Automatic defrosting system and method for new energy automobile

Abstract

The invention relates to the technical field of automobile defrosting, and discloses a new energy automobile automatic defrosting system and method.The method comprises the steps that a first image and a second image of a windshield are obtained in response to a flameout instruction and a starting instruction adjacent to a target automobile; in response to a defrosting instruction of the target vehicle, edge detection is performed on the first image and the second image, and a frosting contour in the second image is determined; projecting a uniform light curtain to the windshield to obtain a third image; segmenting the first image and the third image based on the frosting contour to obtain a feature matrix; dividing a characteristic area, corresponding to the windshield, of the characteristic matrix into M sub-areas with the same size, and determining the frosting thickness of each sub-area based on the characteristic matrix; and based on the frosting thickness of each sub-area, the defrosting angles and the defrosting power of the N defrosting units are dynamically controlled, so that the defrosting time is shortened.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile defrosting, and more particularly to an automatic defrosting system and method for a new energy vehicle. Background Art

[0002] In winter, low temperatures can easily cause frost to form on the windshield surface after a car has been parked for a long time, requiring the car to be defrosted before driving. However, existing defrosting systems typically defrost the entire windshield evenly, resulting in slow defrosting and long waiting times for users.

[0003] Reference publication number CN116442950A discloses an automatic defrosting method, device, and system for a pure electric vehicle. The method comprises receiving a defrost signal and determining whether the plug-in defrost conditions are met; upon determining that the plug-in defrost conditions are met, determining whether the preconditions for plug-in defrost are complete; upon determining that the preconditions for plug-in defrost are complete, limiting the power of the air conditioner and water heater based on the current maximum output power allowed by the onboard charger and executing automatic defrost; and upon determining that the automatic defrost exit conditions are met, exiting the automatic defrost. This method achieves automatic defrost while maximizing battery life and enhancing the user experience.

[0004] A comprehensive analysis of the above-mentioned existing technologies reveals that they control the air conditioner and water heater to perform defrost based on the logical relationship between the defrost signal and the defrost conditions. However, this approach fails to account for uneven frost thickness, for example, defrosting thicker and thinner frost layers differently, resulting in extended defrost times. Summary of the Invention

[0005] The present invention provides an automatic defrosting system and method for a new energy vehicle, which solves the technical problems raised in the background technology.

[0006] In a first aspect, the present invention provides an automatic defrosting method for a new energy vehicle, comprising:

[0007] Step 1, in response to an adjacent shutdown instruction and a startup instruction of a target vehicle, respectively acquiring a first image and a second image of a windshield;

[0008] Step 2, in response to a defrost instruction of the target vehicle, performing edge detection on the first image and the second image respectively, and determining a frosting contour in the second image;

[0009] Step 3: Projecting a uniform light curtain onto the windshield to obtain a third image; and segmenting the first image and the third image based on the frosting contour to obtain a feature matrix;

[0010] Step 4: Divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix;

[0011] Step 5: Based on the frost thickness of each sub-area, dynamically control the defrost angle and defrost power of the N defrost units to shorten the defrost time.

[0012] Furthermore, determining the frosting contour in the second image in step 2 includes:

[0013] Performing edge detection on the first image and the second image to obtain a first edge matrix and a second edge matrix; wherein the first edge matrix and the second edge matrix are both Boolean matrices, and element 1 and element 0 represent edge features and non-edge features;

[0014] Calculate the frosting profile as follows:

[0015] F sp =ReLU(F2-F1); where F sp represents the frosting contour, ReLU represents the ReLU activation function, F1 represents the first edge matrix, and F2 represents the second edge matrix.

[0016] Furthermore, the projection of a uniform light curtain toward the windshield in step 3 is achieved based on the AR-HUD projection unit of the target vehicle.

[0017] Furthermore, the feature matrix in step 3 includes:

[0018] Fitting the frosting contour to the third image, retaining pixels of the third image located inside the frosting contour, to obtain a first initial feature image;

[0019] Fitting the frosting contour to the first image, retaining pixels of the first image located inside the frosting contour, to obtain a second initial feature image;

[0020] The first initial feature image and the second initial feature image are grayscaled respectively, and then the difference is calculated to obtain a feature matrix.

[0021] Furthermore, determining the frosting thickness of each sub-region based on the characteristic matrix in step 4 includes:

[0022] Build a frost thickness prediction model, including sample data and sample labels;

[0023] Repeat steps 1 to 3 for K times in the historical period to obtain K feature matrices.

[0024] Get the normalized element value of the element in the xth row and yth column of each feature matrix as sample data;

[0025] Get the normalized frost thickness of the windshield position corresponding to the element in the xth row and yth column of each feature matrix as the sample label;

[0026] Based on the sample data and sample labels, a frost thickness prediction model is trained;

[0027] The frost thickness prediction model is built based on the support vector machine, and the hyperparameters of the frost thickness prediction model are updated through back propagation of the mean square error loss function.

[0028] Within a preset time period, based on the frost thickness prediction model and the feature matrix, the frost thickness at each corresponding position of the windshield is obtained;

[0029] Obtain the frost thickness at each location in the mth sub-area and calculate the average value to obtain the frost thickness of the mth sub-area; where 1≤m≤M, and m is a positive integer.

[0030] Furthermore, the step 5 of dynamically controlling the defrost angles and defrost powers of the N defrost units includes:

[0031] Step 61, initializing and generating defrost individuals that meet the constraints; wherein the defrost individuals include: defrost angles G and defrost powers P of N defrost units;

[0032] Step 62, the constraints include: G min ≤G≤G max , P min ≤P≤P max Among them, G min and G max Represents the minimum defrost angle and the maximum defrost angle, P min and P max Respectively represent the minimum defrost power and the maximum defrost power;

[0033] Step 63, the steps for obtaining the fitness value of the defrost individual are as follows:

[0034] Step 631: defrost the windshield based on the defrosting individual, and acquire a third image at a fixed time interval to obtain the frost thickness of M sub-regions at each moment;

[0035] Step 632 , sorting the sub-regions based on the frost thickness at the initial moment to obtain a first sorting;

[0036] Step 633 , calculating the difference in frost thickness of each sub-region at adjacent moments, and sorting the M sub-regions based on the difference to obtain a second sorting;

[0037] Step 634 , accumulating the differences between the positions of the M sub-regions in the first sorting and the second sorting, and obtaining the fitness value of the defrost individual;

[0038] Step 64: If the fitness value is less than or equal to the preset fitness threshold, the defrost individual is kept defrosting the windshield; otherwise, the defrost individual is updated based on gradient descent, and steps 61 to 63 are repeated until the fitness value is less than or equal to the preset fitness threshold.

[0039] Furthermore, the step 5 of dynamically controlling the defrost angles and defrost powers of the N defrost units further includes:

[0040] In response to feature matrices corresponding to adjacent shutdown commands and start commands of multiple groups of target vehicles, a standard database is constructed, and a defrost individual corresponding to each feature matrix in the standard database is determined;

[0041] In the target time period, obtain the target feature matrix;

[0042] Calculate the similarity between the target feature matrix and any feature matrix in the standard database, and match the target feature matrix with the feature matrix with the greatest similarity;

[0043] Obtain the defrost individuals corresponding to the matched feature matrix and apply them to the target time period for defrosting.

[0044] Furthermore, the similarity of the feature matrix includes:

[0045] Calculate the Euclidean distance between the target feature matrix and any feature matrix in the standard database;

[0046] The reciprocal of the Euclidean distance is taken as the similarity.

[0047] In a second aspect, an automatic defrosting system for a new energy vehicle is provided, which is applied to any one of the automatic defrosting methods for a new energy vehicle, and comprises:

[0048] a first acquisition module, configured to acquire a first image and a second image of the windshield in response to an adjacent shutdown instruction and a startup instruction of the target vehicle, respectively;

[0049] a second acquisition module, configured to perform edge detection on the first image and the second image respectively in response to a defrost instruction of the target vehicle, and determine a frosting contour in the second image;

[0050] a third acquisition module, configured to project a uniform light curtain toward the windshield to obtain a third image; and segment the first image and the third image based on a frosting profile to obtain a feature matrix;

[0051] a data analysis module, configured to divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix;

[0052] The glass defrost module is used to dynamically control the defrost angle and defrost power of N defrost units based on the frost thickness of each sub-area to shorten the defrost time.

[0053] The beneficial effects of the present invention are: by integrating edge detection, light curtain projection and feature matrix construction technology of images when the engine is turned off and started, accurate identification of the thickness of the frost layer on the front windshield is achieved based on the light refractive index of frost layers of different thicknesses, and differentiated defrosting strategies are adopted for each sub-area using an intelligent model, so as to quickly and accurately defrost, thereby significantly shortening the defrost time and reducing the user's defrost waiting time. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flow chart of an automatic defrosting method for a new energy vehicle according to the present invention;

[0055] Figure 2 This is a module diagram of an automatic defrosting system for new energy vehicles according to the present invention. DETAILED DESCRIPTION

[0056] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed solely to enable those skilled in the art to better understand and implement the subject matter described herein, and that the functions and arrangements of the elements discussed may be varied without departing from the scope of this specification. Various examples may omit, substitute, or add various processes or components as needed. In addition, features described with respect to some examples may also be combined in other examples.

[0057] like Figures 1 and 2 As shown, a new energy vehicle automatic defrosting method includes:

[0058] Step 1, in response to an adjacent shutdown instruction and a startup instruction of a target vehicle, respectively acquiring a first image and a second image of a windshield;

[0059] Step 2, in response to a defrost instruction of the target vehicle, performing edge detection on the first image and the second image respectively, and determining a frosting contour in the second image;

[0060] Step 3: Projecting a uniform light curtain onto the windshield to obtain a third image; and segmenting the first image and the third image based on the frosting contour to obtain a feature matrix;

[0061] Step 4: Divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix;

[0062] Step 5: Based on the frost thickness of each sub-area, dynamically control the defrost angle and defrost power of the N defrost units to shorten the defrost time.

[0063] It should be noted that the "off" command is a control signal issued when the vehicle's power system is turned off, confirming that the vehicle has stopped operating. At this point, there is no frost on the vehicle's windshield (corresponding to the first image). The "start" command is a control signal issued when the vehicle's power system is turned on, confirming that the user has started the vehicle. At this point, the vehicle may have been parked for a long time and frost may be present on the windshield (corresponding to the second image). The "defrost" command indicates that after the user starts the vehicle and visually detects frost on the vehicle's windshield, the user has already started the vehicle.

[0064] In one embodiment of the present invention, the first image, the second image, and the third image are all acquired from a vehicle's driving recorder. The driving recorder acquires images at a fixed angle and position, so the first image, the second image, and the third image are of the same size and comparable.

[0065] In one embodiment of the present invention, determining a frosting contour in the second image includes:

[0066] Performing edge detection on the first image and the second image to obtain a first edge matrix and a second edge matrix; wherein the first edge matrix and the second edge matrix are both Boolean matrices, and element 1 and element 0 represent edge features and non-edge features;

[0067] Calculate the frosting profile as follows:

[0068] F sp =ReLU(F2-F1); where F sp represents the frosting contour, ReLU represents the ReLU activation function, F1 represents the first edge matrix, and F2 represents the second edge matrix.

[0069] Specifically, if there is frost on the windshield, then there is a large area of occlusion on the windshield, so the corresponding crystal edge features can be extracted. Therefore, the unique edge features in the second image are screened out and determined to be frost outlines.

[0070] In one embodiment of the present invention, edge detection includes but is not limited to the Canny edge detection algorithm. The Canny edge detection algorithm first performs Gaussian filtering on the image to remove noise, then calculates the gradient amplitude and direction, and obtains clear edges through non-maximum suppression and double threshold detection.

[0071] In one embodiment of the present invention, the ReLU activation function sets negative values directly to zero, preserving only the positive difference edges (frosting contours) that newly appear in the second image.

[0072] Optionally, the resulting frosting profile matrix F sp Perform morphological operations (such as erosion, dilation, opening, or closing) to remove noise or broken edges and obtain a more coherent and accurate frosting outline.

[0073] In one embodiment of the present invention, projecting a uniform light curtain toward the windshield is achieved based on an AR-HUD projection unit of the target vehicle.

[0074] In one embodiment of the present invention, the AR-HUD projection unit built into the target vehicle is used to project a uniform light curtain onto the windshield. Specifically, new energy vehicles (such as the Lynk & Co 08) are typically equipped with an AR-HUD projection unit. This system is originally used to project navigation, vehicle information, or entertainment content onto the front windshield in an augmented reality manner, thereby improving the driving experience and safety. When the vehicle starts the cinema mode, the AR-HUD can project a preset uniform light curtain onto the windshield.

[0075] In detail, utilizing existing AR-HUD hardware not only reduces the cost of system integration but also gives full play to the versatility of the vehicle's existing equipment.

[0076] In one embodiment of the present invention, the feature matrix includes:

[0077] Fitting the frosting contour to the third image, retaining pixels of the third image located inside the frosting contour, to obtain a first initial feature image;

[0078] Fitting the frosting contour to the first image, retaining pixels of the first image located inside the frosting contour, to obtain a second initial feature image;

[0079] The first initial feature image and the second initial feature image are grayscaled respectively, and then the difference is calculated to obtain a feature matrix.

[0080] In detail, the frosted area of the front windshield is extracted based on the previously extracted frosted contour, thereby obtaining corresponding portions of the frosted area in the first image and the third image.

[0081] In one embodiment of the present invention, determining the frosting thickness of each sub-region based on the characteristic matrix in step 4 includes:

[0082] Build a frost thickness prediction model, including sample data and sample labels;

[0083] Repeat steps 1 to 3 for K times in the historical period to obtain K feature matrices.

[0084] Get the normalized element value of the element in the xth row and yth column of each feature matrix as sample data;

[0085] Get the normalized frost thickness of the windshield position corresponding to the element in the xth row and yth column of each feature matrix as the sample label;

[0086] Based on the sample data and sample labels, a frost thickness prediction model is trained;

[0087] The frost thickness prediction model is built based on the support vector machine, and the hyperparameters of the frost thickness prediction model are updated through back propagation of the mean square error loss function.

[0088] Within a preset time period, based on the frost thickness prediction model and the feature matrix, the frost thickness at each corresponding position of the windshield is obtained;

[0089] Obtain the frost thickness at each location in the mth sub-area and calculate the average value to obtain the frost thickness of the mth sub-area; where 1≤m≤M, and m is a positive integer.

[0090] Because frosting doesn't form uniformly across the windshield, the frost thickness varies. Consequently, different locations on the frosted windshield experience varying occlusion rates. When a uniform light curtain is projected onto the windshield, the varying occlusion rates result in varying grayscale values after grayscale conversion. Consequently, a nonlinear mapping exists between grayscale values and frost thickness.

[0091] Specifically, image acquisition (steps 1-3) is performed multiple times over the historical time period to obtain feature matrices at K different time points. These matrices reflect the grayscale differences caused by different frost occlusion rates under uniform light curtain illumination. For each feature matrix, the normalized grayscale value in the xth row and yth column is extracted as the sample data. At the same time, the actual frost thickness at the known location (which can be obtained through calibration or other measurement methods) is normalized and used as the sample label. This provides the model with a corresponding relationship between input (grayscale information) and output (frost thickness).

[0092] In detail, a support vector machine (SVM) was used to construct a frost thickness prediction model based on sample data and sample labels. SVMs are advantageous in handling nonlinear mapping relationships. Using kernel functions, they can map the original input space into a high-dimensional feature space, thereby better fitting the nonlinear relationship between grayscale values and frost thickness. Mean squared error (MSE) was used as the loss function to measure the error between the model's predicted values and the true labels. By continuously adjusting the model's hyperparameters through a backpropagation algorithm, the prediction error was gradually reduced, ultimately resulting in a highly accurate thickness prediction model.

[0093] Specifically, within a preset time period, the system uses a trained prediction model and a real-time feature matrix to predict the frost thickness at each corresponding location on the windshield. The corresponding feature region of the windshield is divided into M equal-sized subregions. For each subregion, the average of the predicted frost thicknesses at all locations within it is taken as the overall frost thickness indicator for that subregion.

[0094] In one embodiment of the present invention, dynamically controlling the defrost angles and defrost powers of N defrost units includes:

[0095] Step 61, initializing and generating defrost individuals that meet the constraints; wherein the defrost individuals include: defrost angles G and defrost powers P of N defrost units;

[0096] Step 62, the constraints include: G min ≤G≤G max , P min ≤P≤P max Among them, G min and G max Represents the minimum defrost angle and the maximum defrost angle, P min and P max Respectively represent the minimum defrost power and the maximum defrost power;

[0097] Step 63, the steps for obtaining the fitness value of the defrost individual are as follows:

[0098] Step 631: defrost the windshield based on the defrosting individual, and acquire a third image at a fixed time interval to obtain the frost thickness of M sub-regions at each moment;

[0099] Step 632 , sorting the sub-regions based on the frost thickness at the initial moment to obtain a first sorting;

[0100] Step 633 , calculating the difference in frost thickness of each sub-region at adjacent moments, and sorting the M sub-regions based on the difference to obtain a second sorting;

[0101] Step 634 , accumulating the differences between the positions of the M sub-regions in the first sorting and the second sorting, and obtaining the fitness value of the defrost individual;

[0102] Step 64: If the fitness value is less than or equal to the preset fitness threshold, the defrost individual is kept defrosting the windshield; otherwise, the defrost individual is updated based on gradient descent, and steps 61 to 63 are repeated until the fitness value is less than or equal to the preset fitness threshold.

[0103] Alternatively, the power of the N defrost units in a conventional defrost mode is obtained, for example, all Nd, and the total power is N×Nd. Furthermore, the total power of the N defrost units in the defrost method disclosed in this application is also N×Nd, which serves as an optional constraint for step 62 in this application.

[0104] In one embodiment of the present invention, the frost thickness at each location on the windshield is obtained based on a frost prediction model. The difference in frost thickness at adjacent moments is used to represent the frost dissolution rate. Thicker frosted areas should have a faster dissolution rate than thinner frosted areas. Therefore, the first and second rankings represent the initial frost thickness ranking and the dissolution rate ranking of each sub-area, respectively.

[0105] In one embodiment of the present invention, the defrost unit includes, but is not limited to, a warm air defrost device or a resistance wire heating defrost device. For example, a warm air defrost device heats the front windshield to defrost by adjusting the angle of the air outlet and the power of the warm air. A resistance wire heating defrost device heats the front windshield to defrost based on the direction and power of the electromagnetic wave transmission.

[0106] Specifically, the system first uses a frost prediction model to obtain the frost thickness at each location on the windshield. It then compares the thickness changes at each location at adjacent moments to determine the frost dissolution rate. Through continuous image acquisition and the use of the prediction model, the system obtains the frost thickness at each point on the windshield in real time. The difference in frost thickness at each location between consecutive moments is calculated as the frost dissolution rate at that location. The initial frost thickness is ranked (first ranking), and the frost dissolution rates of each sub-area are also ranked (second ranking). Since thick frost areas should theoretically have higher dissolution rates, a discrepancy between the two rankings indicates that the defrosting effect is not meeting expectations. This information can be used to evaluate and optimize the defrost control strategy. Based on these two rankings, the system calculates the fitness value of each sub-area. This is the cumulative difference between the initial thickness ranking and the dissolution rate ranking, which serves as an evaluation metric for the current defrost strategy. When the fitness value exceeds a preset threshold, the system adjusts the defrost unit parameters using algorithms such as gradient descent until the fitness value meets the requirements. This dynamic adjustment ensures that areas with thicker frost layers receive greater defrost power and a more optimal defrost angle, achieving differentiated defrosting and improving overall defrost efficiency. Based on the actual measured frost thickness and dissolution rate, the system automatically selects and adjusts the operating mode and output parameters of the corresponding defrost unit to ensure rapid frost removal.

[0107] In one embodiment of the present invention, dynamically controlling the defrost angles and defrost powers of N defrost units further includes:

[0108] In response to feature matrices corresponding to adjacent shutdown commands and start commands of multiple groups of target vehicles, a standard database is constructed, and a defrost individual corresponding to each feature matrix in the standard database is determined;

[0109] In the target time period, obtain the target feature matrix;

[0110] Calculate the similarity between the target feature matrix and any feature matrix in the standard database, and match the target feature matrix with the feature matrix with the greatest similarity;

[0111] Obtain the defrost individuals corresponding to the matched feature matrix and apply them to the target time period for defrosting.

[0112] In one embodiment of the present invention, a database is generated by storing historical feature matrices and corresponding defrost individuals. Because the shape and thickness of frost on vehicles are self-similar, within a target time period, the target feature matrix is matched with the feature matrix in the database to quickly obtain the corresponding defrost individual. Based on the defrost individual, the defrost unit is controlled to defrost the vehicle-connected windshield.

[0113] In one embodiment of the present invention, the similarity of the feature matrix includes:

[0114] Calculate the Euclidean distance between the target feature matrix and any feature matrix in the standard database;

[0115] The reciprocal of the Euclidean distance is taken as the similarity.

[0116] The system not only adjusts defrost unit parameters based on real-time feature matrices but also builds a historical standard database to further improve the response speed and accuracy of defrost control. In response to multiple sets of adjacent shutdown and start commands from target vehicles, the system collects the corresponding feature matrices and, combined with the defrost control parameters at the corresponding moments (i.e., defrost individual), stores this historical data in the database. Because the shape and thickness of frost on vehicle windshields are self-similar, the feature matrices in the database cover common frost patterns and assign a validated optimal defrost solution to each pattern. During a preset target time period, the system collects a new feature matrix. By calculating the similarity between the newly collected target feature matrix and each feature matrix in the database (using the inverse of the Euclidean distance as the similarity metric), the feature matrix with the highest similarity is selected. The matching standard feature matrix corresponds to a historically validated defrost individual—the optimal defrost parameter combination. The system directly applies this matched defrost individual to defrost control during the target time period. By dynamically adjusting the angles and power of N defrost units, the system achieves rapid and accurate windshield defrost removal. This method utilizes the self-similarity of historical data to achieve rapid parameter matching and response, reduce the complexity of real-time calculations, and improve defrosting efficiency.

[0117] In one embodiment of the present invention, the system is configured with a similarity threshold. When the similarity between the target feature matrix and the most similar feature matrix in the database falls below the threshold, the system performs an online optimization calculation to generate a corresponding target defrost individual and stores the target defrost individual and its feature matrix in the database. Otherwise, the system directly generates the defrost individual through similarity matching.

[0118] An automatic defrosting system for a new energy vehicle, applied to any one of the automatic defrosting methods for a new energy vehicle, comprises:

[0119] a first acquisition module, configured to acquire a first image and a second image of the windshield in response to an adjacent shutdown instruction and a startup instruction of the target vehicle, respectively;

[0120] a second acquisition module, configured to perform edge detection on the first image and the second image respectively in response to a defrost instruction of the target vehicle, and determine a frosting contour in the second image;

[0121] a third acquisition module, configured to project a uniform light curtain toward the windshield to obtain a third image; and segment the first image and the third image based on a frosting profile to obtain a feature matrix;

[0122] a data analysis module, configured to divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix;

[0123] The glass defrost module is used to dynamically control the defrost angle and defrost power of N defrost units based on the frost thickness of each sub-area to shorten the defrost time.

[0124] The above describes the embodiments of this embodiment, but this embodiment is not limited to the above specific implementation methods. The above specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make many forms based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. A new energy vehicle automatic defrosting method, characterized in that: include: Step 1, in response to an adjacent shutdown instruction and a startup instruction of a target vehicle, respectively acquiring a first image and a second image of a windshield; Step 2, in response to a defrost instruction of the target vehicle, performing edge detection on the first image and the second image respectively, and determining a frosting contour in the second image; Step 3: Projecting a uniform light curtain onto the windshield to obtain a third image; and segmenting the first image and the third image based on the frosting contour to obtain a feature matrix; Step 4: Divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix; Step 5: Based on the frost thickness of each sub-area, dynamically control the defrost angle and defrost power of the N defrost units to shorten the defrost time.

2. The automatic defrosting method for new energy vehicles according to claim 1, characterized in that: Determining the frosting contour in the second image in step 2 includes: Performing edge detection on the first image and the second image to obtain a first edge matrix and a second edge matrix; wherein the first edge matrix and the second edge matrix are both Boolean matrices, and element 1 and element 0 represent edge features and non-edge features; Calculate the frosting profile as follows: <h2 style=";text-align:left;direction:ltr">F<h2 style=";text-align:left;direction:ltr"> sp <h2 style=";text-align:left;direction:ltr"> (ReLU(F2-F1)) Among them, F sp represents the frosting contour, ReLU represents the ReLU activation function, F1 represents the first edge matrix, and F2 represents the second edge matrix.

3. The automatic defrosting method for new energy vehicles according to claim 1, characterized in that: Projecting a uniform light curtain toward the windshield in step 3 is achieved based on the AR-HUD projection unit of the target vehicle.

4. The automatic defrosting method for new energy vehicles according to claim 2, characterized in that: The feature matrix in step 3 includes: Fitting the frosting contour to the third image, retaining pixels of the third image located inside the frosting contour, to obtain a first initial feature image; Fitting the frosting contour to the first image, retaining pixels of the first image located inside the frosting contour, to obtain a second initial feature image; The first initial feature image and the second initial feature image are grayscaled respectively, and then the difference is calculated to obtain a feature matrix.

5. The automatic defrosting method for new energy vehicles according to claim 4, characterized in that: Determining the frosting thickness of each sub-region based on the characteristic matrix in step 4 includes: Build a frost thickness prediction model, including sample data and sample labels; Repeat steps 1 to 3 for K times in the historical period to obtain K feature matrices. Get the normalized element value of the element in the xth row and yth column of each feature matrix as sample data; Get the normalized frost thickness of the windshield position corresponding to the element in the xth row and yth column of each feature matrix as the sample label; Based on the sample data and sample labels, a frost thickness prediction model is trained; The frost thickness prediction model is built based on the support vector machine, and the hyperparameters of the frost thickness prediction model are updated through back propagation of the mean square error loss function. Within a preset time period, based on the frost thickness prediction model and the feature matrix, the frost thickness at each corresponding position of the windshield is obtained; Obtain the frost thickness at each location in the mth sub-area and calculate the average value to obtain the frost thickness of the mth sub-area; where 1≤m≤M, and m is a positive integer.

6. The automatic defrosting method for new energy vehicles according to claim 5, characterized in that: The step 5 of dynamically controlling the defrost angles and defrost powers of the N defrost units includes: Step 61, initializing and generating defrost individuals that meet the constraints; wherein the defrost individuals include: defrost angles G and defrost powers P of N defrost units; Step 62, the constraints include: G min ≤G≤G max , P min ≤P≤P max Among them, G min and G max Represents the minimum defrost angle and the maximum defrost angle, P min and P max Respectively represent the minimum defrost power and the maximum defrost power; Step 63, the steps for obtaining the fitness value of the defrost individual are as follows: Step 631: defrost the windshield based on the defrosting individual, and acquire a third image at a fixed time interval to obtain the frost thickness of M sub-regions at each moment; Step 632 , sorting the sub-regions based on the frost thickness at the initial moment to obtain a first sorting; Step 633 , calculating the difference in frost thickness of each sub-region at adjacent moments, and sorting the M sub-regions based on the difference to obtain a second sorting; Step 634 , accumulating the differences between the positions of the M sub-regions in the first sorting and the second sorting, and obtaining the fitness value of the defrost individual; Step 64: If the fitness value is less than or equal to the preset fitness threshold, the defrost individual is kept defrosting the windshield; otherwise, the defrost individual is updated based on gradient descent, and steps 61 to 63 are repeated until the fitness value is less than or equal to the preset fitness threshold.

7. The automatic defrosting method for new energy vehicles according to claim 6, characterized in that: The step 5 of dynamically controlling the defrost angles and defrost powers of the N defrost units further includes: In response to feature matrices corresponding to adjacent shutdown commands and start commands of multiple groups of target vehicles, a standard database is constructed, and a defrost individual corresponding to each feature matrix in the standard database is determined; In the target time period, obtain the target feature matrix; Calculate the similarity between the target feature matrix and any feature matrix in the standard database, and match the target feature matrix with the feature matrix with the greatest similarity; Obtain the defrost individuals corresponding to the matched feature matrix and apply them to the target time period for defrosting.

8. The automatic defrosting method for new energy vehicles according to claim 7, characterized in that: The similarity of the feature matrix includes: Calculate the Euclidean distance between the target feature matrix and any feature matrix in the standard database; The reciprocal of the Euclidean distance is taken as the similarity.

9. An automatic defrosting system for a new energy vehicle, applied to an automatic defrosting method for a new energy vehicle according to any one of claims 1 to 8, characterized in that: include: a first acquisition module, configured to acquire a first image and a second image of the windshield in response to an adjacent shutdown instruction and a startup instruction of the target vehicle, respectively; a second acquisition module, configured to perform edge detection on the first image and the second image respectively in response to a defrost instruction of the target vehicle, and determine a frosting contour in the second image; a third acquisition module, configured to project a uniform light curtain toward the windshield to obtain a third image; and segment the first image and the third image based on a frosting profile to obtain a feature matrix; a data analysis module, configured to divide the characteristic region of the windshield corresponding to the characteristic matrix into M subregions of equal size, and determine the frost thickness of each subregion based on the characteristic matrix; The glass defrost module is used to dynamically control the defrost angle and defrost power of N defrost units based on the frost thickness of each sub-area to shorten the defrost time.

Images