Vehicle sensor window cleaning method and device and vehicle system

Abstract

The invention provides a window cleaning method and device for a sensor of a vehicle and a vehicle system. According to the method, a smudginess judgment result and a current smudginess type are obtained by processing an image collected by a shooting device, a point cloud collected by a laser radar, weather information and current rainfall, the smudginess judgment result represents whether a window of a sensor is smudginess or not, and the current smudginess type is sludge smudginess or rainwater smudginess; and under the condition that the dirt judgment result represents that the window of the sensor is dirty, the first processing mode or the second processing mode is executed, so that compared with an existing scheme, the window of the sensor can be cleaned more accurately, the detection result of the sensor is more fit with the actual working condition, and the detection efficiency is improved. Therefore, the problems that the sight of a driver is shielded and serious potential safety hazards are easily caused due to the fact that the smudginess of the window of the sensor of the vehicle is recognized by using a single image in an existing scheme and the smudginess judgment result is easily inconsistent with the actual situation are solved.

Description

Technical Field

[0001] The present application relates to the technical field of vehicle control. Specifically, it relates to a method, device, and vehicle system for cleaning the window of a vehicle's sensor. Background Art

[0002] Autonomous vehicles are equipped with a large number of optical sensors, such as cameras and lidar, to achieve perception of the external environment. These optical sensors have high requirements for the cleanliness of their window surfaces. If the window surfaces of the optical sensors on an autonomous vehicle are covered by pollutants such as rainwater and mud during driving, they will be unable to normally perceive the real external environment information, further resulting in the inability to make correct judgments and execution strategies, causing serious safety hazards.

[0003] Existing solutions use single-image recognition to detect dirt on the vehicle's window, which easily leads to a mismatch between the dirt determination result and the actual situation, further causing an obstructive effect on the driver's line of sight and serious safety hazards. Summary of the Invention

[0004] The main purpose of the present application is to provide a method, device, and vehicle system for cleaning the window of a vehicle's sensor, so as to at least solve the problem that existing solutions use single-image recognition to detect dirt on the vehicle's window, which easily leads to a mismatch between the dirt determination result and the actual situation, further causing an obstructive effect on the driver's line of sight and serious safety hazards.

[0005] To achieve the above object, according to one aspect of the present application, a method for cleaning the window of a vehicle's sensor is provided. The method includes:

[0006] Obtain an image collected by a photographing device, point cloud collected by a lidar, weather information, and current rainfall. The current rainfall is the rainfall at the current moment collected by a rainfall sensor. The weather information includes at least a rainfall monitoring result and a wind speed. The rainfall monitoring result indicates whether it is raining in the environment where the vehicle is located;

[0007] Process the image collected by the photographing device, the point cloud collected by the lidar, the weather information, and the current rainfall to obtain a dirt determination result and the current dirt type. The dirt determination result indicates whether dirt appears on the window of the sensor. The current dirt type is sludge-based dirt or rainwater-based dirt;

[0008] When the dirt determination result indicates that the window of the sensor is dirty, a first processing method or a second processing method is executed. The first processing method represents controlling the wiper to clean the window, and the second processing method represents cleaning the window by means of spraying water and jetting air, or only jetting air. Wherein, when the second processing method is executed, according to the current dirt type, the window is cleaned by means of spraying water and jetting air, or only jetting air.

[0009] Optionally, executing the first processing method or the second processing method includes: when the installation position of the target sensor is the windshield of the vehicle, executing the first processing method, where the target sensor represents the sensor whose window is dirty; when the installation position of the target sensor is other positions in the vehicle except the windshield, executing the second processing method.

[0010] Optionally, according to the current dirt type, cleaning the window by means of spraying water and jetting air, or only jetting air includes: when it is determined that the current dirt type is sludge-like dirt, first use the method of spraying water to wash it away, and then use the method of jetting air to clean the remaining water droplets; when it is determined that the current dirt type is rain-like dirt, use the high-pressure gas generated by a blower or an air compressor to clean the dirty window.

[0011] Optionally, processing the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall amount to obtain a dirt determination result and the current dirt type includes: using a deep learning model to process the images collected by the imaging device to obtain a first dirt degree and a first dirt type determination coefficient; using the method of cross-validation to process the point clouds collected by the lidar to obtain a second dirt degree and a second dirt type determination coefficient; processing the weather information in the form of a mapping relationship to obtain a third dirt degree and a third dirt type determination coefficient; processing the current rainfall amount in the form of a mapping relationship to obtain a fourth dirt degree and a fourth dirt type determination coefficient; performing a weighted summation process on the first dirt degree, the second dirt degree, the third dirt degree, and the fourth dirt degree to obtain a final dirt result judgment value, and performing a weighted summation process on the first dirt type determination coefficient, the second dirt type determination coefficient, the third dirt type determination coefficient, and the fourth dirt type determination coefficient to obtain a final dirt type judgment value; determining the dirt determination result according to the magnitude of the final dirt result judgment value, and determining the current dirt type according to the magnitude of the final dirt type judgment value.

[0012] Optionally, perform a weighted sum processing on the first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt to obtain a final dirt result judgment value, including:

[0013] Assign different weight values to the first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt;

[0014] Based on the assigned weight values, perform a weighted sum processing on the first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt to obtain the final dirt result judgment value.

[0015] Optionally, assigning different weight values to the first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt includes: determining a first preset weight value corresponding to the first degree of dirt, a second preset weight value corresponding to the second degree of dirt, a third preset weight value corresponding to the third degree of dirt, and a fourth preset weight value corresponding to the fourth degree of dirt in a mapping relationship manner.

[0016] Optionally, the method further includes: when receiving a cleaning instruction, performing the first processing method or the second processing method based on the cleaning instruction.

[0017] Optionally, the method further includes: obtaining the temperature of each actuator and the pressure and flow rate of each water spraying system to obtain monitoring data; generating a prompt message to prompt the magnitude of the monitoring data or generating a warning message to prompt that the monitoring data exceeds the standard according to the magnitude of the monitoring data.

[0018] According to another aspect of the present application, there is provided a window cleaning device for a sensor of a vehicle, and the device includes:

[0019] An acquisition unit, configured to acquire images collected by a photographing device, point clouds collected by a lidar, weather information, and current rainfall, where the current rainfall is the rainfall at the current moment collected by a rainfall sensor, and the weather information includes at least a rainfall monitoring result and a wind speed, and the rainfall monitoring result indicates whether it is raining in the environment where the vehicle is located;

[0020] A first processing unit, configured to process the images collected by the photographing device, the point clouds collected by the lidar, the weather information, and the current rainfall to obtain a dirt judgment result and a current dirt type, where the dirt judgment result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt;

[0021] A second processing unit, configured to execute a first processing mode or a second processing mode when the dirt determination result indicates that the window of the sensor is dirty. The first processing mode represents controlling a wiper to clean the window, and the second processing mode represents cleaning the window by using a water spraying and air jetting method, or only using an air jetting method. Wherein, when executing the second processing mode, according to the current dirt type, the window is cleaned by using a water spraying and air jetting method, or only using an air jetting method.

[0022] According to another aspect of the present application, there is provided a vehicle system, including: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include those for executing any one of the methods.

[0023] Applying the technical solution of the present application, by processing the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall, a dirt determination result and the current dirt type are obtained. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt; when the dirt determination result indicates that the window of the sensor is dirty, the first processing mode or the second processing mode is executed, so that the present application can clean the window of the sensor more accurately compared with the existing solutions, making the detection result of the sensor more consistent with the actual working conditions, and further solving the problem that the existing solutions use single image recognition to identify the dirt on the window of the vehicle sensor, which easily makes the dirt determination result inconsistent with the actual situation, further causing an occlusion effect on the driver's line of sight and easily causing serious safety hazards. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The specification drawings forming a part of the present application are used to provide a further understanding of the present application. The schematic embodiments and descriptions thereof of the present application are used to explain the present application and do not constitute an improper limitation to the present application. In the drawings:

[0025] Figure 1 It shows a schematic flowchart of a method for cleaning a window of a vehicle sensor provided according to an embodiment of the present application;

[0026] Figure 2 It shows a schematic diagram of a method for cleaning a window of a vehicle sensor provided according to an embodiment of the present application;

[0027] Figure 3 It shows a structural block diagram of a device for cleaning a window of a vehicle sensor provided according to an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts shall fall within the protection scope of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances for the embodiments of the present application described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0031] As introduced in the background art, in order for an autonomous vehicle to perceive the external environment, a large number of optical sensors are installed, such as cameras, lidar, etc. These optical sensors have relatively high requirements for the cleanliness of their window surfaces. If the window surfaces of the optical sensors of an autonomous vehicle are covered by pollutants such as rainwater and mud during driving, it will be unable to normally perceive the real external environment information, further resulting in the inability to make correct judgments and execution strategies, causing serious safety hazards. To solve the problem that the existing solution uses single image recognition to detect the dirt on the vehicle window, which easily makes the dirt determination result inconsistent with the actual situation, further causing an occlusion effect on the driver's line of sight and easily causing serious safety hazards, the embodiments of the present application provide a method, device and vehicle system for cleaning the window of a vehicle sensor.

[0032] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0033] In this embodiment, a method for cleaning the window of a vehicle sensor is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0034] Figure 1 It is a schematic flowchart of a method for cleaning the window of a vehicle sensor according to an embodiment of the present application. As Figure 1 shown, the method includes the following steps:

[0035] Step S101, obtain the image collected by the imaging device, the point cloud collected by the lidar, the weather information, and the current rainfall amount. The current rainfall amount is the rainfall amount at the current moment collected by the rainfall sensor. The weather information includes at least the rainfall monitoring result and the wind speed. The rainfall monitoring result indicates whether it is raining in the environment where the vehicle is located;

[0036] The weather information may further include whether the weather is sunny.

[0037] Specifically, as Figure 2The window cleaning architecture of the sensors of the vehicle shown includes a camera imaging quality judgment module, a lidar point cloud quality judgment module, a weather information module, a rain sensor module, and a manual trigger button module. It judges the current dirty scene, clarifies the current dirty type and degree of dirt, selects the corresponding cleaning means and parameters, and transmits the corresponding cleaning instructions to the execution / cleaning link. For the images collected by the imaging device: The camera imaging quality judgment module uses photos with manually annotated camera imaging quality results and an algorithm model obtained through deep learning training. It can process each frame of the images of each camera of the autonomous vehicle and judge whether there is dirt on the frame. If there is dirt, it further outputs information in two dimensions: the type of dirt and the degree of dirt. For the point cloud collected by the lidar: The lidar point cloud quality judgment module statistically analyzes the number of points and the degree of uniform distribution of the point cloud received by the lidar, and cross-verifies with the detection data of sensors such as cameras and millimeter-wave radars to determine whether there is dirt / obstruction on the surface of the lidar. In a typical traffic scene, there is an interval range for the number of point clouds received by the lidar. When the number of points of the point cloud received by the lidar continuously and significantly is less than the lower limit of this interval, and other sensors (such as cameras, millimeter-wave radars, etc.) can detect obstacles / traffic participants in a certain azimuth around, and there is a lack of point cloud at the corresponding position of the lidar, it can be considered that there is dirt / obstruction at the corresponding position of the lidar. For the weather information: The weather information module constructs a real-time monitoring and forecasting system for meteorological information for autonomous vehicle fleets and individual vehicles that can be accurate to 1 km in spatial granularity and 5 minutes or even shorter in time granularity. This system will push the weather information to the autonomous vehicle fleet and individual vehicles in real time. For the current rainfall: The rain sensor module judges whether there is rainfall and the size of the rainfall through the installed rain sensor.

[0038] The upper computer includes a camera imaging quality judgment module, a lidar point cloud quality judgment module, a weather information module, a rain sensor module, and a manual trigger button module. The lower computer includes the controller of the cleaning equipment (composed of a water spraying device, an air jetting device, and a windshield wiper). The monitoring equipment includes sensors such as pressure, flow rate, and temperature. The monitoring equipment communicates with the lower computer through a serial port such as RS485 or Lin communication protocols, and the upper computer and the lower computer communicate with each other through the CAN method.

[0039] Step S102, process the images collected by the above imaging device, the point cloud collected by the above lidar, the above weather information, and the above current rainfall to obtain a dirt judgment result and the current dirt type. The dirt judgment result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt;

[0040] The above-mentioned deep learning model includes a convolutional neural network, a recurrent neural network, an attention mechanism model, and a point cloud processing network.

[0041] Specifically, the deep learning model can improve accuracy through a large amount of training data, effectively identify and analyze the images, point cloud images, weather information, and rainfall data collected by the imaging device, so as to obtain a more accurate evaluation result of the dirtiness degree. The deep learning model can also realize the automatic processing and analysis of multiple data sources, save labor costs and time costs, improve work efficiency. The deep learning model can also quickly process a large amount of data and update the evaluation result of the dirtiness degree in real time, and provide decision support in a timely manner.

[0042] In an embodiment of the present application, the images collected by the above-mentioned imaging device, the point cloud collected by the above-mentioned lidar, the above-mentioned weather information, and the above-mentioned current rainfall are processed to obtain a dirtiness judgment result and the current dirtiness type, including: using the deep learning model to process the images collected by the above-mentioned imaging device to obtain the first dirtiness degree and the first dirtiness type determination coefficient; using the cross-validation method to process the point cloud collected by the above-mentioned lidar to obtain the second dirtiness degree and the second dirtiness type determination coefficient; processing the above-mentioned weather information in the form of a mapping relationship to obtain the third dirtiness degree and the third dirtiness type determination coefficient; processing the above-mentioned current rainfall in the form of a mapping relationship to obtain the fourth dirtiness degree and the fourth dirtiness type determination coefficient; performing weighted summation processing on the above-mentioned first dirtiness degree, the above-mentioned second dirtiness degree, the above-mentioned third dirtiness degree, and the above-mentioned fourth dirtiness degree to obtain a final dirtiness result judgment value, and performing weighted summation processing on the above-mentioned first dirtiness type determination coefficient, the above-mentioned second dirtiness type determination coefficient, the above-mentioned third dirtiness type determination coefficient, and the above-mentioned fourth dirtiness type determination coefficient to obtain a final dirtiness type judgment value; determining the above-mentioned dirtiness judgment result according to the magnitude of the above-mentioned final dirtiness result judgment value, and determining the above-mentioned current dirtiness type according to the magnitude of the above-mentioned final dirtiness type judgment value.

[0043] Specifically, similar to the dirtiness degree and dirt type determination coefficient, it is also represented in the form of a numerical value. The image collected by the camera is input into a pre-trained deep learning model, which can identify and quantify the dirtiness degree and type, and the output is the first dirtiness degree and the first dirt type determination coefficient. One or more algorithms are used to cross-validate the lidar point cloud data, and the dirtiness degree of the lidar surface is judged through indicators such as the decrease in point cloud density and uneven distribution, and the output is the second dirtiness degree and the second dirt type determination coefficient. In the way of using a mapping relationship, the corresponding third dirtiness degree and third dirt type determination coefficient are mapped according to the weather information. The current rainfall is the same as the weather information and will not be elaborated here. By setting corresponding thresholds, the above-mentioned dirtiness judgment result is determined according to the size of the numerical value of the final dirtiness result, and the current dirt type is determined according to the size of the numerical value of the final dirt type judgment, which will not be elaborated here.

[0044] Combining the analysis of images by the deep learning model, the cross-validation of point cloud data, the mapping of weather information, and the real-time monitoring of rainfall, the dirtiness degree and type are comprehensively judged from multiple dimensions. This method of multi-source data fusion can effectively improve the accuracy and reliability of judgment, and reduce the possibility of misjudgment and missed judgment. The introduction of weather information and rainfall monitoring enables the system to make a quick response according to the real-time environmental conditions, predict potential dirtiness situations, especially for the upcoming rain-type dirtiness, and can prepare and respond in advance, enhancing the environmental adaptability and early warning ability of the system. By weighting and summing the dirtiness determination coefficients from different sources through algorithms, the system can automatically make the optimal cleaning strategy selection according to the specific weights of each dirt type and degree without manual intervention, improving the intelligent level of the cleaning system. Through the effective judgment of the dirtiness degree and type, the system can intelligently select the cleaning method and intensity, avoid the situations of over-cleaning or insufficient cleaning, rationally utilize the cleaning resources, and reduce the operating cost of the cleaning system. The real-time monitoring and fast processing mechanism enable the system to quickly respond to dirtiness changes, start the cleaning process in time, reduce the perception delay, and ensure the real-time and accurate perception of the external environment information by the autonomous driving vehicle. Through real-time monitoring and intelligent analysis, the system can not only judge the dirtiness situation, but also diagnose the operating state of the cleaning system, discover potential faults in time, and perform preventive maintenance, reducing the sensor performance degradation caused by system failures. The accurate judgment of the dirt type and degree, combined with an efficient cleaning plan, ensures that the sensor can maintain a good working state at any time. For an autonomous driving vehicle, this directly relates to the vehicle's perception ability of the external environment and driving safety, significantly improving the safety of the overall system operation.

[0045] Among them, weighted summation processing is performed on the above-mentioned first dirtiness degree, second dirtiness degree, third dirtiness degree, and fourth dirtiness degree to obtain a final dirtiness result judgment numerical value, including:

[0046] Assign different weights to the above first degree of dirtiness, the above second degree of dirtiness, the above third degree of dirtiness, and the above fourth degree of dirtiness;

[0047] Based on the assigned weights, perform a weighted summation process on the above first degree of dirtiness, the above second degree of dirtiness, the above third degree of dirtiness, and the above fourth degree of dirtiness to obtain the above final dirtiness result judgment value.

[0048] Among them, assigning different weights to the above first degree of dirtiness, the above second degree of dirtiness, the above third degree of dirtiness, and the above fourth degree of dirtiness includes: determining a first preset weight corresponding to the above first degree of dirtiness, a second preset weight corresponding to the above second degree of dirtiness, a third preset weight corresponding to the above third degree of dirtiness, and a fourth preset weight corresponding to the above fourth degree of dirtiness in the form of a mapping relationship.

[0049] Specifically, different preset weights can be assigned to different degrees of dirtiness in the form of a mapping relationship. Different sensors and sources of dirt have different degrees of influence on autonomous driving vehicles. Assigning weights can more accurately reflect the influence of various dirt information on the overall perception system, thereby making a cleaning decision more in line with the actual situation. By adjusting the weights, the system can dynamically optimize the cleaning strategy according to changes in different environments and conditions. For example, in rainy days, the weight of rainfall information can be increased to pay more attention to the influence of rain on sensors. Even if a certain sensor or information source fails, the mechanism of assigning weights and performing weighted summation can still ensure that the system makes a reasonable cleaning judgment based on other valid information, enhancing the stability and reliability of the system. According to the specific degree and type of dirt of each sensor, the system can intelligently decide the allocation of cleaning resources, such as the amount of cleaning agent used, water pressure, and air flow size, to avoid waste of resources and improve cleaning efficiency. Through refined management, the system can respond more quickly and accurately to different dirt situations, reduce the impact on sensor performance, and thus improve the overall operation efficiency of autonomous driving vehicles and the sense of security and comfort of users. Under different operating environments and time conditions, the influence of dirt on sensors may vary greatly. The dynamic weight assignment mechanism enables the system to better adapt to various complex scenarios and ensure that effective cleaning decisions can be made in any situation.

[0050] Step S103, when the above-mentioned dirt determination result indicates that the window of the above-mentioned sensor is dirty, execute the first processing method or the second processing method. The above-mentioned first processing method represents controlling the wiper to clean the above-mentioned window, and the above-mentioned second processing method represents cleaning the above-mentioned window by means of spraying water and jetting air, or only jetting air to clean the above-mentioned window. Among them, when executing the above-mentioned second processing method, according to the above-mentioned current dirt type, clean the above-mentioned window by means of spraying water and jetting air, or only jetting air to clean the above-mentioned window.

[0051] In the above steps, by processing the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall, a dirt determination result and the current dirt type are obtained. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt; when the dirt determination result indicates that the window of the sensor is dirty, execute the first processing method or the second processing method, so that the present application can clean the window of the sensor more accurately than the existing solution, making the detection result of the sensor more consistent with the actual working conditions, and thus solving the problem that the existing solution uses single image recognition to identify the dirt on the window of the vehicle's sensor, which easily makes the dirt determination result inconsistent with the actual situation, thereby causing an occlusion effect on the driver's line of sight and easily causing serious safety hazards.

[0052] Among them, the execution of the first processing method or the second processing method in step S103 includes: when the installation position of the target sensor is the windshield of the above-mentioned vehicle, execute the above-mentioned first processing method, and the above-mentioned target sensor represents the sensor with a dirty window; when the installation position of the above-mentioned target sensor is other positions in the above-mentioned vehicle except the windshield, execute the above-mentioned second processing method.

[0053] The sensors on the windshield can be cleaned by using the existing windshield wiper system of the vehicle. This not only reduces the use of additional water spraying and air jetting devices but also effectively utilizes the existing vehicle hardware resources, reducing the complexity and cost of the cleaning system. Cleaning the sensors on the windshield by using the windshield wiper can fully consider the specific requirements of the windshield during driving, such as removing raindrops, dust, etc., to ensure good decontamination effects. For sensors at other positions of the vehicle, the method of water spraying and / or air jetting can be used to directly and efficiently clean the specific environment where the target sensor is located. For sensors at non-windshield positions, directly spraying water may leave water droplets, which may affect the optical performance of the sensors. The strategy of spraying water first and then jetting air can effectively remove the water droplets, avoiding secondary contamination and ensuring that the sensor surface is dry and clean. Sensors at different positions of the vehicle may face different forms of contamination and cleaning requirements. For example, sensors behind the windshield may be mainly affected by raindrops and dust, while sensors on the roof or at the rear of the vehicle may be more easily contaminated by mud. Through intelligent selection of the cleaning method, this design can adapt to various complex environments and ensure the cleanliness of all sensors. The selection and execution of this processing method are completely automatically judged and controlled by the controller according to the sensor position, reducing the need for manual intervention, improving the automation and intelligence level of the cleaning process, and ensuring the timeliness and accuracy of cleaning. The cleaning of the sensors on the windshield is completed by the windshield wiper system without causing additional interference to the vehicle driving, while the cleaning of sensors at other positions is quickly completed during vehicle driving or when the vehicle is stationary without affecting the normal operation of autonomous driving.

[0054] Among them, according to the above-mentioned current dirt type, the above-mentioned window is cleaned by using the method of water spraying and air jetting, or only by using the method of air jetting, including: in the case where it is determined that the above-mentioned current dirt type is the sludge-based dirt, first use the method of water spraying to wash and remove it, and then use the method of air jetting to clean up the remaining water droplets; in the case where it is determined that the above-mentioned current dirt type is the rainwater-based dirt, use the high-pressure gas generated by a blower or an air compressor to clean the above-mentioned window where dirt appears.

[0055] Specifically, for sludge-like dirt, which is usually viscous and difficult to remove only by airflow, water spraying can effectively remove the mud first, and the subsequent air jet step can quickly remove water marks, avoid water stains on the surface of the optical sensor, and improve cleaning efficiency. For rainwater-like dirt, air jet cleaning can be used directly to quickly evaporate or blow away water droplets, avoid unnecessary waste of water resources, and quickly restore sensor performance. Taking the right cleaning method is crucial to protecting the optical surface of the sensor. Water spray cleaning is suitable for removing sticky dirt, while air jet is suitable for removing liquid residues. This strategy ensures that the sensor surface is not damaged while maintaining the best optical performance. By intelligently judging and selecting the cleaning method, unnecessary cleaning steps are avoided, the consumption of cleaning materials (such as water) is reduced, and the maintenance cost and energy consumption of the cleaning system are reduced. At the same time, efficient cleaning reduces the frequency of failures caused by sensor dirt, and indirectly reduces the maintenance and repair costs of the vehicle. This strategy can flexibly adjust the cleaning method according to the actual pollution situation. Whether it is muddy roads or rainy days, it can quickly adapt and effectively clean, ensuring the consistent performance of the sensor in different environments. Timely and correct cleaning methods can quickly restore the normal working state of the sensor and reduce the perception errors caused by dirt, thereby improving the safety and reliability of autonomous vehicles, especially in severe weather conditions.

[0056] In one embodiment of the present application, the method further includes: upon receiving a cleaning instruction, executing the first processing method or the second processing method based on the cleaning instruction.

[0057] Specifically, for the manual trigger button module, physical / virtual buttons are provided on the autonomous driving vehicle side and the remote monitoring side, so that people on the vehicle or remote monitoring personnel can trigger the corresponding cleaning function by pressing the buttons when necessary to send cleaning instructions.

[0058] In one embodiment of the present application, the above method also includes: obtaining the temperature of each actuator and the pressure and flow of each water spray system to obtain monitoring data; based on the size of the above monitoring data, generating prompt information to prompt the size of the above monitoring data or generating early warning information to prompt that the above monitoring data exceeds the standard.

[0059] Specifically, for monitoring equipment, the pressure and flow sensors arranged in the main and branch lines of the jet / water spray system are used to monitor and record the pressure and flow of the main and important branches in real time; in addition, temperature sensors are installed for the main actuators (air compressor / blower / water pump, etc.) of the cleaning link / module to monitor and record the temperature rise of these actuators in real time. When the pressure, flow or temperature is abnormal, the corresponding fault diagnosis will be made, and the personnel on the vehicle, remote monitoring personnel and vehicle maintenance personnel will be reminded to carry out maintenance through screen pop-ups, voice, text messages, etc.

[0060] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0061] The embodiment of the present application also provides a window cleaning device for a vehicle sensor. It should be noted that the window cleaning device for the vehicle sensor in the embodiment of the present application can be used to execute the window cleaning method for the vehicle sensor provided by the embodiment of the present application. This device is used to implement the above embodiments and preferred implementation manners, and those that have been described will not be repeated. As used hereinafter, the term "module" can be a combination of software and / or hardware that realizes a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0062] The following introduces the window cleaning device for a vehicle sensor provided by the embodiment of the present application.

[0063] Figure 3 It is a structural block diagram of a window cleaning device for a vehicle sensor provided according to an embodiment of the present application. As Figure 3 shown, the device includes:

[0064] An acquisition unit 31, configured to acquire images collected by a photographing device, point clouds collected by a lidar, weather information, and current rainfall. The above current rainfall is the rainfall at the current moment collected by a rainfall sensor. The above weather information includes at least a rainfall monitoring result and a wind speed. The above rainfall monitoring result characterizes whether it is raining in the environment where the vehicle is located;

[0065] A first processing unit 32, configured to process the images collected by the photographing device, the point clouds collected by the lidar, the weather information, and the current rainfall to obtain a dirt determination result and a current dirt type. The above dirt determination result characterizes whether dirt appears on the window of the sensor. The above current dirt type is sludge-based dirt or rainwater-based dirt;

[0066] A second processing unit 33, configured to execute a first processing method or a second processing method when the above dirt determination result characterizes that dirt appears on the window of the sensor. The above first processing method characterizes controlling the windshield wiper to clean the above window. The above second processing method characterizes cleaning the above window by means of spraying water and jetting air, or only jetting air. Wherein, when executing the above second processing method, according to the above current dirt type, clean the above window by means of spraying water and jetting air, or only jetting air.

[0067] In the above device, by processing the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall, a dirt determination result and the current dirt type are obtained. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt. When the dirt determination result indicates that the window of the sensor is dirty, the first processing method or the second processing method is executed, so that the present application can clean the window of the sensor more accurately than the existing solutions, making the detection result of the sensor more consistent with the actual working conditions. Furthermore, the problem of the existing solutions using single image recognition to identify the dirt on the window of the vehicle sensor, which easily leads to a mismatch between the dirt determination result and the actual situation, thereby causing an occlusion effect on the driver's line of sight and easily posing a serious safety hazard, is solved.

[0068] In an embodiment of the present application, the second processing unit includes a first processing module and a second processing module. The first processing module is configured to execute the above first processing method when the installation position of the target sensor is the windshield of the above vehicle, and the above target sensor represents the sensor with a dirty window. The second processing module is configured to execute the above second processing method when the installation position of the above target sensor is other positions in the above vehicle except the windshield.

[0069] In an embodiment of the present application, the second processing unit includes a third processing module and a fourth processing module. The third processing module is configured to first use a water spraying method to wash and remove the dirt when it is determined that the current dirt type is the sludge-based dirt, and then use a gas jetting method to clean the remaining water droplets. The fourth processing module is configured to use the high-pressure gas generated by a blower or an air compressor to clean the dirty window when it is determined that the current dirt type is the rainwater-based dirt.

[0070] In an embodiment of the present application, the first processing unit includes a fifth processing module, a sixth processing module, a seventh processing module, an eighth processing module, a ninth processing module, and a tenth processing module. The fifth processing module is configured to process the images collected by the above-mentioned imaging device using a deep learning model to obtain a first degree of dirt and a first determination coefficient for the type of dirt. The sixth processing module is configured to process the point cloud collected by the above-mentioned lidar using a cross-validation method to obtain a second degree of dirt and a second determination coefficient for the type of dirt. The seventh processing module is configured to process the above-mentioned weather information in the form of a mapping relationship to obtain a third degree of dirt and a third determination coefficient for the type of dirt. The eighth processing module is configured to process the above-mentioned current rainfall in the form of a mapping relationship to obtain a fourth degree of dirt and a fourth determination coefficient for the type of dirt. The ninth processing module is configured to perform a weighted summation process on the above-mentioned first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt to obtain a final determination value for the dirt result, and perform a weighted summation process on the above-mentioned first determination coefficient for the type of dirt, the second determination coefficient for the type of dirt, the third determination coefficient for the type of dirt, and the fourth determination coefficient for the type of dirt to obtain a final determination value for the type of dirt. The tenth processing module is configured to determine the above-mentioned dirt determination result based on the magnitude of the above-mentioned final determination value for the dirt result, and determine the above-mentioned current type of dirt based on the magnitude of the above-mentioned final determination value for the type of dirt.

[0071] In an embodiment of the present application, the ninth processing module includes a first processing sub-module and a second processing sub-module. The first processing sub-module is configured to assign different weight values to the above-mentioned first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt. The second processing sub-module is configured to perform a weighted summation process on the above-mentioned first degree of dirt, the second degree of dirt, the third degree of dirt, and the fourth degree of dirt based on the assigned weight values to obtain the above-mentioned final determination value for the dirt result.

[0072] In an embodiment of the present application, the first processing sub-module includes a third processing sub-module, which is configured to determine a first preset weight value corresponding to the above-mentioned first degree of dirt, a second preset weight value corresponding to the above-mentioned second degree of dirt, a third preset weight value corresponding to the above-mentioned third degree of dirt, and a fourth preset weight value corresponding to the above-mentioned fourth degree of dirt in the form of a mapping relationship.

[0073] In an embodiment of the present application, the above-mentioned device further includes a third processing unit, which is configured to execute the above-mentioned first processing method or the above-mentioned second processing method based on the above-mentioned cleaning instruction when the cleaning instruction is received.

[0074] In one embodiment of the present application, the above device further includes a fourth processing unit and a fifth processing unit: the fourth processing unit is configured to obtain the temperatures of the actuators and the pressures and flows of the water spraying systems to obtain monitoring data; the fifth processing unit is configured to generate a prompt message to prompt the size of the above monitoring data or generate a warning message to prompt that the above monitoring data exceeds the standard according to the size of the above monitoring data.

[0075] The window cleaning device of the sensor of the above vehicle includes a processor and a memory. The above acquisition unit, the first processing unit, the second processing unit, etc. are all stored in the memory as program units, and the corresponding functions are implemented by the processor executing the above program units stored in the memory. The above modules are all located in the same processor; alternatively, the above modules are respectively located in different processors in any combination form.

[0076] The processor contains a kernel, and the kernel is used to retrieve the corresponding program units from the memory. One or more kernels can be set. By adjusting the kernel parameters, the problem that the existing solution uses a single image to identify the dirt on the window of the vehicle, which easily makes the dirt determination result inconsistent with the actual situation, further causes an occlusion effect on the driver's line of sight and easily causes serious safety hazards can be solved.

[0077] The memory may include non-permanent memory in a computer-readable medium, forms such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM), and the memory includes at least one storage chip.

[0078] The embodiment of the present invention provides a computer-readable storage medium. The above computer-readable storage medium includes a stored program. When the above program runs, it controls the device where the above computer-readable storage medium is located to execute the window cleaning method of the sensor of the above vehicle.

[0079] The embodiment of the present invention provides a processor. The above processor is used to run a program. When the above program runs, it executes the window cleaning method of the sensor of the above vehicle.

[0080] An embodiment of the present invention provides a device, which includes a processor, a memory, and a program stored on the memory and executable on the processor. When the processor executes the program, it implements at least the following steps: obtaining an image collected by a shooting device, a point cloud collected by a lidar, weather information, and current rainfall amount. The image collected by the shooting device is a video of the window of a vehicle captured by the shooting device, and the point cloud collected by the lidar is a point cloud image of the window of the vehicle scanned by the lidar; using a deep learning model to process the image collected by the shooting device, the point cloud collected by the lidar, the weather information, and the current rainfall amount respectively to obtain a first degree of dirtiness, a second degree of dirtiness, a third degree of dirtiness, and a fourth degree of dirtiness, and performing weighted summation processing on the first degree of dirtiness, the second degree of dirtiness, the third degree of dirtiness, and the fourth degree of dirtiness to obtain a final judgment value, and determining whether the window is dirty according to the size of the final judgment value; in the case of determining that the window is dirty, executing a first processing method or a second processing method. The first processing method represents controlling the windshield wiper to clean the window, and the second processing method represents cleaning the window by using a water spraying and air jetting method, or only using an air jetting method to clean the window. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.

[0081] The present application also provides a computer program product, which is suitable for executing a program initialized with at least the following method steps when executed on a data processing device: obtaining an image collected by a shooting device, a point cloud collected by a lidar, weather information, and current rainfall amount. The image collected by the shooting device is a video of the window of a vehicle captured by the shooting device, and the point cloud collected by the lidar is a point cloud image of the window of the vehicle scanned by the lidar; using a deep learning model to process the image collected by the shooting device, the point cloud collected by the lidar, the weather information, and the current rainfall amount respectively to obtain a first degree of dirtiness, a second degree of dirtiness, a third degree of dirtiness, and a fourth degree of dirtiness, and performing weighted summation processing on the first degree of dirtiness, the second degree of dirtiness, the third degree of dirtiness, and the fourth degree of dirtiness to obtain a final judgment value, and determining whether the window is dirty according to the size of the final judgment value; in the case of determining that the window is dirty, executing a first processing method or a second processing method. The first processing method represents controlling the windshield wiper to clean the window, and the second processing method represents cleaning the window by using a water spraying and air jetting method, or only using an air jetting method to clean the window.

[0082] The present application also provides a vehicle system, including: one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include those for executing any of the above methods. By processing the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall, a dirt determination result and the current dirt type are obtained. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-type dirt or rain-type dirt; in the case where the dirt determination result indicates that the window of the sensor is dirty, a first processing method or a second processing method is executed, so that the present application can clean the window of the sensor more precisely than the existing solutions, making the detection result of the sensor more consistent with the actual working conditions, thereby solving the problem that the existing solutions use single image recognition to identify the dirt on the window of the vehicle sensor, which easily makes the dirt determination result inconsistent with the actual situation, further causing an occlusion effect on the driver's line of sight and easily posing a serious safety hazard.

[0083] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general-purpose computing device. They can be concentrated on a single computing device or distributed on a network composed of multiple computing devices. They can be implemented by program codes executable by the computing device, so that they can be stored in a storage device and executed by the computing device. And in some cases, the steps shown or described can be executed in a different order from here, or they can be separately fabricated into individual integrated circuit modules, or multiple modules or steps among them can be fabricated into a single integrated circuit module for implementation. In this way, the present invention is not limited to any specific combination of hardware and software.

[0084] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memories, CD-ROMs, optical memories, etc.) containing computer-usable program codes.

[0085] This application is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a means for implementing the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 a means for implementing the functions specified in one block or multiple blocks.

[0086] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including an instruction means that implements the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 a means for implementing the functions specified in one block or multiple blocks.

[0087] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in the flow Figure 1 one flow or multiple flows and / or blocks Figure 1 a means for implementing the functions specified in one block or multiple blocks.

[0088] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and a memory.

[0089] The memory may include non-permanent memory in the form of computer-readable media, random access memory (RAM), and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of computer-readable media.

[0090] A computer-readable medium includes permanent and non-permanent, removable and non-removable media that can implement information storage by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, disk storage, or other magnetic storage devices, or any other non-transitory medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer-readable medium does not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0091] It should also be noted that the term "comprising", "including", or any other variant thereof is intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a series of elements includes not only those elements but also other elements not expressly listed, or elements that are inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0092] From the above description, it can be seen that the above embodiments of the present application achieve the following technical effects:

[0093] 1), The window cleaning method for the sensors of the vehicle in the present application processes the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall amount to obtain a dirt determination result and the current dirt type. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt; in the case where the dirt determination result indicates that the window of the sensor is dirty, the first processing method or the second processing method is executed, so that the present application can clean the window of the sensor more accurately than the existing solutions, making the detection result of the sensor more consistent with the actual working conditions, and thus solving the problem that the existing solutions use single image recognition to identify the dirt on the window of the vehicle sensor, which easily makes the dirt determination result inconsistent with the actual situation, thereby blocking the driver's line of sight and easily causing serious safety hazards.

[0094] 2) The window cleaning device of the vehicle sensor in this application processes the images collected by the imaging device, the point clouds collected by the lidar, the weather information, and the current rainfall amount to obtain a dirt determination result and the current dirt type. The dirt determination result indicates whether the window of the sensor is dirty, and the current dirt type is sludge-based dirt or rainwater-based dirt. When the dirt determination result indicates that the window of the sensor is dirty, the first processing method or the second processing method is executed. As a result, compared with the existing solutions, this application can clean the window of the sensor more precisely, making the detection result of the sensor more consistent with the actual working conditions. Furthermore, it solves the problem that the existing solutions use single image recognition to detect the dirt on the window of the vehicle sensor, which easily leads to a discrepancy between the dirt determination result and the actual situation, thus blocking the driver's line of sight and posing a serious safety hazard.

[0095] The foregoing are only the preferred embodiments of this application and are not intended to limit this application. For those skilled in the art, this application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this application shall be included within the protection scope of this application.

Claims

1. A method for cleaning the window of a vehicle sensor, characterized in that: include: Acquire images collected by a camera, point clouds collected by a laser radar, weather information, and current rainfall, wherein the current rainfall is the rainfall at the current moment collected by a rainfall sensor, and the weather information includes at least rainfall monitoring results and wind speed, wherein the rainfall monitoring results indicate whether it is raining in the environment where the vehicle is located; The image collected by the shooting device, the point cloud collected by the laser radar, the weather information and the current rainfall are processed to obtain a dirt judgment result and a current dirt type, wherein the dirt judgment result indicates whether dirt appears on the sensor window, and the current dirt type is mud dirt or rain water dirt; When the dirt judgment result indicates that the window of the sensor is dirty, a first processing method or a second processing method is executed, wherein the first processing method indicates controlling the wiper to clean the window, and the second processing method indicates using water spray and air jet to clean the window, or only using air jet to clean the window, wherein when executing the second processing method, according to the current dirt type, the window is cleaned by using water spray and air jet, or only using air jet to clean the window.

2. The method according to claim 1, characterized in that Executing the first processing method or the second processing method includes: In the case where the target sensor is installed at the windshield of the vehicle, the first processing method is performed, and the target sensor represents a sensor where the window is dirty; When the target sensor is installed at a location other than the windshield in the vehicle, the second processing method is performed.

3. The method according to claim 1, characterized in that According to the current dirt type, the window is cleaned by spraying water and jetting air, or the window is cleaned by only jetting air, including: When it is determined that the current dirt type is the sludge dirt, firstly flushing and removing it by spraying water, and then cleaning the remaining water droplets by jetting air; When it is determined that the current dirt type is rain-related dirt, the dirty window is cleaned using high-pressure gas generated by a blower or an air compressor.

4. The method according to claim 1, characterized in that The image collected by the shooting device, the point cloud collected by the laser radar, the weather information and the current rainfall are processed to obtain a dirt determination result and a current dirt type, including: Using a deep learning model to process the image captured by the shooting device to obtain a first dirtiness degree and a first dirtiness type determination coefficient; The laser radar collected point cloud is processed by a cross-validation method to obtain a second dirtiness degree and a second dirtiness type determination coefficient; Processing the weather information in a mapping manner to obtain a third dirtiness degree and a third dirtiness type determination coefficient; Processing the current rainfall in a mapping manner to obtain a fourth pollution degree and a fourth pollution type determination coefficient; Performing weighted sum processing on the first dirtiness degree, the second dirtiness degree, the third dirtiness degree, and the fourth dirtiness degree to obtain a final dirtiness result judgment value, and performing weighted sum processing on the first dirtiness type determination coefficient, the second dirtiness type determination coefficient, the third dirtiness type determination coefficient, and the fourth dirtiness type determination coefficient to obtain a final dirtiness type judgment value; The contamination judgment result is determined according to the size of the final contamination result judgment value, and the current contamination type is determined according to the size of the final contamination type judgment value.

5. The method according to claim 4, characterized in that The first dirtiness level, the second dirtiness level, the third dirtiness level and the fourth dirtiness level are weighted and summed to obtain a final dirtiness result judgment value, including: assigning different weights to the first dirtiness level, the second dirtiness level, the third dirtiness level, and the fourth dirtiness level; Based on the assigned weights, a weighted summation process is performed on the first dirtiness level, the second dirtiness level, the third dirtiness level and the fourth dirtiness level to obtain the final dirtiness result judgment value.

6. The method according to claim 5, characterized in that Assigning different weights to the first dirtiness level, the second dirtiness level, the third dirtiness level, and the fourth dirtiness level comprises: In a mapping manner, a first preset weight corresponding to the first degree of dirtiness, a second preset weight corresponding to the second degree of dirtiness, a third preset weight corresponding to the third degree of dirtiness, and a fourth preset weight corresponding to the fourth degree of dirtiness are determined.

7. The method according to claim 1, characterized in that The method further comprises: When a cleaning instruction is received, the first processing method or the second processing method is executed based on the cleaning instruction.

8. The method according to any one of claims 1 to 7, characterized in that The method further comprises: Obtain the temperature of each actuator and the pressure and flow of each water spray system to obtain monitoring data; According to the size of the monitoring data, a prompt message is generated to prompt the size of the monitoring data or an early warning message is generated to prompt that the monitoring data exceeds the limit.

9. A window cleaning device for a sensor of a vehicle, characterized in that: include: An acquisition unit, used to acquire images collected by a shooting device, point clouds collected by a laser radar, weather information and current rainfall, wherein the current rainfall is the rainfall at the current moment collected by a rainfall sensor, and the weather information at least includes a rainfall monitoring result and a wind speed, wherein the rainfall monitoring result indicates whether it is raining in the environment where the vehicle is located; A first processing unit is used to process the image collected by the shooting device, the point cloud collected by the laser radar, the weather information and the current rainfall to obtain a dirt judgment result and a current dirt type, wherein the dirt judgment result indicates whether dirt appears on the window of the sensor, and the current dirt type is mud dirt or rain water dirt; The second processing unit is used to execute the first processing method or the second processing method when the dirt judgment result indicates that the window of the sensor is dirty. The first processing method indicates controlling the wiper to clean the window, and the second processing method indicates using water spray and air jet to clean the window, or only using air jet to clean the window. When executing the second processing method, the window is cleaned by water spray and air jet, or only using air jet to clean the window according to the current dirt type.

10. A vehicle system, characterized in that: include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the one or more processors, and the one or more programs include methods for executing any one of claims 1 to 8.

Images