controlled to collect images of the vehicle's forward road and surrounding environment in real time under severe weather conditions; The collected images of the vehicle's forward road and surrounding environment are marked and transmitted to the image processing center.

4. The artificial intelligence-based severe weather automatic driving assistance method according to claim 3, characterized in that: The collected images of the vehicle's forward road and surrounding environment are marked and transmitted to the image processing center, including: Establish independent communication links between each sensor and the image processing center, and transmit the vehicle's road ahead image and surrounding environment image collected by different sensors to the image processing center based on the independent communication links. The independent communication links mean that the communication links between each sensor and the image processing center are independent and do not affect each other; Based on the image processing center, the vehicle's forward road images and surrounding environment images transmitted by different sensors are distinguished and cached.

5. The artificial intelligence-based severe weather automatic driving assistance method according to claim 1, characterized in that: In step 1, the vehicle's forward road image and surrounding environment image are preprocessed, including: Obtaining the obtained vehicle forward road image and surrounding environment image, and inputting the vehicle forward road image and surrounding environment image into the filter for denoising; Collect image datasets under different severe weather scenarios and generate labels corresponding to different severe weather conditions; Annotate the corresponding image dataset based on the marked labels to obtain an annotated image dataset, and train a trainable model framework based on the annotated image dataset to obtain an inclement weather recognition model; The denoised images of the vehicle's forward road and surrounding environment are input into the severe weather recognition model for identification to obtain the severe weather type the vehicle is currently in. Determining interference objects in the vehicle's forward road image and the surrounding environment image based on the type of severe weather, and determining first morphological features of the interference objects; Locking a target pixel of an interference object in an image of the vehicle's forward road and an image of the surrounding environment based on the first morphological feature, and determining a second morphological feature of the main object based on neighboring pixels of the target pixel; Determining depth information of the vehicle's forward road image and the vehicle's surrounding environment image based on the first morphological feature and the second morphological feature, and performing layer separation on the vehicle's forward road image and the vehicle's surrounding environment image based on the depth information; Based on the layer splitting result, morphological processing is performed on the target pixel of the interference object corresponding to the first morphological feature, where the morphological processing includes corrosion, expansion, opening operation and closing operation.

6. The artificial intelligence-based severe weather automatic driving assistance method according to claim 1, characterized in that: In step 2, the driving characteristics of the vehicle are determined based on the preprocessing results, and the current automatic driving assistance strategy of the vehicle is determined based on the driving characteristics, including: Object extraction is performed on the preprocessed images of the vehicle's forward road and surrounding environment to obtain the distribution of road obstacles in the vehicle's current forward direction and the road parameters of the vehicle's current travel. The road parameters include road width and road flatness, and road obstacles include pedestrians, road vehicles, and fixed, immovable objects. Determine the relative position and relative distance between the current vehicle and the road obstacle based on the distribution of road obstacles, and use the relative position and relative distance as the first consideration parameters, while using the road parameters as the second consideration parameters; Based on the first consideration parameter and the second consideration parameter, available automatic driving assistance strategies are matched from a predefined series of driving strategies. At the same time, the influence coefficient of severe weather on automatic driving is determined, and local parameters in the available automatic driving assistance strategies are adaptively corrected based on the influence coefficient to obtain the final automatic driving assistance strategy.

7. The artificial intelligence-based severe weather automatic driving assistance method according to claim 6, characterized in that: The final automated driving assistance strategy is obtained, including: The automated driving assistance strategy is split into task nodes to obtain a set of task nodes, and the specific data content corresponding to each task node is verified based on safe driving standards. After the parameter value verification is passed, the automatic driving assistance strategy is authorized to take effect.

8. The artificial intelligence-based severe weather automatic driving assistance method according to claim 1, characterized in that: In step 3, the vehicle is assisted in driving based on the automatic driving assistance strategy, and the automatic driving assistance strategy is dynamically updated based on the vehicle's driving status, including: The automated driving assistance strategy is decomposed based on the vehicle operating components to obtain the control parameters of each vehicle operating component. The control parameters are then divided into stages based on the vehicle state change, which includes the improvement stage and the fully automated assisted driving stage. The vehicle state change is the degree of change in vehicle speed and steering amplitude. Adjust the control parameters at different stages, and coordinately control the vehicle's operating components based on the parameter adjustment results at each stage to change the vehicle's state to assisted driving; Based on the change results, the vehicle is controlled to perform trial assisted driving in the perfect driving phase, and the automatic driving assistance strategy is improved based on the trial assisted driving results. After a fixed mileage without accidents, the vehicle enters the fully automatic assisted driving phase; At the same time, during the fully automatic assisted driving stage, the vehicle's driving status, the vehicle's forward road status and the surrounding environment status are monitored in real time, and the automatic driving assistance strategy is adjusted locally and dynamically in real time according to the vehicle's driving status, the vehicle's forward road status and the surrounding environment status until the assisted driving is completed.

9. An artificial intelligence-based bad weather automatic driving assistance system, characterized in that: include: An image acquisition module is used to collect images of the vehicle's road ahead and surrounding environment in real time under severe weather conditions, and to pre-process the images of the vehicle's road ahead and surrounding environment; A strategy formulation module, configured to determine the driving characteristics of the vehicle based on the preprocessing results, and determine the current automatic driving assistance strategy of the vehicle based on the driving characteristics; The control module is used to assist the vehicle in driving based on the automatic driving assistance strategy and dynamically update the automatic driving assistance strategy based on the vehicle's driving status.

10. The artificial intelligence-based severe weather automatic driving assistance system according to claim 9, characterized in that: Image acquisition module, including: An image pre-capture unit is used to pre-capture images of the corresponding working area based on sensors deployed on the vehicle, and to perform spatial range recognition on the pre-captured images to obtain the image acquisition angles of each sensor; The device configuration unit is used to determine the image capture range of each sensor under the image acquisition angle based on the spatial range recognition result, and when the image capture range does not meet the vehicle's automatic driving assistance requirements, perform multiple adaptive corrections on the image capture angle of each sensor until the image capture range meets the vehicle's automatic driving assistance requirements.

Images