Air conditioning defogging method and device of vehicle, electronic equipment and storage medium

Abstract

The application discloses a vehicle air conditioner defogging method and device, electronic equipment and a storage medium, and relates to the technical field of vehicle monitoring. The method comprises the following steps: obtaining image data of a preset windshield area in a vehicle after the whole vehicle is powered on; determining fogging state information in the preset windshield area based on the image data and a trained fogging identification model; wherein the fogging state information comprises a fogging area in the preset windshield area and a fogging intensity corresponding to the fogging area; determining a current fogging level of the preset windshield area based on the fogging state information; generating an air conditioner control instruction based on the current fogging level, and controlling the air conditioner in the vehicle to defog the preset windshield area according to the air conditioner control instruction; wherein the air conditioner control instruction is used to instruct the air conditioner to perform a defogging action matched with the current fogging level. The embodiments provided in the application can timely identify the fogging state of the vehicle window, reduce the time difference of air conditioner defogging, and thus improve the user experience.

Description

Technical Field

[0001] This application relates to the field of vehicle monitoring technology, and in particular to a method, device, electronic device and storage medium for defogging the air conditioning of a vehicle. Background Technology

[0002] Currently, vehicle defogging functions are typically based on monitoring and processing data such as temperature and humidity inside and outside the vehicle. They determine the vehicle's dew point temperature and the air conditioning defogging setting to be tested based on environmental analysis, and then control the air conditioning to perform defogging according to the preset defogging setting.

[0003] However, the temperature and humidity data collected by sensors and other devices inside the vehicle are lagging indicators, while windshield fogging is rapid. This results in a time lag in the air conditioning defogging function, which monitors temperature and humidity, and can only perform defogging after fog has formed, affecting the user experience. Summary of the Invention

[0004] This application provides a vehicle air conditioning defogging method, device, electronic device, and storage medium. The embodiments provided by this application solve the technical problem that the existing air conditioning defogging has a certain time difference and can only be defogging after fog has formed, which affects the user experience. The embodiments provided by this application can identify the fogging state of the car window in a timely manner, reduce the time difference of air conditioning defogging, and thus improve the user experience.

[0005] In a first aspect, this application provides a method for defogging an air conditioner in a vehicle, the method comprising: After the vehicle is powered on, image data of a preset windshield area inside the vehicle is acquired; Based on the image data and the trained fogging recognition model, the fogging state information within the preset windshield area is determined; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area; Based on the atomization status information, the current atomization level of the preset windshield area is determined; Based on the current fogging level, an air conditioning control command is generated, and the air conditioning in the vehicle is controlled to defog the preset windshield area according to the air conditioning control command; wherein, the air conditioning control command is used to instruct the air conditioning to perform a defogging action that matches the current fogging level.

[0006] In one feasible implementation, determining the fogging state information within the preset windshield area based on the image data and the trained fogging recognition network model includes: The image data is input into a trained fog recognition network model to determine the fog area under the preset windshield area; Based on the area ratio of the atomized region in the preset windshield area and the average opacity of the atomized region, the global atomization index of the atomized region is determined, and the global atomization index is used as the atomization intensity corresponding to the atomized region.

[0007] In one feasible implementation, the step of inputting the image data into a trained fogging recognition network model to determine the fogging area under the preset windshield area includes: The trained fogging recognition model is used to perform semantic segmentation on the image data to determine the label corresponding to each sub-region in the image data; the label is used to characterize whether fogging exists on the surface of the sub-region. Based on the label corresponding to each sub-region in the image data, the fogging area under the preset windshield area is determined.

[0008] In one feasible implementation, the trained fog recognition model is obtained by training in the following manner: Obtain a sample training set, wherein the sample training set includes sample image data of multiple preset windshield areas; each sample image data contains corresponding standard fogging state information; The sample image data is input into the initial fog recognition model to obtain the predicted labels corresponding to multiple sub-regions; Based on standard atomization state information, determine the standard label corresponding to each of the sub-regions; The training loss is determined based on the differences between the predicted labels and standard labels corresponding to the multiple sub-regions. The parameters of the initial fog recognition model are adjusted based on the training loss to obtain a trained fog recognition model.

[0009] In one feasible implementation, determining the current atomization level of the preset windshield area based on the atomization state information includes: When the atomization intensity is less than the first intensity threshold, the current atomization level of the preset windshield area is determined to be the first level, wherein the first level indicates that the preset windshield area is in a clear state. When the degree of fogging is greater than or equal to the first intensity threshold and less than the second intensity threshold, the current fogging level of the preset windshield area is determined to be the second level, wherein the second level indicates that the preset windshield area is in a slightly fogged state; When the degree of atomization is greater than or equal to the second intensity threshold and less than the third intensity threshold, the current atomization level of the preset windshield area is determined to be the third level, wherein the third level indicates that the preset windshield area is in a moderate atomization state. When the degree of fogging is greater than or equal to the third intensity threshold, the current fogging level of the preset windshield area is determined to be the fourth level, wherein the fourth level indicates that the preset windshield area is in a state of severe fogging; the third intensity threshold is greater than the second intensity threshold, and the second intensity threshold is greater than the first intensity threshold.

[0010] In one feasible implementation, generating the air conditioning control command based on the current atomization level includes: Based on the vehicle's environmental parameters and the current atomization level, the target operating parameters of the air conditioner are determined; Based on the target operating parameters, air conditioning control commands are generated.

[0011] In one feasible implementation, the air conditioner includes multiple air outlets, each air outlet having at least two air outlet directions, and the target operating parameters include the target air outlet direction; determining the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level includes: Based on the vehicle's environmental parameters, the current atomization level, and the position of the atomization area in the preset windshield area, a target air outlet is determined from multiple air outlets, as well as the target air outlet's target air outlet direction.

[0012] In one feasible implementation, the environmental parameters include at least one of in-vehicle temperature, out-of-vehicle temperature, in-vehicle relative humidity, and out-of-vehicle relative humidity; the target operating parameters include at least one of air outlet mode, fan speed setting, target temperature, and air source mode; and determining the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level includes: When the current atomization level is determined to be the third level and the in-vehicle temperature is greater than the dew point temperature, the air source mode of the air conditioner is determined to be external circulation, the fan speed is determined to be the first preset speed, the target temperature is to reduce the current in-vehicle temperature to the first preset temperature, and the air outlet mode is determined to be the windshield air outlet mode.

[0013] In one feasible implementation, determining the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level includes: When the current atomization level is determined to be the fourth level, and the atomization intensity is greater than the rate of change of a preset rate of change threshold, the air source mode of the air conditioner is determined to be internal circulation, the fan speed is determined to be the maximum speed, the air outlet mode is determined to be the front windshield air outlet mode, and the target temperature is determined to raise the current temperature inside the vehicle to the second preset temperature, and the electric heating fan is activated, wherein the second preset temperature is greater than the first preset temperature, and the maximum speed is greater than the first preset speed.

[0014] In one feasible implementation, the method further includes: During the defogging process, updated image data of the preset windshield area after defogging is acquired; Based on the updated image data, the clarity status of the preset windshield area is determined; When the clarity status meets the preset recovery conditions, an air conditioning recovery command is generated; wherein, the air conditioning recovery command is used to instruct the air conditioner to stop defogging the preset windshield area and exit the defogging mode.

[0015] In one feasible implementation, the sharpness state is determined to meet a preset recovery condition by the following method: When the degree of fogging corresponding to the updated image data is determined to be less than the fourth intensity threshold, the clarity state is determined to meet the preset recovery condition.

[0016] In one feasible implementation, before determining the fogging state information within the preset windshield area based on the image data and the trained fogging recognition model, the method further includes: Based on external rainfall sensor data, determine whether the current weather is rainy or snowy; When the current weather is determined to be rainy or snowy weather, the image data is subjected to occlusion identification to determine the dynamic occlusions in the image data; When the motion characteristics of the dynamic occlusion satisfy the preset raindrop motion characteristics, the corresponding pixel area of ​​the dynamic occlusion is excluded from the preset windshield area.

[0017] In one feasible implementation, the method further includes: When the position change of the atomized area within a preset time window is less than a preset position change threshold, and the change in atomization intensity is less than a preset intensity change threshold, the atomized area is marked as a dirty area, and a cleaning prompt message is generated.

[0018] In one feasible implementation, before acquiring image data of a preset windshield area inside the vehicle, the method further includes: Obtain the current light intensity of the external environment; When the current light intensity is less than the preset light threshold intensity, the lighting equipment inside the vehicle is turned on to assist in acquiring the image data of the preset windshield area inside the vehicle.

[0019] In a second aspect, this application provides a vehicle air conditioning defogging device, the vehicle air conditioning defogging device comprising: The acquisition module is used to acquire image data of a preset windshield area inside the vehicle after the vehicle is powered on; The first determining module is used to determine the fogging state information within the preset windshield area based on the image data and the trained fogging recognition model; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area; The second determining module is used to determine the current atomization level of the preset windshield area based on the atomization state information; The generation module is used to generate air conditioning control instructions based on the current atomization level, and control the air conditioning in the vehicle to defog the preset windshield area according to the air conditioning control instructions; wherein, the air conditioning control instructions are used to instruct the air conditioning to perform a defogging action that matches the current atomization level.

[0020] In a third aspect, this application provides an electronic device, including a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the machine-readable instructions are executed by the processor to perform steps of the vehicle air conditioning defogging method described above.

[0021] In a fourth aspect of this application, an embodiment of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the vehicle air conditioning defogging method described above.

[0022] Compared with the prior art, the vehicle air conditioning defogging method, device, electronic device, and storage medium provided in this application have the following advantages: After the vehicle is powered on, the embodiments of this application acquire image data of a preset windshield area inside the vehicle, and based on the image data and a trained fogging recognition model, determine the fogging state information of the preset windshield area. Then, based on the fogging state information, determine the current fogging level of the preset windshield area, and based on the current fogging level, generate an air conditioning control command and control the air conditioning in the vehicle to defog the preset windshield area according to the air conditioning control command. This application can realize timely and early identification of the fogging state of the window and control the air conditioning in the vehicle to perform intelligent defogging, reduce the time lag of air conditioning defogging, and thus improve the user experience. Attached Figure Description

[0023] Figure 1 A flowchart of a vehicle air conditioning defogging method provided in an embodiment of this application is shown; Figure 2 This paper shows a structural block diagram of a vehicle air conditioning defogging device according to an embodiment of the present application; Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown.

[0024] Figure 2 and Figure 3 The correspondence between the figure labels and figure titles in the accompanying drawings is as follows: 200 Vehicle air conditioning defrosting device; 210 Auxiliary acquisition module; 220 Acquisition module; 230 First determination module; 240 Second determination module; 250 Generation module; 260 Update module; 270 Obstruction recognition module; 280 Dirty area recognition module; 300 Electronic device; 310 Processor; 320 Memory; 330 Bus. Detailed Implementation

[0025] To better understand the technical solutions provided in the embodiments of this specification, the technical solutions of the embodiments of this specification will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0026] In this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element. The term "two or more" includes two or more cases.

[0027] First, the applicable application scenarios of this application are introduced. The embodiments provided in this application are applicable to the field of vehicle monitoring technology, and can be specifically applied to new energy vehicles, autonomous vehicles and intelligent cockpit systems, especially relating to a vehicle air conditioning defogging method, device, electronic device and storage medium.

[0028] Currently, the temperature and humidity data collected by sensors and other devices inside vehicles are lagging indicators, while windshield fogging is rapid. This results in a time lag in the defogging function of the air conditioning system, which monitors temperature and humidity. Furthermore, defogging can only be performed after fog has formed, which affects the user experience and consumes a lot of energy.

[0029] Based on this, the embodiments of this application provide a method, device, electronic device and storage medium for defogging air conditioning in vehicles. The embodiments provided by this application solve the technical problem that there is a certain time difference in air conditioning defogging in the prior art, and that defogging can only be performed after fog has formed, which affects the user experience. The embodiments provided by this application can detect fog in a timely manner, reduce the time difference of air conditioning defogging, and thus improve the user experience.

[0030] Figure 1 A flowchart illustrating a vehicle air conditioning defogging method according to an embodiment of this application is shown. Figure 1 As shown, the vehicle's air conditioning defogging method includes the following steps: S101. After the vehicle is powered on, acquire image data of the preset windshield area inside the vehicle.

[0031] In this step, in the embodiments provided in this application, after determining that the vehicle to be detected is powered on, the camera or radar inside the vehicle starts to work and acquires real-time video stream data of the vehicle's windshield. Then, the video stream data is used to extract features according to the preset key area of ​​windshield fogging to determine the preset windshield area inside the vehicle and the original image data of the preset windshield area. Then, the original image data is standardized to determine the image data of the preset windshield area inside the vehicle.

[0032] It should be noted that the embodiments provided in this application define the preset windshield fogging key area as the preset windshield area inside the vehicle.

[0033] It is understood that the camera provided in the embodiments of this application can be customized according to different application scenarios and usage conditions. In the embodiments provided in this application, the camera can be specifically set as a dedicated cabin camera or a forward-facing monocular camera. The field of view of the camera is set to be greater than 100°, and the forward-facing monocular camera has the function of high dynamic range monitoring.

[0034] Furthermore, the image data acquisition and image data processing operations of the embodiments provided in this application can be integrated into the vehicle's Advanced Driver Assistance Systems (ADAS).

[0035] The preset windshield area provided in this application can be customized according to different application scenarios and usage conditions. The preset windshield area provided in this application includes at least the central main viewing area of ​​the windshield and the side area near the A-pillar.

[0036] Here, in the embodiments provided in this application, the central main viewing area of ​​the windshield can specifically be the fan-shaped area directly above the steering wheel and in the lower middle part of the windshield, that is, the core area directly in front of the driver; the side area of ​​the A-pillar is specifically the triangular area where the left and right edges of the windshield are close to the A-pillar and intersect with it; the preset windshield area in the embodiments provided in this application also includes the defrost airflow blind spot and the field of view of the autonomous driving forward-view camera, and the defrost airflow blind spot is specifically the area at the bottom of the windshield close to the hood, as well as the blind spot area that the wipers cannot sweep; the field of view of the autonomous driving forward-view camera is specifically the upper central area of ​​the windshield directly opposite the back of the rearview mirror.

[0037] In the embodiments provided in this application, the frequency at which the camera acquires image data of a preset windshield area inside the vehicle is customized according to different application scenarios and usage conditions.

[0038] S102. Based on image data and a trained fogging recognition model, determine the fogging state information within a preset windshield area; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area.

[0039] In this step, based on image data and a trained fogging recognition model, the fogging state information within a preset windshield area is determined. This is achieved by inputting the collected windshield area image into a deep learning model that has been trained to convergence. The model automatically extracts visual features from the image and performs inference and judgment, ultimately outputting structured information that reflects whether fog exists and its distribution in the preset windshield area. The fogging state information includes at least two dimensions: fogging area and fogging intensity, which are used to quantify the current degree of obstruction of the windshield.

[0040] It should be noted that, in the embodiments provided in this application, the trained fog recognition model has learned the visual feature distribution patterns of surface states such as clear, fog, water droplet condensation, and frost by a large number of sample windshield images labeled with pixel-level sub-region labels during the training phase. Therefore, in the actual inference phase, the trained fog recognition model can classify the input image data pixel by pixel, and the output classification result is a label map reflecting the surface state of the preset windshield area. Furthermore, the trained fog recognition model in the embodiments provided in this application can specifically analyze the image data of the preset windshield area once every 500ms.

[0041] Understandably, the label image is a pixel-level classification result image corresponding to the resolution of the input image data. Each pixel is labeled with a label representing the surface state category at that location. Based on this label image, the region consisting of all pixels labeled as fog, water droplet condensation, or frost in the preset windshield area can be determined. The continuous area of ​​fog is the fogging area, which represents the specific range of the windshield actually obscured by fog. At the same time, based on the preset mathematical algorithm, the fogging intensity corresponding to the fogging area is output.

[0042] Among them, the atomization intensity is a comprehensive quantitative index calculated based on the area ratio of the atomized area in the preset windshield area and the average opacity of the atomized area. The atomization intensity and the atomized area together constitute the atomization state information, which is used to provide numerical basis for the subsequent atomization level determination step, so as to support the air conditioning system to execute differentiated defogging control strategies according to different degrees of atomization.

[0043] S103. Based on the atomization status information, determine the current atomization level of the preset windshield area.

[0044] In this step, in the embodiments provided in this application, the fogging state information determined in the previous stage is comprehensively evaluated according to the preset level classification rules to determine the current fogging severity of the preset windshield area, so as to determine the fogging level to which the preset windshield area currently belongs. The fogging level is used to characterize the fogging severity of the current windshield area and to provide a graded basis for the selection of subsequent air conditioning control strategies.

[0045] It should be noted that the atomization level is divided based on the atomization intensity in the atomization state information. The atomization intensity value is a continuous numerical range, and different numerical ranges correspond to different atomization levels. In the embodiments provided in this application, multiple atomization levels are preset, and each atomization level corresponds to a range of atomization intensity values. The current atomization level can be determined by judging the numerical range in which the atomization intensity is located.

[0046] Understandably, the higher the fog intensity, the denser or wider the fog coverage in the preset windshield area, and the higher the current fog level is determined, indicating a more severe fog condition. Different fog levels correspond to different defogging response strategies. Lower levels may only require reminders or preparatory actions, while higher levels require triggering the air conditioning system's active defogging intervention.

[0047] In the embodiments provided in this application, the atomization level can be customized and used according to different application scenarios and usage conditions. The atomization level in the embodiments provided in this application can be specifically set to 3 levels, including at least the first level, the second level and the third level, with the atomization intensity threshold of each level increasing sequentially.

[0048] Here, once the current atomization level is determined, it is output to the air conditioner for decision-making to trigger an air conditioner control strategy that matches the current atomization level.

[0049] S104. Based on the current atomization level, generate an air conditioning control command and control the air conditioning in the vehicle to defog the preset windshield area according to the air conditioning control command; wherein, the air conditioning control command is used to instruct the air conditioning to perform a defogging action that matches the current atomization level.

[0050] In this step, in the embodiment provided in this application, based on the current atomization level determined in the previous stage, an air conditioning control command matching the current atomization level is generated, and the air conditioning system in the vehicle is controlled to perform corresponding defogging actions according to the air conditioning control command, so as to perform targeted defogging treatment on the preset windshield area. The air conditioning control command is used to instruct the air conditioning system to perform defogging operation according to the operating parameters corresponding to the current atomization level.

[0051] It should be noted that the generation of air conditioning control commands is based on at least the current atomization level and the vehicle's environmental parameters. The environmental parameters include at least one of the following: in-vehicle temperature, outside vehicle temperature, in-vehicle relative humidity, and outside vehicle relative humidity. The embodiments provided in this application determine the target air conditioning operating parameters that are suitable for the current atomization scenario by comprehensively matching the current atomization level with the environmental parameters. The target air conditioning operating parameters include at least one or more of the following: air outlet mode, air volume level, target temperature, and air source mode. Air conditioning control commands are generated based on the target air conditioning operating parameters.

[0052] Understandably, different atomization levels correspond to different defogging control strategies. The lower the atomization level, the lower the intensity of the defogging action and the less vehicle energy is consumed. The higher the atomization level, the higher the intensity of the defogging action. The air conditioning system will perform defogging with stronger operating parameters to quickly restore clear visibility in the preset windshield area. After the air conditioning control command is sent to the air conditioning system, the air conditioning system will parse the air conditioning control command and drive the corresponding actuator to perform the defogging action.

[0053] The target air conditioning operating parameters are determined through a preset control strategy. When the current atomization level is level one, the air conditioning control command is used to maintain the current air conditioning operation or only perform a prompt operation. When the current atomization level is level two, the air conditioning control command is used to control the air conditioning system to perform defogging with the first preset operating parameters. When the current atomization level is level three, the air conditioning control command is used to control the air conditioning system to perform defogging with the second preset operating parameters, where the defogging intensity of the second preset operating parameters is higher than that of the first preset operating parameters. During the defogging process, the atomization state of the preset windshield area is continuously monitored to determine whether the defogging effect has reached the preset recovery conditions.

[0054] Compared with the prior art, the vehicle air conditioning defogging method provided in this application acquires image data of a preset windshield area after the vehicle is powered on. Based on the image data and a trained fogging recognition model, it determines the fogging state information of the preset windshield area, determines the current fogging level of the preset windshield area based on the fogging state information, and generates an air conditioning control command based on the current fogging level. The air conditioning in the vehicle is then controlled to defog the preset windshield area according to the air conditioning control command. This application can realize timely and early identification of the fogging state of the windows and control the air conditioning in the vehicle to perform intelligent defogging, reducing the time lag of air conditioning defogging and thus improving the user experience.

[0055] Optionally, based on image data and a trained fog recognition network model, the fog state information within a preset windshield area is determined, including: inputting image data into the trained fog recognition network model to determine the fog area under the preset windshield area; determining the global fog index of the fog area based on the area ratio of the fog area in the preset windshield area and the average opacity of the fog area, and using the global fog index as the fog intensity corresponding to the fog area.

[0056] In the embodiments provided in this application, image data is input into a trained fog recognition network model. The fog recognition network model performs feature extraction and pixel-by-pixel classification on the image data to determine the fog area under the preset windshield area. Then, based on the fog area, the area ratio of the fog area in the preset windshield area is further determined, that is, the ratio of the number of pixels contained in the fog area to the total number of pixels corresponding to the preset windshield area. The area ratio is used to quantitatively characterize the coverage of fog in the preset windshield area. At the same time, based on the image feature values ​​of each pixel corresponding to the fog area, the average opacity of the fog area can be determined. Finally, based on the area ratio and the average opacity, the global fog index is calculated by multiplying the area ratio and the average opacity. The global fog index is a continuous numerical index used to comprehensively characterize the current overall fog severity of the preset windshield area. That is, the global fog index is the fog intensity corresponding to the fog area.

[0057] Specifically, the trained fog recognition model in the embodiments provided in this application is a neural network model built based on a deep learning semantic segmentation network. Its input is preprocessed image data, and its output is a pixel-level label map corresponding to the resolution of the image data. The label value of each pixel in the label map is used to characterize the surface state category to which the pixel position belongs. The surface state category includes at least clear state and fog state. By extracting all pixels in the label map with the label value of fog state, the connected region formed by the continuously distributed pixels is determined as the fog region.

[0058] Understandably, the fogged area represents the specific spatial range currently covered by fog within the preset windshield area. Based on the fogged area, the area ratio of the fogged area within the preset windshield area can be further determined, that is, the ratio of the number of pixels contained in the fogged area to the total number of pixels corresponding to the preset windshield area. The area ratio is used to quantitatively characterize the coverage of fog within the preset windshield area.

[0059] The neural network models in the embodiments provided in this application can be customized and selected according to different application scenarios and usage conditions. Specifically, the neural network models in the embodiments provided in this application can be semantic segmentation network models, such as Bilateral Segmentation Network (BiSeNet-V2) and Efficient Spatial Pyramid Network for Semantic Segmentation (ESPNet-seg). The model size of the semantic segmentation network is compressed to less than 5MB, and the inference speed is greater than 20 milliseconds / frame, using INT8 precision inference.

[0060] In the above-mentioned implementation method provided in this application, the windshield image data is classified by region, i.e. by pixel, through a deep learning semantic segmentation network, which can accurately identify the spatial boundary of the fogging area and avoid the problem of response lag in traditional temperature and humidity detection methods. At the same time, the spatial distribution characteristics and concentration characteristics of fog are integrated into a quantitative index by multiplying the area ratio and the average opacity, making the characterization of fogging intensity more comprehensive and objective.

[0061] Optionally, the image data is input into a trained fogging recognition network model to determine the fogging area under a preset windshield area, including: The trained fogging recognition model performs semantic segmentation on the image data to determine the label corresponding to each sub-region in the image data; the label is used to characterize whether fogging exists on the surface of the sub-region; based on the label corresponding to each sub-region in the image data, the fogging area under the preset windshield area is determined.

[0062] In the above process, image data is input into a trained fogging recognition network model, which performs semantic segmentation on the image data to determine the label corresponding to each sub-region in the image data. The label is used to characterize whether fogging exists at the windshield surface corresponding to the sub-region, that is, whether the surface state of the sub-region is clear or fogged. Based on the label corresponding to each sub-region, all sub-regions marked as fogged can be identified from the preset windshield area, and these sub-regions marked as fogged together constitute the fogged area.

[0063] It should be noted that the sub-region is the basic processing unit when performing semantic segmentation on image data. In the embodiments provided in this application, the sub-region is specifically a single pixel in the image data, that is, the semantic segmentation operation processes the image data in a pixel-by-pixel classification manner; the label is the classification result output by the fog recognition network model for each sub-region. The classification result is used to indicate the surface state category to which the sub-region belongs. The surface state category includes at least the clear state that is not fogged, the fogged state that is fogged, the water vapor condensation state, and / or the frosting state, etc.

[0064] Understandably, a well-trained fog recognition network model has learned the visual feature distribution patterns of different surface state categories in images during the training phase. Therefore, the trained fog recognition network model can predict the probability of each sub-region belonging to each surface state category based on the image features of each sub-region in the input image data, and determine the category with the highest probability as the label of that sub-region. Then, based on the label distribution map formed by the labels corresponding to all sub-regions in the image data, the spatial distribution of sub-regions marked as foggy in the preset windshield area can be determined. The foggy area consists of all interconnected sub-regions marked as foggy within the preset windshield area.

[0065] The spatial boundary of the fogging area is defined by the dividing line between the fogging status label and the clear status label in the label distribution map. After the fogging area is determined, it is output and used in the subsequent fogging intensity calculation step as the basis for calculating the area ratio of the fogging area in the preset windshield area and the average opacity of the fogging area.

[0066] This application uses a pixel-level classification-based method for determining fogged areas, which reduces interference from non-fog-obstructing factors such as raindrops and dust, effectively improving the accuracy and reliability of fogged state information determination.

[0067] Optionally, the trained fog recognition model is obtained by training in the following way: A sample training set is obtained, which includes sample image data of multiple preset windshield areas; each sample image data contains corresponding standard fogging state information; the sample image data is input into an initial fogging recognition model to obtain predicted labels corresponding to multiple sub-regions; based on the standard fogging state information, the standard label corresponding to each sub-region is determined; based on the difference between the predicted labels and standard label information corresponding to multiple sub-regions, the training loss is determined; based on the training loss, the parameters of the initial fogging recognition model are adjusted to obtain a trained fogging recognition model.

[0068] In the embodiments provided in this application, a sample training set is obtained, which includes sample image data of multiple preset windshield areas. Each sample image data contains corresponding standard fogging state information, which is used to characterize the surface state category label corresponding to each sub-region in the sample image data. The sample image data is input into an initial fogging recognition model, which performs feature extraction and sub-region classification processing on the sample image data and outputs a predicted label corresponding to each sub-region in the sample image data. The predicted label is the prediction result of the initial fogging recognition model for the surface state category of the sub-region. Based on the standard fogging state information, the sample image is determined. For each sub-region in the data, a standard label is generated, which is the manually labeled true value of the surface state category of that sub-region. For each sub-region in the sample image data, the difference between the predicted label and the standard label of that sub-region is calculated. Based on the comprehensive result of the differences corresponding to all sub-regions in the sample image data, the training loss corresponding to the sample image data is determined. With minimizing the training loss as the optimization objective, the backpropagation algorithm is used to iteratively adjust the model parameters of the initial fog recognition model until the training loss reaches the preset convergence condition and training stops. The model parameters obtained when training stops are used as the final parameters of the trained fog recognition model, thus obtaining the trained fog recognition model.

[0069] It should be noted that the sample training set consists of images acquired under various acquisition conditions from preset windshield areas of different vehicles (which can be the same vehicle, vehicles of the same type, or vehicles of different types). These acquisition conditions include at least various lighting conditions, various weather scenarios, and various levels of fogging. Lighting conditions include daytime light, low-light nighttime light, and tunnel light; weather scenarios include sunny days, rainy days, and snowy days; fogging levels include no fog, slight fogging, moderate fogging, and severe fogging; standard fogging state information is the annotation information obtained after labeling the surface state category of each sub-region in each sample image data; surface state categories include clear state, fogged state, water droplet condensation state, and frosted state; the initial fogging recognition model contains initialized model parameters, which control the behavior of the initial fogging recognition model in feature extraction and classification mapping of the input image.

[0070] Understandably, the process by which the initial fog recognition model outputs a predicted label for each sub-region essentially involves calculating the probability distribution of the sub-region's belonging to each surface state category based on the image features of that sub-region under the current model parameters, and using the category with the highest probability as the predicted label for that sub-region. The training loss reflects the overall deviation between the initial fog recognition model's predicted labels for all sub-regions and their corresponding standard labels under the current model parameters. A larger training loss indicates a greater deviation between the model's prediction results and the manually labeled true values, resulting in poorer model performance. Through the backpropagation algorithm, the gradient of the training loss is propagated layer by layer forward to the parameters of each layer of the initial fog recognition model, and the model parameters are updated based on the gradient, so that the updated model parameters can reduce the training loss.

[0071] The standard fogging state information is used to characterize the surface state category label corresponding to each sub-region in the sample image data.

[0072] The fog recognition model training method provided in this application constructs a large-scale labeled sample training set containing multiple acquisition conditions and multiple fog levels. This allows the model to fully learn the visual feature distribution patterns of various surface states under different lighting, climates, and fog levels during the training phase, effectively improving the model's generalization ability and recognition robustness in real-world complex scenarios. Simultaneously, the supervised learning method using sub-regions (pixels) as basic processing units enables the model to accurately learn the mapping relationship between the surface state category and image features at each pixel location, ensuring fine-grained model output. This provides a reliable model foundation for subsequent accurate delineation of fogged regions and precise quantification of fog intensity based on pixel-level classification results. Furthermore, the backpropagation algorithm iteratively optimizes the model parameters until convergence, enabling the trained fog recognition model to achieve high classification accuracy while maintaining a lightweight structure, meeting the dual requirements of model accuracy and inference speed for vehicle-mounted embedded deployments.

[0073] Optionally, based on the atomization status information, the current atomization level of the preset windshield area is determined, including: when the atomization intensity is less than a first intensity threshold, the current atomization level of the preset windshield area is determined to be level one, wherein level one indicates that the preset windshield area is in a clear state; when the atomization degree is greater than or equal to the first intensity threshold and less than a second intensity threshold, the current atomization level of the preset windshield area is determined to be level two, wherein level two indicates that the preset windshield area is in a slightly atomized state; when the atomization degree is greater than or equal to the second intensity threshold and less than a third intensity threshold, the current atomization level of the preset windshield area is determined to be level three, wherein level three indicates that the preset windshield area is in a moderately atomized state; when the atomization degree is greater than or equal to the third intensity threshold, the current atomization level of the preset windshield area is determined to be level four, wherein level four indicates that the preset windshield area is in a severely atomized state; wherein the third intensity threshold is greater than the second intensity threshold, and the second intensity threshold is greater than the first intensity threshold.

[0074] In the embodiments provided in this application, the current atomization level of the preset windshield area is determined by comparing the atomization intensity with a first intensity threshold, a second intensity threshold, and a third intensity threshold, respectively.

[0075] It should be noted that the first, second, and third intensity thresholds divide the continuous range of fogging intensity into four progressively increasing numerical intervals. Fog intensity is a global fogging index calculated based on the area ratio of the fogged region within the preset windshield area and the average opacity of the fogged region. The fogging intensity value increases monotonically as the fogging coverage area or fog density within the preset windshield area increases. The first level corresponds to the lowest fogging intensity interval, indicating that the surface of the preset windshield area is clear and fog-free. The second level corresponds to the second lowest fogging intensity interval, indicating that there is slight fogging in the preset windshield area but it has not yet seriously affected the driver's vision. The third level corresponds to the second highest fogging intensity interval, indicating that there is significant fogging in the preset windshield area and it has already obstructed the driver's vision to some extent. The fourth level corresponds to the highest fogging intensity interval, indicating that there is severe fogging in the preset windshield area and it has significantly obstructed the driver's vision. At this time, maximum defogging needs to be forcibly activated, and the driver will be prompted by voice.

[0076] Understandably, the determination of the current fog level is based on the numerical comparison between fog intensity and each preset threshold, rather than on subjective judgment or empirical estimation of the image. Therefore, it can objectively classify the fog state at different times and in different scenarios using a unified quantitative standard.

[0077] The specific values ​​of the first intensity threshold, the second intensity threshold, and the third intensity threshold can be customized and adjusted according to different types of vehicles, different regional climate conditions, or different users' needs for defogging response sensitivity. In this application, the first intensity threshold is 0.1; the second intensity threshold is 0.3; and the third intensity threshold is 0.6.

[0078] As the fogging intensity increases, the current fogging level gradually increases. Correspondingly, the severity of fogging in the preset windshield area increases step by step. The clear state is a normal vision state without fogging interference. The slightly fogged state is a preparatory state in which fogging has initially formed but has not yet triggered forced defogging. The moderate fogged state is a state in which fogging has significantly affected the vision and requires active defogging intervention. The severe fogged state is a state in which fogging has severely obscured the vision and requires rapid intervention with maximum defogging capability.

[0079] The first level corresponds to a state where no defogging action is required; the second level corresponds to a state where a reminder or preparatory operation can be performed; the third level corresponds to a state where an active defogging mode needs to be activated; and the fourth level corresponds to a state where the maximum power defogging mode needs to be forcibly activated. Once the current atomization level is determined, it is output to the air conditioner. The air conditioner generates an air conditioning control command that matches the current atomization level to instruct the air conditioning system to perform defogging actions according to the target air conditioning operating parameters corresponding to the current atomization level, thereby realizing graded differentiated defogging control based on the actual severity of atomization.

[0080] This application determines the current atomization level by using a threshold comparison method. The judgment logic is simple and clear, without the need for complex reasoning calculations, which helps to reduce the computational overhead of the vehicle controller and improve the system response speed. In addition, the values ​​of each intensity threshold can be customized and adjusted according to the actual application scenario, making the sensitivity of the atomization level judgment configurable and adaptable to the differentiated needs of different vehicle models, different climate conditions and different user preferences for defogging response sensitivity.

[0081] Optionally, based on the current atomization level, generate air conditioning control commands, including: Based on the vehicle's environmental parameters and current atomization level, the target operating parameters for the air conditioning are determined; based on the target operating parameters, air conditioning control commands are generated.

[0082] In the embodiments provided in this application, the current atomization level is comprehensively matched with environmental parameters and then a target operating parameter combination adapted to the current scene is output. After the air conditioning control command is generated, it is sent to the electronic control unit of the air conditioning system through the vehicle controller local area network to drive the actuator of the air conditioning system to perform defogging action according to the target operating parameters.

[0083] It should be noted that when an air conditioner performs defogging for the same atomization level under different environmental parameter conditions, the configuration of the target operating parameters will differ. This is because the physical mechanism of fog formation is different under different temperature and humidity conditions, and the required defogging methods are also different. For example, under low temperature and high humidity conditions, a combination of heating and external circulation is required to quickly eliminate fog, while under high temperature and high humidity conditions, a combination of cooling dehumidification and internal circulation is required. The current atomization level reflects the actual severity of fogging in the preset windshield area, while the environmental parameters reflect the external conditions that cause fogging to occur and persist. Combining the two allows the determination of the target operating parameters to take into account both the actual severity of the current fogging and its causes.

[0084] The environmental parameters include at least one or more of the following: in-vehicle temperature, outside vehicle temperature, in-vehicle relative humidity, and outside vehicle relative humidity; the target operating parameters include one or more of the following: air outlet mode, fan speed setting, target temperature, and air source mode; the air conditioning control command is a data format conforming to the vehicle communication protocol, used to instruct the air conditioning system to perform defogging action according to the target operating parameters.

[0085] This application uses both the current fogging level and environmental parameters as inputs to determine the target operating parameters. This allows the generation of air conditioning control commands to consider both the actual severity of fogging and its causes. Compared to single-dimensional control methods that rely solely on fogging level or environmental parameters, this enables the defogging strategy to more accurately adapt to the current actual defogging needs. At the same time, the preset control strategy is pre-installed in the decision module in a rule-based manner, eliminating the need for complex online optimization calculations. This helps reduce the computational load on the vehicle controller and improve the decision response speed, meeting the real-time requirements of defogging control.

[0086] Optionally, the air conditioner includes multiple air outlets, each with at least two air outlet directions, and the target operating parameters include the target air outlet direction; based on the vehicle's environmental parameters and the current atomization level, the target operating parameters of the air conditioner are determined, including: based on the vehicle's environmental parameters, the current atomization level, and the position of the atomization area in the preset windshield area, determining the target air outlet from the multiple air outlets, and the target air outlet direction.

[0087] In the above process, based on environmental parameters, the current atomization level, and the location information of the atomization area within the preset windshield area, a target air outlet is determined from multiple air outlets of the air conditioner according to a preset control strategy. The target air outlet direction is determined based on the location information of the atomization area. The target air outlet direction is associated with the location of the atomization area within the preset windshield area, so that the target air outlet delivers airflow toward the atomization area in the target air outlet direction. After determining the target air outlet and the target air outlet direction, an air conditioning control command is generated based on the target air outlet and the target air outlet direction. The air conditioning control command is used to instruct the air conditioning system to perform a defogging action through the target air outlet in the target air outlet direction.

[0088] It should be noted that the multiple air vents include at least the central air vent, the left air vent, and the right air vent of the windshield. The installation position of each air vent in the vehicle corresponds to different spatial sections of the preset windshield area. Each air vent has at least two airflow directions, which are adjusted by the deflection angle of the air guide plate or air guide grille set inside the air vent. Different airflow directions correspond to different deflection angles of the air guide plate or air guide grille, thereby achieving directional delivery of airflow to different positions in the preset windshield area. The position information of the atomizing area in the preset windshield area is determined by the spatial distribution position of the atomizing area in the label diagram. The position information includes at least the horizontal and / or vertical position of the atomizing area relative to the preset windshield area. The preset control strategy includes the correspondence between the position of the atomizing area and the target air vent and airflow direction. The correspondence is pre-established based on the installation position of each air vent, the airflow direction adjustment range of each air vent, and the spatial division of the preset windshield area.

[0089] It is understood that the multiple air vents include at least the central air vent of the windshield, the left air vent of the windshield, and the right air vent of the windshield. The installation position of each air vent in the vehicle corresponds to different spatial sections of the preset windshield area. Each air vent has at least two air outlet directions, which are adjusted by the deflection angle of the air guide plate or air guide grille set inside the air vent. Different air outlet directions correspond to different deflection angles of the air guide plate or air guide grille, thereby achieving directional delivery of airflow to different positions in the preset windshield area. The position information of the atomizing area in the preset windshield area is determined by the spatial distribution position of the atomizing area in the label diagram. The position information includes at least the horizontal position and / or vertical position of the atomizing area relative to the preset windshield area. The preset control strategy includes the correspondence between the position of the atomizing area and the target air vent and air outlet direction. The correspondence is pre-established based on the installation position of each air vent, the airflow adjustment range of each air vent, and the spatial division of the preset windshield area.

[0090] The target operating parameter determination method provided in this application uses the location information of the atomized area within the preset windshield area as the basis for determining the target air outlet and airflow direction, thereby achieving precise positioning and air delivery to the atomized area. This allows the defogging airflow to be directionally delivered to the actual fogging location. Compared to using a unified target air outlet and a unified target airflow direction to fully cover the entire windshield, this method effectively reduces ineffective air delivery to non-atomized areas. While ensuring the defogging effect, it reduces the energy consumption of the air conditioning system and minimizes interference with driver comfort. Furthermore, by activating multiple air outlets when the atomized area covers multiple locations and configuring corresponding target airflow directions for each, the coordinated action of airflow from multiple points improves the defogging efficiency and coverage integrity for large-scale or irregularly distributed fog.

[0091] Optionally, environmental parameters include at least one of the following: in-vehicle temperature, outside vehicle temperature, in-vehicle relative humidity, and outside vehicle relative humidity. Based on target operating parameters, including at least one of the following: air outlet mode, fan speed setting, target temperature, and air source mode, and based on the vehicle's environmental parameters and current atomization level, the target operating parameters for the air conditioning are determined, including: When the current atomization level is determined to be level three and the interior temperature is greater than the dew point temperature, the air conditioning is set to external circulation mode, fan speed to the first preset level, target temperature to reduce the current interior temperature to the first preset temperature, and air outlet mode to windshield air outlet mode.

[0092] In the above, environmental parameters include at least one of in-vehicle temperature, out-of-vehicle temperature, in-vehicle relative humidity, and out-of-vehicle relative humidity. Based on target operating parameters, including at least one of air outlet mode, fan speed setting, target temperature, and air source mode, and based on the vehicle's environmental parameters and current atomization level, the target operating parameters of the air conditioner are determined, including: When the current atomization level is determined to be level three and the interior temperature is greater than the dew point temperature, the air conditioning is set to external circulation mode, fan speed to the first preset level, target temperature to reduce the current interior temperature to the first preset temperature, and air outlet mode to windshield air outlet mode.

[0093] When the current atomization level is determined to be the fourth level, and the atomization intensity is greater than the rate of change is greater than the preset rate of change threshold, the air source mode of the air conditioner is determined to be internal circulation, the fan speed is set to the maximum level, the air outlet mode is set to the front windshield air outlet mode, and the target temperature is set to raise the current temperature inside the vehicle to the second preset temperature. The electric heating fan is then activated, wherein the second preset temperature is greater than the first preset temperature, and the maximum level is greater than the first preset level.

[0094] In the embodiments provided in this application, the third level represents that the preset windshield area is in a moderate fogging state, and the fourth level represents that the preset windshield area is in a severe fogging state. The dew point temperature is the critical temperature at which water vapor in the air begins to condense into dew under the current in-vehicle temperature and relative humidity conditions. Therefore, when the in-vehicle temperature is greater than the dew point temperature, it indicates that the water vapor in the in-vehicle air has not yet reached a saturated state. At this time, by introducing relatively dry air from outside the vehicle and appropriately lowering the in-vehicle temperature, the absolute humidity of the in-vehicle air can be reduced, thereby disrupting the thermodynamic conditions for fog to condense on the inner surface of the windshield. Therefore, for the current fogging level of the third level, the strategy of introducing dry air from the external circulation and coordinating with moderate cooling is adopted to achieve the defogging effect with lower energy consumption.

[0095] It should be noted that in the embodiments provided in this application, the rate of change of atomization intensity is used to characterize how fast the atomization intensity changes per unit time. When the rate of change is greater than the preset rate of change threshold, it indicates that the atomization degree of the preset windshield area is rapidly deteriorating. Even if the current atomization intensity has not yet reached the highest value of the fourth level, intervention is still required according to the highest defogging capability of the fourth level to prevent the atomization from spreading rapidly to the point of seriously affecting the driver's visibility in a short period of time. Therefore, for the fourth level or the rapidly deteriorating atomization condition, a strategy of internal circulation combined with a large increase in temperature and electric heating of the windshield is adopted to quickly eliminate the severe fog that has been formed with the maximum defogging power and prevent the atomization from continuing to intensify.

[0096] Understandably, when using the external circulation mode, the air conditioning system draws air from outside the vehicle. In winter or rainy weather, the relative humidity of the air outside the vehicle is usually lower than that inside the vehicle. Replacing the humid air inside the vehicle with dry air from outside can reduce the water vapor content in the air inside the vehicle from the source. Combined with the windshield vent mode, the airflow is directly delivered to the inner surface of the preset windshield area. By lowering the interior temperature to the first preset temperature, the temperature difference between the inner surface of the windshield and the interior air temperature is reduced, thereby weakening the driving force for water vapor to condense on the inner surface of the windshield. This achieves a reduction in the fogging state of the preset windshield area from level three to level one or two with lower energy consumption.

[0097] When the internal circulation mode is used with the maximum fan speed, the interior temperature is raised to the second preset temperature, and the electrically heated windshield is activated, the air conditioning system recirculates and continuously heats the air inside the vehicle. The high-temperature, dry airflow is directly delivered to the inner surface of the preset windshield area through the windshield vent mode, raising the temperature of the inner surface of the windshield to above the dew point temperature of the air inside the vehicle. At the same time, the electrically heated windshield directly heats the windshield glass physically. The temperature of the windshield glass is raised from both the inner and outer heat sources, causing the water droplets that have condensed on the windshield surface to evaporate quickly. Furthermore, the high-temperature airflow further enhances the water vapor carrying capacity of the air near the inner surface of the windshield, allowing the evaporated water vapor to be carried away by the airflow without recondensing. This reduces the fogging state of the preset windshield area from level four or rapidly deteriorating state to a safe level in the shortest possible time.

[0098] Among them, the first preset temperature, the second preset temperature, the first preset gear, the maximum gear, and the preset rate of change threshold are all preset boundary values ​​of air conditioning operating parameters. The first preset temperature and the second preset temperature correspond to the target temperature values ​​applicable to the third level and the fourth level, respectively. The first preset gear is the air volume gear corresponding to the third level, and the maximum gear is the highest air volume gear that the air conditioning system can provide. The second preset temperature is greater than the first preset temperature to ensure that a faster demisting rate is achieved through a higher heating temperature under the fourth level condition. The maximum gear is greater than the first preset gear to ensure that a faster demisting rate is achieved through a larger air supply volume under the fourth level condition.

[0099] This application configures differentiated target operating parameters for different fogging levels and environmental parameter combinations to precisely match the defogging strategy with the actual severity of the current fogging and its thermodynamic causes. For level three conditions where the in-vehicle temperature is higher than the dew point, an external circulation strategy combined with moderate cooling is adopted. This utilizes relatively dry outside air to replace the moisture inside the vehicle and reduce the temperature difference between the inside and outside of the windshield, achieving a defogging effect with lower energy consumption and avoiding energy waste caused by relying solely on heating for defogging. For level four conditions or conditions where fogging worsens rapidly, a combined strategy of internal circulation, significant heating, and electrically heated windshield is adopted. Through the synergistic effect of continuous heating of the inner surface of the windshield by high-temperature airflow and direct heating of the windshield glass by the electrically heated windshield, a rapid defogging response is achieved under extreme fogging conditions, effectively ensuring the speed of recovery of driving visibility and driving safety.

[0100] Optionally, the method also includes: During the defogging process, updated image data after defogging the preset windshield area is acquired; based on the updated image data, the clarity status of the preset windshield area is determined; when the clarity status meets the preset recovery conditions, an air conditioning recovery command is generated; wherein, the air conditioning recovery command is used to instruct the air conditioner to stop defogging the preset windshield area and exit the defogging mode.

[0101] In the embodiments provided in this application, during the process of the air conditioning system performing defogging on a preset windshield area according to the air conditioning control command, the vehicle-mounted camera continuously acquires updated image data of the preset windshield area after defogging at a preset acquisition frequency. Then, the updated image data is input into a trained fogging recognition network model. The trained fogging recognition network model performs pixel-by-pixel semantic segmentation processing on the updated image data, determines the update label corresponding to each sub-region in the updated image data, and determines the updated fogging area and updated fogging intensity of the preset windshield area after defogging based on the update labels corresponding to each sub-region, thereby determining the clarity state of the preset windshield area. When the aforementioned clarity state meets the preset recovery conditions, an air conditioning recovery command is generated. The air conditioning recovery command instructs the air conditioner to stop defogging the preset windshield area. The air conditioning recovery command is based on pre-set criteria for determining whether the preset windshield area has been restored to a clear state. The criteria include that the fogging intensity is less than a fourth preset threshold, and the duration of the fogging intensity being less than the fourth preset threshold is greater than or equal to a preset duration threshold. The fourth preset threshold is less than a first intensity threshold. The duration being greater than or equal to the preset duration threshold is used to ensure that the clarity state of the preset windshield area is stable, avoiding misjudgment as a restored clear state due to image noise or occasional local slight fogging.

[0102] It should be noted that the air conditioning recovery command carries information to instruct the air conditioning system to restore the air conditioning operating parameters before the defrosting operation. The air conditioning operating parameters before the defrosting operation are the parameter values ​​of the air outlet mode, fan speed, target temperature, and air source mode used by the air conditioning system before receiving the air conditioning control command. The air conditioning recovery command causes the air conditioning system to exit the defrosting mode and restore the operating state before defrosting, so as to avoid the air conditioning system directly shutting down or switching to an operating mode that does not match the driver's preference after exiting the defrosting mode.

[0103] This application enables the air conditioning system to accurately grasp the actual effect of the defogging action by continuously acquiring and updating image data and calculating the clarity status of the preset windshield area in real time during the defogging process. This reduces the premature cessation of defogging by the air conditioning system before the fogging is completely cleared, and also reduces the energy waste and cabin temperature interference caused by the air conditioning continuing to run the defogging mode after the fogging has been cleared.

[0104] For example, the method further includes: obtaining user preference settings information before generating air conditioning control commands; when the user preference settings information contains a disallowed automatic temperature adjustment indicator, generating a prompt message to replace the air conditioning control commands, the prompt message being used to suggest that the driver manually turn on the defroster.

[0105] This application enables the air conditioning system to automatically return to the user's preferred operating state after defogging is stopped by carrying the air conditioning operating parameter information before defogging in the recovery command, thus achieving a seamless connection between defogging intervention and user personalization settings and improving the user experience.

[0106] Optionally, the sharpness status can be determined to meet the preset recovery conditions in the following ways: When the degree of fogging corresponding to the updated image data is determined to be less than the fourth intensity threshold, the sharpness state is determined to meet the preset recovery conditions.

[0107] This application effectively reduces the interference of single-frame misjudgment or instantaneous fluctuations on the recovery judgment result by setting dual recovery conditions: the fogging intensity is less than the fourth preset threshold and the duration reaches the preset duration threshold. This improves the reliability and stability of the recovery judgment.

[0108] Optionally, before determining the fogging state information within the preset windshield area based on image data and a trained fogging recognition model, the method further includes: Based on external rainfall sensor data, determine whether the current weather is rainy or snowy; when the current weather is determined to be rainy or snowy, identify occlusions in the image data and identify dynamic occlusions in the image data; when the motion characteristics of the dynamic occlusions meet the preset raindrop motion characteristics, exclude the corresponding pixel area of ​​the dynamic occlusions from the preset windshield area.

[0109] In the embodiments provided in this application, before inputting image data into the fogging recognition model to determine fogging state information, rainfall sensing data is acquired through an external rain sensor, and the current weather is determined to be rainy or snowy based on the rainfall sensing data. When the current weather is determined to be rainy or snowy, an occlusion recognition operation is performed on the image data to identify dynamic occlusions in the image data. Dynamic occlusions are image occlusion areas formed by raindrops or snowflakes adhering to or sliding off the outer surface of a preset windshield area. When the motion characteristics of the dynamic occlusions meet the preset raindrop motion characteristics, before inputting the image data into the fogging recognition model, the pixel area corresponding to the dynamic occlusions in the image data is marked as a raindrop interference area, so that the fogging recognition model ignores the raindrop interference area when performing semantic segmentation on the image data.

[0110] It should be noted that the external rain sensor is a rain sensing device independently installed in the windshield area of ​​the vehicle. The rain sensing data is used to directly characterize whether there is precipitation and the intensity of precipitation. The determination of whether the current weather is rainy or snowy is based on whether the rain sensing data is greater than a preset rain threshold. When the rain sensing data is greater than the preset rain threshold, the current weather is determined to be rainy or snowy; otherwise, the current weather is determined to be non-rainy or non-snowy. The obstruction recognition operation is based on the detection of moving targets in the image data using image processing algorithms. The recognition of dynamic obstructions is based on the periodic or directional changes in pixel intensity or texture features in the image data over time.

[0111] Understandably, in rainy or snowy weather conditions, raindrops or snowflakes adhere to the outer surface of a pre-defined windshield area, appearing as locally blurred areas or semi-transparent moving patches in image data. If image data containing raindrop interference areas is directly input into a fog recognition model for semantic segmentation, the fog recognition model may misclassify the blurred areas formed by raindrops or snowflakes as fog areas on the inner surface of the pre-defined windshield area, thus producing false fog area detection results. Therefore, when the current weather is determined to be rainy or snowy, before inputting image data into the fog recognition model, dynamic occlusions corresponding to raindrops or snowflakes are identified through dynamic detection, and interference markers are added to the dynamic occlusions. This allows the fog recognition model to exclude pixel areas marked with interference markers from the classification range during semantic segmentation, thereby preventing the fog recognition model from misclassifying raindrops or snowflakes as fog.

[0112] The preset raindrop motion characteristics include the motion direction of the dynamic obstruction on the outer surface of the preset windshield area being either falling along the direction of gravity or sliding along the direction of airflow. The motion trajectory of the dynamic obstruction has directional consistency and velocity continuity in multiple consecutive frames of image data, and the shape of the dynamic obstruction is teardrop-shaped or striped. When the motion characteristics of the dynamic obstruction match the preset raindrop motion characteristics, the dynamic obstruction is determined to be a raindrop interference area.

[0113] In this process, after marking dynamic occlusions as raindrop interference areas, the remaining areas in the image data other than the raindrop interference areas are retained as valid analysis areas. The fog recognition model only performs semantic segmentation processing on the valid analysis areas. After marking is completed, the marked image data is input into the fog recognition model, and the fog recognition model determines the fog state information within the valid analysis areas.

[0114] In the above-mentioned rain and snow interference elimination method provided in this application, by determining whether there is precipitation based on external rain sensor data before fogging recognition, the obstruction recognition operation is triggered only under rain and snow conditions, which reduces the invalid computing overhead caused by performing obstruction recognition in sunny weather, thereby reducing the computing load of the vehicle controller.

[0115] Optionally, the method also includes: When the position change of the atomized area within a preset time window is less than a preset position change threshold, and the change in atomization intensity is less than a preset intensity change threshold, the atomized area is marked as a dirty area, and a cleaning prompt message is generated.

[0116] In the embodiments provided in this application, after the fogging state information is determined, the fogging area and fogging intensity corresponding to each of the consecutive frames of image data are obtained. The position change of the fogging area within a preset time window is determined based on the fogging area corresponding to the consecutive frames of image data, and the change of fogging intensity within a preset time window is determined based on the fogging intensity corresponding to the consecutive frames of image data. When the position change is less than a preset position change threshold and the change of fogging intensity is less than a preset intensity change threshold, the fogging area is marked as a dirty area, and a cleaning prompt message is generated. The cleaning prompt message is used to prompt the driver to perform physical cleaning operations on the preset windshield area.

[0117] It should be noted that in the embodiments provided in this application, the preset time window is a pre-set time length used to limit the effective time period for calculating the changes in position and atomization intensity.

[0118] The time span between the start and end times of the time window is no less than the reciprocal of the preset frame rate of the image data acquired by the vehicle-mounted camera multiplied by the preset number of frames.

[0119] It is understandable that the actual fog in the preset windshield area is formed by the condensation of water vapor in the air inside the vehicle on the inner surface of the windshield. The intensity and coverage of this condensation are directly affected by changes in the temperature and humidity conditions inside the vehicle. When the air conditioning system performs defogging, the area and position of the fogged area will shrink or shift due to the drying or heating airflow delivered by the air conditioning system. Therefore, the changes in position and fogging intensity usually show significant variations within the preset time window. In contrast, physical contaminants such as dust or oil films adhere to the outer or inner surface of the preset windshield area; their attachment position is fixed and does not change with the vehicle's position. The fogging effect is not affected by changes in internal temperature and humidity conditions, nor will it disappear or move due to the defogging action of the air conditioning system. Therefore, when the fogging area is actually obstructed by physical dirt such as dust or oil film, the corresponding change in position approaches zero within a preset time window, and the corresponding change in fogging intensity also approaches zero. Therefore, by setting preset position change thresholds and preset intensity change thresholds, fogging areas with position change less than the preset position change threshold and fogging intensity change less than the preset intensity change threshold are identified as dirty areas, which can effectively distinguish between real fog obstruction and physical dirt obstruction.

[0120] Here, in the embodiments provided in this application, after the atomized area is marked as a dirty area, the atomized area is excluded in the subsequent atomization state information determination process, that is, the pixel area corresponding to the dirty area is no longer included in the calculation range of the atomized area and atomization intensity; the cleaning prompt information is output through the vehicle's central control display screen or instrument panel in at least one of the following ways: text prompt, icon prompt or voice prompt, to inform the driver that there are physical stains in the preset windshield area and to suggest cleaning.

[0121] This application achieves accurate differentiation between physical dirt such as dust and oil film and actual fog, effectively reducing the occurrence of misjudging physical dirt as fog and triggering invalid air conditioning defogging operations. This reduces energy waste and meaningless operation of the air conditioning system caused by misjudgment. At the same time, by generating cleaning prompts, it actively reminds the driver to physically clean the preset windshield area, helping the driver to promptly detect and remove physical dirt on the windshield surface. This eliminates the hidden danger of dirt obstructing the driver's vision from the source, further improving driving safety. In addition, this application can exclude pixel areas marked as dirty from the calculation of subsequent fogging state information, reducing the continuous interference of physical dirt on the fogging intensity calculation results and improving the accuracy of fogging state information determination.

[0122] For example, after generating the air conditioning control command, this application sends the air conditioning control command to the air conditioning electronic control unit through the controller local area network. The air conditioning control command includes a mode field, an air volume field, a temperature field, and an air source field. The mode field is used to indicate the air outlet mode, the air volume field is used to indicate the air volume level, the temperature field is used to indicate the target temperature, and the air source field is used to indicate the air source mode.

[0123] The mode field occupies 2 bits, the air volume field occupies 3 bits, the temperature field occupies 4 bits, and the air source field occupies 1 bit. The stored value of the temperature field is the ratio of the target temperature to the temperature accuracy, with a temperature accuracy of 0.5℃.

[0124] For example, the embodiments provided in this application can also obtain the vehicle's driving status information and determine the air source mode based on the driving status information, wherein when the vehicle is in motion, the external circulation mode is selected first.

[0125] Here, driving status information is used to indicate whether the vehicle is in motion.

[0126] Figure 2 This diagram illustrates a structural block diagram of a vehicle air conditioning defogging device according to an embodiment of this application. Figure 2 As shown, the vehicle's air conditioning defrosting device 200 includes: The auxiliary acquisition module 210 is used to acquire the current light intensity of the external environment; when the current light intensity is less than the preset light threshold intensity, it controls the lighting equipment inside the vehicle to turn on in order to assist in acquiring image data of the preset windshield area inside the vehicle.

[0127] The acquisition module 220 is used to acquire image data of a preset windshield area inside the vehicle after the vehicle is powered on.

[0128] The first determining module 230 is used to determine the fogging state information within a preset windshield area based on image data and a trained fogging recognition model; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area.

[0129] The second determining module 240 is used to determine the current atomization level of the preset windshield area based on the atomization status information.

[0130] The generation module 250 is used to generate air conditioning control commands based on the current atomization level, and control the air conditioning in the vehicle to defog the preset windshield area according to the air conditioning control commands; wherein, the air conditioning control commands are used to instruct the air conditioning to perform defogging actions that match the current atomization level.

[0131] The update module 260 is used to acquire updated image data after defogging the preset windshield area during the defogging process; determine the clarity status of the preset windshield area based on the updated image data; and generate an air conditioning recovery command when the clarity status meets the preset recovery conditions. The air conditioning recovery command is used to instruct the air conditioner to stop defogging the preset windshield area and exit the defogging mode.

[0132] The occlusion recognition module 270 is used to determine whether the current weather is rainy or snowy based on external rainfall sensing data; when the current weather is determined to be rainy or snowy, it performs occlusion recognition on the image data to identify dynamic occlusions in the image data; when the motion characteristics of the dynamic occlusions meet the preset raindrop motion characteristics, the corresponding pixel area of ​​the dynamic occlusions is excluded from the preset windshield area.

[0133] The dirty area identification module 280 is used to mark the atomized area as a dirty area and generate cleaning prompt information when the position change of the atomized area within a preset time window is less than a preset position change threshold and the change of atomization intensity is less than a preset intensity change threshold.

[0134] Optionally, the first determining module 230 includes: The first determination submodule is used to input image data into the trained fog recognition network model to determine the fog area under the preset windshield area.

[0135] The second determining submodule is used to determine the global fogging index of the fogging area based on the area ratio of the fogging area in the preset windshield area and the average opacity of the fogging area, and to use the global fogging index as the fogging intensity corresponding to the fogging area.

[0136] Optionally, the first determining module 230 further includes: The trained fogging recognition model performs semantic segmentation on the image data to determine the label corresponding to each sub-region in the image data; the label is used to characterize whether fogging exists on the surface of the sub-region.

[0137] Based on the label corresponding to each sub-region in the image data, the fogging area under the preset windshield area is determined.

[0138] Optionally, the fog recognition model trained in the first determining module 230 is obtained by training in the following manner: A sample training set is obtained, which includes sample image data of multiple preset windshield areas; each sample image data contains corresponding standard fogging state information; the sample image data is input into an initial fogging recognition model to obtain predicted labels corresponding to multiple sub-regions; based on the standard fogging state information, the standard label corresponding to each sub-region is determined; based on the difference between the predicted labels and standard label information corresponding to multiple sub-regions, the training loss is determined; based on the training loss, the parameters of the initial fogging recognition model are adjusted to obtain a trained fogging recognition model.

[0139] Optionally, the second determining module 240 includes: The third determining submodule is used to determine the current atomization level of the preset windshield area as the first level when the atomization intensity is less than the first intensity threshold. The first level indicates that the preset windshield area is in a clear state. The fourth determination submodule is used to determine the current fogging level of the preset windshield area as the second level when the fogging degree is greater than or equal to the first intensity threshold and less than the second intensity threshold. The second level indicates that the preset windshield area is in a slightly fogged state. The fifth determination submodule is used to determine the current fog level of the preset windshield area as the third level when the fog level is greater than or equal to the second intensity threshold and less than the third intensity threshold. The third level indicates that the preset windshield area is in a moderate fog state. The sixth determination submodule is used to determine the current fogging level of the preset windshield area as the fourth level when the fogging degree is greater than or equal to the third intensity threshold. The fourth level indicates that the preset windshield area is in a state of severe fogging. The third intensity threshold is greater than the second intensity threshold, and the second intensity threshold is greater than the first intensity threshold.

[0140] Optionally, the third determining module 240 includes: The seventh determination submodule is used to determine the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level.

[0141] The generation module is used to generate air conditioning control commands based on the target operating parameters.

[0142] Optionally, the air conditioner includes multiple air outlets, each with at least two air outlet directions. The target operating parameters include the target air outlet direction. The seventh determination submodule includes: The first determining subunit is used to determine the target air outlet and the target air outlet direction from multiple air outlets based on the vehicle's environmental parameters, the current atomization level, and the position of the atomization area in the preset windshield area.

[0143] Optionally, environmental parameters include at least one of the following: in-vehicle temperature, outside vehicle temperature, in-vehicle relative humidity, and outside vehicle relative humidity. Based on target operating parameters, including at least one of the following: air outlet mode, fan speed setting, target temperature, and air source mode, the seventh determining submodule includes: The second determining subunit is used to determine the air source mode of the air conditioner as external circulation, the fan speed as the first preset speed, the target temperature as reducing the current temperature inside the vehicle to the first preset temperature, and the air outlet mode as the windshield air outlet mode when the current atomization level is determined to be the third level and the temperature inside the vehicle is greater than the dew point temperature.

[0144] The third determining subunit is used to determine the air source mode of the air conditioner as internal circulation, the fan speed as maximum, the air outlet mode as front windshield air outlet mode, and the target temperature as raising the current temperature inside the vehicle to the second preset temperature when the current atomization level is determined to be the fourth level and the atomization intensity is greater than the preset change rate threshold. The electric heating fan is then activated. The second preset temperature is greater than the first preset temperature, and the maximum speed is greater than the first preset speed.

[0145] Compared with the prior art, the vehicle air conditioning defogging device 200 provided in this application acquires image data of a preset windshield area inside the vehicle after the vehicle is powered on. Based on the image data and a trained fogging recognition model, it determines the fogging state information of the preset windshield area, then determines the current fogging level of the preset windshield area based on the fogging state information, and generates an air conditioning control command based on the current fogging level. The air conditioning in the vehicle is then controlled to defog the preset windshield area according to the air conditioning control command. This application can realize timely and early identification of the fogging state of the car window and control the air conditioning in the vehicle to perform intelligent defogging, reducing the time lag of air conditioning defogging and thus improving the user experience.

[0146] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Figure 3 As shown, the electronic device 300 includes a processor 310, a memory 320, and a bus 330.

[0147] Memory 320 stores machine-readable instructions executable by processor 310. When electronic device 300 is running, processor 310 and memory 320 communicate via bus 330. When the machine-readable instructions are executed by processor 310, they can perform the operations described above. Figure 1 The steps of the vehicle air conditioning defogging method in the method embodiment shown are specifically implemented in the method embodiment and will not be repeated here.

[0148] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can perform the above-described actions. Figure 1 The steps of the vehicle air conditioning defogging method in the method embodiment shown are specifically implemented in the method embodiment and will not be repeated here.

[0149] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0150] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0151] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-readable program code.

[0152] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create a machine for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0153] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0154] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0155] This application also provides a computer program product, which includes computer software instructions that, when executed on a processing device, cause the processing device to perform a vehicle air conditioning defogging method.

[0156] A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0157] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0158] In the several embodiments provided in this application, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between devices or units, and may be electrical, mechanical, or other forms.

[0159] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0160] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0161] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0162] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

[0163] Although preferred embodiments have been described in this specification, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this specification.

[0164] Obviously, those skilled in the art can make various modifications and variations to this specification without departing from its spirit and scope. Therefore, if such modifications and variations fall within the scope of the claims and their equivalents, this specification is also intended to include such modifications and variations.

Claims

1. A method for defogging a vehicle's air conditioning system, characterized in that, The vehicle's air conditioning defogging method includes: After the vehicle is powered on, image data of a preset windshield area inside the vehicle is acquired; Based on the image data and the trained fogging recognition model, the fogging state information within the preset windshield area is determined; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area; Based on the atomization status information, the current atomization level of the preset windshield area is determined; Based on the current fogging level, an air conditioning control command is generated, and the air conditioning in the vehicle is controlled to defog the preset windshield area according to the air conditioning control command; wherein, the air conditioning control command is used to instruct the air conditioning to perform a defogging action that matches the current fogging level.

2. The vehicle air conditioning defogging method according to claim 1, characterized in that, The step of determining the fogging state information within the preset windshield area based on the image data and the trained fogging recognition network model includes: The image data is input into a trained fog recognition network model to determine the fog area under the preset windshield area; Based on the area ratio of the atomized region in the preset windshield area and the average opacity of the atomized region, the global atomization index of the atomized region is determined, and the global atomization index is used as the atomization intensity corresponding to the atomized region.

3. The vehicle air conditioning defogging method according to claim 2, characterized in that, The step of inputting the image data into the trained fogging recognition network model to determine the fogging area under the preset windshield area includes: The trained fogging recognition model is used to perform semantic segmentation on the image data to determine the label corresponding to each sub-region in the image data; the label is used to characterize whether fogging exists on the surface of the sub-region. Based on the label corresponding to each sub-region in the image data, the fogging area under the preset windshield area is determined.

4. The vehicle air conditioning defogging method according to claim 1, characterized in that, The trained fog recognition model was obtained through the following method: Obtain a sample training set, wherein the sample training set includes sample image data of multiple preset windshield areas; each sample image data contains corresponding standard fogging state information; The sample image data is input into the initial fog recognition model to obtain the predicted labels corresponding to multiple sub-regions; Based on standard atomization state information, determine the standard label corresponding to each of the sub-regions; The training loss is determined based on the differences between the predicted labels and standard labels corresponding to the multiple sub-regions. The parameters of the initial fog recognition model are adjusted based on the training loss to obtain a trained fog recognition model.

5. The vehicle air conditioning defogging method according to claim 2, characterized in that, Determining the current atomization level of the preset windshield area based on the atomization state information includes: When the atomization intensity is less than the first intensity threshold, the current atomization level of the preset windshield area is determined to be the first level, wherein the first level indicates that the preset windshield area is in a clear state. When the degree of fogging is greater than or equal to the first intensity threshold and less than the second intensity threshold, the current fogging level of the preset windshield area is determined to be the second level, wherein the second level indicates that the preset windshield area is in a slightly fogged state; When the degree of atomization is greater than or equal to the second intensity threshold and less than the third intensity threshold, the current atomization level of the preset windshield area is determined to be the third level, wherein the third level indicates that the preset windshield area is in a moderate atomization state. When the degree of fogging is greater than or equal to the third intensity threshold, the current fogging level of the preset windshield area is determined to be the fourth level, wherein the fourth level indicates that the preset windshield area is in a state of severe fogging; the third intensity threshold is greater than the second intensity threshold, and the second intensity threshold is greater than the first intensity threshold.

6. The vehicle air conditioning defogging method according to claim 5, characterized in that, The step of generating air conditioning control commands based on the current atomization level includes: Based on the vehicle's environmental parameters and the current atomization level, the target operating parameters of the air conditioner are determined; Based on the target operating parameters, air conditioning control commands are generated.

7. The vehicle air conditioning defogging method according to claim 6, characterized in that, The air conditioner includes multiple air outlets, each of which has at least two air outlet directions. The target operating parameters include the target air outlet direction. Determining the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level includes: Based on the vehicle's environmental parameters, the current atomization level, and the position of the atomization area in the preset windshield area, a target air outlet is determined from multiple air outlets, as well as the target air outlet's target air outlet direction.

8. The vehicle air conditioning defogging method according to claim 6, characterized in that, The environmental parameters include at least one of the following: in-vehicle temperature, outside vehicle temperature, in-vehicle relative humidity, and outside vehicle relative humidity. The target operating parameters include at least one of the following: air outlet mode, fan speed setting, target temperature, and air source mode. Determining the target operating parameters of the air conditioner based on the vehicle's environmental parameters and the current atomization level includes: When the current atomization level is determined to be the third level and the in-vehicle temperature is greater than the dew point temperature, the air source mode of the air conditioner is determined to be external circulation, the fan speed is determined to be the first preset speed, the target temperature is to reduce the current in-vehicle temperature to the first preset temperature, and the air outlet mode is determined to be the windshield air outlet mode.

9. The vehicle air conditioning defogging method according to claim 8, characterized in that, The determination of the target operating parameters for the air conditioner based on the vehicle's environmental parameters and current atomization level includes: When the current atomization level is determined to be the fourth level, and the atomization intensity is greater than the rate of change of a preset rate of change threshold, the air source mode of the air conditioner is determined to be internal circulation, the fan speed is determined to be the maximum speed, the air outlet mode is determined to be the front windshield air outlet mode, and the target temperature is determined to raise the current temperature inside the vehicle to the second preset temperature, and the electric heating fan is activated, wherein the second preset temperature is greater than the first preset temperature, and the maximum speed is greater than the first preset speed.

10. The vehicle air conditioning defogging method according to claim 5, characterized in that, The method further includes: During the defogging process, updated image data of the preset windshield area after defogging is acquired; Based on the updated image data, the clarity status of the preset windshield area is determined; When the clarity status meets the preset recovery conditions, an air conditioning recovery command is generated; wherein, the air conditioning recovery command is used to instruct the air conditioner to stop defogging the preset windshield area and exit the defogging mode.

11. The vehicle air conditioning defogging method according to claim 10, characterized in that, The sharpness state is determined to meet the preset recovery conditions using the following method: When the degree of fogging corresponding to the updated image data is determined to be less than the fourth intensity threshold, the clarity state is determined to meet the preset recovery condition.

12. The vehicle air conditioning defogging method according to claim 1, characterized in that, Before determining the fogging state information within the preset windshield area based on the image data and the trained fogging recognition model, the method further includes: Based on external rainfall sensor data, determine whether the current weather is rainy or snowy; When the current weather is determined to be rainy or snowy weather, the image data is subjected to occlusion identification to determine the dynamic occlusions in the image data; When the motion characteristics of the dynamic occlusion satisfy the preset raindrop motion characteristics, the corresponding pixel area of ​​the dynamic occlusion is excluded from the preset windshield area.

13. The vehicle air conditioning defogging method according to claim 2, characterized in that, The method further includes: When the position change of the atomized area within a preset time window is less than a preset position change threshold, and the change in atomization intensity is less than a preset intensity change threshold, the atomized area is marked as a dirty area, and a cleaning prompt message is generated.

14. The vehicle air conditioning defogging method according to claim 1, characterized in that, Before acquiring image data of a preset windshield area inside the vehicle, the method further includes: Obtain the current light intensity of the external environment; When the current light intensity is less than the preset light threshold intensity, the lighting equipment inside the vehicle is turned on to assist in acquiring the image data of the preset windshield area inside the vehicle.

15. A vehicle air conditioning defrosting device, characterized in that, The vehicle's air conditioning defrosting device includes: The acquisition module is used to acquire image data of a preset windshield area inside the vehicle after the vehicle is powered on; The first determining module is used to determine the fogging state information within the preset windshield area based on the image data and the trained fogging recognition model; wherein, the fogging state information includes the fogging area within the preset windshield area and the fogging intensity corresponding to the fogging area; The second determining module is used to determine the current atomization level of the preset windshield area based on the atomization state information; The generation module is used to generate air conditioning control instructions based on the current atomization level, and control the air conditioning in the vehicle to defog the preset windshield area according to the air conditioning control instructions; wherein, the air conditioning control instructions are used to instruct the air conditioning to perform a defogging action that matches the current atomization level.

16. An electronic device, characterized in that, include: The device includes a processor, a memory, and a bus. The memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor communicates with the memory via the bus, and the machine-readable instructions are executed by the processor to perform the steps of the air conditioning defogging method for a vehicle as described in any of claims 1-14.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the vehicle air conditioning defogging method as described in any of claims 1-14.

18. A computer program product comprising a computer program or computer-executable instructions, characterized in that, When the computer program or computer-executable instructions are executed by a processor, the steps of the air conditioning defogging method for a vehicle as described in any one of claims 1 to 14 are implemented.

Images