Intelligent cleaning method and system for engineering vehicle

Abstract

The invention discloses an intelligent cleaning method and system for an engineering vehicle, and the method comprises the steps: capturing a vehicle surface image of a current to-be-cleaned engineering vehicle, carrying out the preprocessing of the vehicle surface image, and obtaining vehicle image data; according to the vehicle image data and the trained machine learning model, obtaining a stain identification result and vehicle type information of the engineering vehicle; based on a stain identification result and vehicle type information, an optimal cleaning scheme is comprehensively calculated, and cleaning parameters are intelligently adjusted; the cleaning state is monitored in real time, abnormal conditions are detected, and when water source insufficiency and machine faults occur, an alarm is given out immediately, and corresponding treatment measures are taken. The cleaning efficiency of the engineering vehicle is remarkably improved, the cleaning time is shortened, and the labor cost is reduced; through the synergistic effect of stain analysis and intelligent cleaning, targeted cleaning and comprehensive deep cleaning are achieved, and the cleaning quality is improved.

Description

Technical Field

[0001] The invention relates to the technical field of engineering vehicle cleaning, and in particular discloses an intelligent cleaning method and system for an engineering vehicle. Background Art

[0002] With the rapid development of engineering construction, construction vehicles (such as muck trucks, large trucks, excavators, loaders, mixer trucks, etc.) play a vital role in various construction sites. However, these vehicles are easily contaminated with stubborn stains such as mud and oil during use. The traditional manual cleaning method is not only inefficient, but also labor-intensive, and it is difficult to meet the high efficiency and environmental protection needs of modern construction.

[0003] Therefore, the present invention aims to provide an intelligent cleaning device for engineering vehicles, which can achieve efficient, thorough and environmentally friendly cleaning of engineering vehicles through automated and intelligent technical means. Summary of the invention

[0004] The present invention provides an intelligent cleaning method and system for an engineering vehicle, aiming to solve at least one defect existing in the above-mentioned prior art.

[0005] One aspect of the present invention relates to an intelligent cleaning method for an engineering vehicle, comprising the following steps:

[0006] Capturing the vehicle surface image of the engineering vehicle to be cleaned, and preprocessing the vehicle surface image to obtain vehicle image data;

[0007] Obtain stain recognition results and vehicle model information of engineering vehicles based on vehicle image data and trained machine learning models;

[0008] Based on the stain recognition results and vehicle model information, the optimal cleaning solution is comprehensively calculated and the cleaning parameters are intelligently adjusted;

[0009] Monitor the cleaning status in real time and detect abnormal situations. When water shortage or machine failure occurs, an alarm will be issued immediately and corresponding treatment measures will be taken.

[0010] Furthermore, the vehicle surface image of the engineering vehicle to be cleaned is captured, and the vehicle surface image is preprocessed to obtain the vehicle image data, including the following steps:

[0011] Image acquisition: Use a multispectral camera to capture the vehicle surface image of the engineering vehicle to be cleaned;

[0012] Noise Removal: Apply filters to remove image noise on vehicle surface images;

[0013] Image enhancement: adjust the contrast and brightness of the vehicle surface image to enhance the characteristics of stains;

[0014] In the noise removal step, a Gaussian kernel is used to remove the noise of the vehicle surface image:

[0015]

[0016] Where G(x, y) is the Gaussian kernel, (x, y) is the pixel coordinate, and σ is the standard deviation.

[0017] Furthermore, the steps of obtaining the stain recognition result and vehicle model information of the engineering vehicle according to the vehicle image data and the trained machine learning model include:

[0018] Model training: Use convolutional neural networks to train machine learning models;

[0019] Model evaluation: Use validation sets to evaluate machine learning model performance;

[0020] Deployment: Deploy the trained machine learning model to the actual engineering vehicle cleaning application;

[0021] Stain classification: The extracted stain recognition results and vehicle model information of the engineering vehicle are input into the classifier of the machine learning model to identify the stain type;

[0022] Region segmentation: Segment the vehicle surface image into different stain regions based on stain type or color difference;

[0023] Pollution degree assessment: Calculate the coverage area, density or intensity index of the stains in each stain area to obtain the stain identification results of the engineering vehicle;

[0024] In the model training step, the cross entropy loss function is used to train the machine learning model. The cross entropy loss function is:

[0025]

[0026] Among them, L is the cross entropy loss, N is the number of samples, M is the number of categories, and y ic is the true label, p ic is the predicted probability.

[0027] Furthermore, the stain recognition results and vehicle model information include vehicle shape and stain distribution. Based on the stain recognition results and vehicle model information, the optimal cleaning solution is comprehensively calculated. In the step of intelligently adjusting the cleaning parameters, the spray angle and intensity are automatically adjusted according to the vehicle shape and stain distribution to perform fine cleaning on the engineering vehicles.

[0028] Furthermore, the stain recognition result and vehicle model information include the stain type, distribution area and pollution degree, the stain recognition result and vehicle model information include the vehicle shape and stain distribution, and based on the stain recognition result and vehicle model information, the optimal cleaning solution is comprehensively calculated, and the steps of intelligently adjusting the cleaning parameters include:

[0029] Parameter mapping: Establishing the mapping rules of cleaning parameters according to the stain type, distribution area and degree of contamination;

[0030] Parameter adjustment: Based on the mapping rules of washing parameters, dynamically adjust the type, concentration, spraying amount and cleaning time of the cleaning agent;

[0031] Output recommendation: Output recommended car washing materials and their usage suggestions based on stain type, distribution area and degree of contamination;

[0032] Real-time monitoring: Real-time monitoring of cleaning effect and water quality. When it is found that the cleaning effect is poor or the water quality deteriorates, the cleaning parameters are automatically adjusted;

[0033] In the parameter adjustment step, the cleaning agent concentration C is adjusted according to the contamination level P:

[0034] C=C base +α·P

[0035] Where C is the cleaning agent concentration, C base is the basic concentration, α is the adjustment coefficient, and P is the pollution degree index;

[0036] The spraying volume Q and cleaning time T are adjusted accordingly, taking into account the area of ​​the stain A:

[0037]

[0038] T=T base +β·A

[0039] Where Q is the spraying amount, T is the cleaning time, A is the area of ​​the stain, and Q base and T base are the basic spraying amount and time, A ref is the reference area, β is the time adjustment factor;

[0040] In the step of outputting recommendations, the comprehensive score of the output car wash materials is:

[0041]

[0042] Among them, Score i is the comprehensive score of the i-th car wash material, n is the number of environmental factors considered; w j is the weight of the jth environmental factor; f ij (x j) is the i-th material under the j-th environmental factor χ j The fitness function under j is the current value of the jth environmental factor.

[0043] Another aspect of the present invention relates to an intelligent cleaning system for an engineering vehicle, comprising:

[0044] The first acquisition module is used to capture the vehicle surface image of the engineering vehicle to be cleaned, and pre-process the vehicle surface image to obtain vehicle image data;

[0045] The second acquisition module is used to obtain the stain recognition result and vehicle model information of the engineering vehicle based on the vehicle image data and the trained machine learning model;

[0046] Intelligent adjustment module, which is used to comprehensively calculate the optimal cleaning solution based on the stain recognition results and vehicle model information, and intelligently adjust the cleaning parameters;

[0047] The monitoring and processing module is used to monitor the cleaning status in real time and detect abnormal situations. When water shortage or machine failure occurs, an alarm will be issued immediately and corresponding treatment measures will be taken.

[0048] Furthermore, the first acquisition module includes:

[0049] An image acquisition unit, used for capturing a vehicle surface image of the engineering vehicle to be cleaned by using a multi-spectral camera;

[0050] a noise removal unit for applying a filter to remove image noise to the vehicle surface image;

[0051] An image enhancement unit, used to adjust the contrast and brightness of the vehicle surface image and enhance the stain characteristics;

[0052] In the noise removal unit, a Gaussian kernel is used to remove the noise of the vehicle surface image:

[0053]

[0054] Where G(x, y) is the Gaussian kernel, (x, y) is the pixel coordinate, and σ is the standard deviation.

[0055] Furthermore, the second acquisition module includes:

[0056] A model training unit, used to train a machine learning model using a convolutional neural network;

[0057] Model evaluation unit, used to evaluate the performance of machine learning models using a validation set;

[0058] A deployment unit, used to deploy the trained machine learning model to the actual engineering vehicle cleaning application;

[0059] A stain classification unit, used to input the extracted stain recognition result of the engineering vehicle and the features of the vehicle model information into the classifier of the machine learning model to identify the stain type;

[0060] A region segmentation unit, used for segmenting the vehicle surface image into different stain regions based on stain types or color differences;

[0061] A pollution degree assessment unit, used to calculate the coverage area, density or intensity index of the stains in each stain area, and obtain the stain identification result of the engineering vehicle;

[0062] In the model training unit, the cross entropy loss function is used to train the machine learning model. The cross entropy loss function is:

[0063]

[0064] Among them, L is the cross entropy loss, N is the number of samples, M is the number of categories, and y ic is the true label, p ic is the predicted probability.

[0065] Furthermore, in the intelligent adjustment module, the spray angle and intensity are automatically adjusted according to the shape of the vehicle and the distribution of stains, so as to carry out fine cleaning of the engineering vehicles.

[0066] Furthermore, the intelligent adjustment module includes:

[0067] A parameter mapping unit, for establishing a mapping rule for cleaning parameters according to the type of stain, distribution area and degree of contamination;

[0068] A parameter adjustment unit is used to dynamically adjust the type, concentration, spraying amount and cleaning time of the cleaning agent based on the mapping rules of the cleaning parameters;

[0069] An output recommendation unit, used for outputting recommended car washing materials and usage suggestions according to the stain type, distribution area and contamination degree;

[0070] Real-time monitoring unit, used to monitor the cleaning effect and water quality in real time. When it is found that the cleaning effect is not good or the water quality deteriorates, the cleaning parameters are automatically adjusted;

[0071] In the parameter adjustment unit, the cleaning agent concentration C is adjusted according to the contamination level P:

[0072] C=C base +α·P

[0073] Where C is the cleaning agent concentration, C base is the basic concentration, α is the adjustment coefficient, and P is the pollution degree index;

[0074] The spraying volume Q and cleaning time T are adjusted accordingly, taking into account the area of ​​the stain A:

[0075]

[0076] T=T base +β·A

[0077] Where Q is the spraying amount, T is the cleaning time, A is the area of ​​the stain, and Q base and T base are the basic spraying amount and time, A ref is the reference area, β is the time adjustment factor;

[0078] In the output recommendation unit, the comprehensive score of the output car wash materials is:

[0079]

[0080] Among them, Score i is the comprehensive score of the i-th car wash material, n is the number of environmental factors considered; w j is the weight of the jth environmental factor; f ij (x j ) is the i-th material under the j-th environmental factor χ j The fitness function under j is the current value of the jth environmental factor.

[0081] The beneficial effects achieved by the present invention are:

[0082] The present invention provides an intelligent cleaning method and system for engineering vehicles. The method captures the vehicle surface image of the engineering vehicle to be cleaned and pre-processes the vehicle surface image to obtain vehicle image data. The stain recognition result and vehicle model information of the engineering vehicle are obtained based on the vehicle image data and the trained machine learning model. Based on the stain recognition result and vehicle model information, the optimal cleaning solution is comprehensively calculated and the cleaning parameters are intelligently adjusted. The cleaning status is monitored in real time and abnormal conditions are detected. When water shortage or machine failure occurs, an alarm is immediately issued and corresponding treatment measures are taken. The beneficial effects achieved by the intelligent cleaning method and system for engineering vehicles provided by the present invention are specifically as follows:

[0083] 1. Improve cleaning efficiency: Automated and intelligent cleaning methods significantly improve the cleaning efficiency of engineering vehicles, shorten cleaning time, and reduce labor costs.

[0084] 2. Improve cleaning quality: Through the synergy of stain analysis and intelligent cleaning, targeted cleaning and comprehensive deep cleaning are achieved, improving the cleaning quality.

[0085] 3. Environmental protection and energy saving: The application of environmental protection treatment system reduces the impact of the cleaning process on the environment. At the same time, the use of low VOC (volatile organic compounds) environmentally friendly cleaning agents is more in line with environmental protection requirements.

[0086] 4. Enhance user experience: The intelligent design of the user interaction interface makes the car washing process more convenient and comfortable, improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 The present invention is a flow chart of an embodiment of an intelligent cleaning method for an engineering vehicle. DETAILED DESCRIPTION

[0088] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0089] like Figure 1 As shown, the first embodiment of the present invention provides an intelligent cleaning method for an engineering vehicle, comprising the following steps:

[0090] Step S100: Capture the vehicle surface image of the engineering vehicle to be cleaned, and pre-process the vehicle surface image to obtain vehicle image data.

[0091] Using high-definition cameras and image processing technology, the type, size and license plate information of engineering vehicles entering the cleaning area can be automatically identified. Through laser radar or visual positioning technology, the position of the vehicle in the cleaning area can be accurately measured to ensure that the cleaning device can accurately align the vehicle for cleaning.

[0092] Vehicle detection sensors: Through the installation of devices such as infrared sensors or lidar, when an engineering vehicle enters its detection range, the sensor will capture the presence of the vehicle and send a signal to the control system. These sensors are usually highly sensitive and fast-response, ensuring that the cleaning process is triggered immediately when a vehicle enters the cleaning area.

[0093] Environmental sensors: Environmental sensors such as temperature sensors and humidity sensors are integrated inside or around the cleaning device. They measure environmental parameters in real time through physical sensing principles and transmit data to the control system. The control system adjusts cleaning parameters such as water temperature and detergent concentration based on these data to adapt to different environmental conditions.

[0094] Stain detection sensors: Stain detection sensors such as multispectral cameras capture images of the vehicle surface and use spectral analysis technology to identify the type and distribution of stains. These cameras are able to capture light of different wavelengths and generate image data containing rich information. By analyzing this data through image processing algorithms, the type of stain and the degree of contamination can be accurately identified.

[0095] Image preprocessing: The collected image data first goes through the preprocessing stage, including denoising, contrast enhancement, etc. These operations are designed to improve image quality, reduce noise and interference factors, and provide clear image data for subsequent analysis.

[0096] Step S200: Obtain stain recognition results and vehicle model information of the engineering vehicle based on the vehicle image data and the trained machine learning model.

[0097] The pre-processed image data is input into the image processing algorithm in the machine learning model for analysis. The image processing algorithm uses image segmentation, feature extraction and other technologies to identify the type, distribution area and degree of contamination of the stains. These analysis results will serve as the decision-making basis for the intelligent cleaning system.

[0098] Use the labeled stain image data to train classifiers, such as SVM (Support Vector Machine) and CNN (Convolutional Neural Network).

[0099] In this embodiment, “Convolutional Neural Networks (CNN)” is used to identify the engineering vehicle.

[0100] Convolutional Neural Networks: CNN is particularly suitable for image recognition tasks because it can automatically learn and extract hierarchical feature representations from raw images.

[0101] Step S300: Based on the stain recognition result and vehicle model information, the optimal cleaning solution is comprehensively calculated and the cleaning parameters are intelligently adjusted.

[0102] Based on the stain recognition results and vehicle model information, the machine learning algorithm in the machine learning model will intelligently adjust the cleaning parameters. The machine learning algorithm will consider factors such as the type of stain, distribution area, degree of contamination, and vehicle model characteristics, and comprehensively calculate the optimal cleaning solution. These solutions include the setting of parameters such as the type of detergent, concentration, spraying amount, and cleaning time.

[0103] In this embodiment, according to the stain analysis results, cleaning parameters such as the type, concentration, spraying amount and cleaning time of the cleaning agent are intelligently adjusted to achieve targeted cleaning.

[0104] The stain recognition results and vehicle model information include vehicle shape and stain distribution. Based on the stain recognition results and vehicle model information, the optimal cleaning plan is comprehensively calculated. In the step of intelligently adjusting the cleaning parameters, the spray angle and intensity are automatically adjusted according to the vehicle shape and stain distribution to perform fine cleaning on the engineering vehicles.

[0105] Step S400: monitor the cleaning status in real time and detect abnormal conditions. When water shortage or machine failure occurs, an alarm is immediately sounded and corresponding treatment measures are taken.

[0106] The system status is monitored in real time and abnormal situations are detected. When the system encounters problems such as insufficient water supply and machine failure, an alarm will be immediately issued and appropriate treatment measures will be taken to ensure the normal progress of the cleaning operation.

[0107] Further, see Figure 1 In the intelligent cleaning method for engineering vehicles provided in this embodiment, step S100 includes:

[0108] Step S110, image acquisition: using a multispectral camera to capture the vehicle surface image of the engineering vehicle to be cleaned.

[0109] With the help of multispectral imaging technology, the stain image on the vehicle surface is captured, and the stain type (such as dirt, oil, cement stains, etc.), distribution area and degree of contamination are analyzed.

[0110] Step S120, noise removal: applying a filter to remove image noise from the vehicle surface image.

[0111] Apply filters (such as Gaussian filtering, median filtering) to remove image noise.

[0112] Use a Gaussian kernel to remove noise from the image of the vehicle surface:

[0113]

[0114] In formula (1), G(x, y) is the Gaussian kernel, (x, y) is the pixel coordinate, and σ is the standard deviation.

[0115] Step S130, image enhancement: adjusting the contrast and brightness of the vehicle surface image to enhance the stain features.

[0116] Adjust the contrast, brightness, etc. of the vehicle surface image to enhance the stain characteristics.

[0117] Preferably, see Figure 1 In the intelligent cleaning method for engineering vehicles provided in this embodiment, step S200 includes:

[0118] Step S210, model training: use a convolutional neural network to train the machine learning model.

[0119] The CNN model is trained with the goal of minimizing the classification error (such as cross entropy loss).

[0120] The machine learning model is trained using the cross entropy loss function, which is:

[0121]

[0122] In formula (2), L is the cross entropy loss, N is the number of samples, M is the number of categories, and y ic is the true label, p ic is the predicted probability.

[0123] Step S220, model evaluation: Use the validation set to evaluate the performance of the machine learning model.

[0124] Use the validation set to evaluate model performance, such as accuracy and recall.

[0125] Step S230, deployment: deploying the trained machine learning model to the actual engineering vehicle cleaning application.

[0126] Apply the trained model to practical applications, such as actual engineering vehicle cleaning applications.

[0127] Step S240, stain classification: input the extracted stain recognition result of the engineering vehicle and the features of the vehicle model information into the classifier of the machine learning model to identify the stain type.

[0128] The extracted stain recognition results of the engineering vehicle and the features of the vehicle model information are input into the classifier to identify the stain type (mud, oil, cement, etc.).

[0129] Step S250: region segmentation: segmenting the vehicle surface image into different stain regions based on stain types or color differences.

[0130] Step S260, pollution degree assessment: calculating the coverage area, density or intensity index of the stains in each stain area, and obtaining the stain recognition result of the engineering vehicle.

[0131] The stain coverage area is calculated as follows:

[0132] Coverage Area=∑ i∈stained pixels Area of ​​pixel i (3)

[0133] In formula (3), Coverage Area is the area covered by the stain, Area of ​​pixel iis the area of ​​each stain pixel, and i traverses all pixels identified as stains.

[0134] Further, see Figure 1 In the intelligent cleaning method for engineering vehicles provided in this embodiment, step S300 includes:

[0135] Step S310, parameter mapping: establishing a mapping rule for cleaning parameters according to the stain type, distribution area and contamination degree.

[0136] A mapping table of stain type, distribution area, contamination degree and cleaning parameters is established in the database.

[0137] Step S320, parameter adjustment: based on the mapping rules of the washing parameters, dynamically adjust the type, concentration, spraying amount and washing time of the cleaning agent.

[0138] The cleaning agent concentration C is adjusted according to the degree of contamination P: C = C base +α·P(4)

[0139] In formula (4), C is the cleaning agent concentration, C base is the basic concentration, α is the adjustment coefficient, and P is the pollution degree index;

[0140] The spraying volume Q and cleaning time T are adjusted accordingly, taking into account the area of ​​the stain A:

[0141]

[0142] T=T base +β·A (6)

[0143] In formulas (5) and (6), Q is the spraying amount, T is the cleaning time, A is the area of ​​the stain, and Q is base and T base are the basic spraying amount and time, A ref is the reference area and β is the time adjustment factor.

[0144] Step S330, output recommendation: output recommended car wash materials and usage suggestions based on the stain type, distribution area and pollution degree.

[0145] The comprehensive rating of the output car wash materials is:

[0146]

[0147] In formula (7), Score i is the comprehensive score of the i-th car wash material, n is the number of environmental factors considered; w j is the weight of the jth environmental factor, indicating the importance of this factor in selecting car wash materials, and fij (x j ) is the i-th material under the j-th environmental factor χ j The fitness function under j is the current value of the jth environmental factor.

[0148] Parameter definition:

[0149] w j : Weight, determined by experts based on experience or data analysis, indicating the degree of influence of different environmental factors on the selection of car wash materials.

[0150] f ij (x j ): Fitness function is a function designed according to specific environmental factors and material properties. For example, for temperature factors, if a material is prone to freezing at low temperatures, the function will output a lower value at low temperatures; conversely, it will output a higher value within a suitable temperature range.

[0151] χ j :The current values ​​of environmental factors, such as current temperature, humidity, etc., are provided in real time by environmental sensors.

[0152] Step S340, real-time monitoring: real-time monitoring of the cleaning effect and water quality. When it is found that the cleaning effect is not good or the water quality deteriorates, the cleaning parameters are automatically adjusted.

[0153] Feedback adjustment and optimization: Smart cleaning equipment monitors the cleaning effect and water quality in real time through sensors and control systems. When it is found that the cleaning effect is poor or the water quality deteriorates, the control system will automatically adjust the cleaning parameters (such as the concentration of the cleaning agent, the spraying amount, the cleaning time, etc.) to ensure that the cleaning effect is optimal. This feedback adjustment mechanism makes the smart cleaning equipment more adaptable and stable.

[0154] The present invention relates to an intelligent cleaning system for engineering vehicles, comprising a first acquisition module, a second acquisition module, an intelligent adjustment module and a monitoring and processing module, wherein the first acquisition module is used to capture the vehicle surface image of the engineering vehicle to be cleaned, and pre-process the vehicle surface image to obtain vehicle image data. The second acquisition module is used to obtain the stain recognition result and vehicle model information of the engineering vehicle based on the vehicle image data and a trained machine learning model. The intelligent adjustment module is used to comprehensively calculate the optimal cleaning solution based on the stain recognition result and vehicle model information, and intelligently adjust the cleaning parameters. The monitoring and processing module is used to monitor the cleaning status in real time and detect abnormal conditions. When there is insufficient water source or machine failure, an alarm is immediately issued and corresponding treatment measures are taken.

[0155] The first acquisition module uses high-definition cameras and image processing technology to automatically identify the type, size and license plate information of engineering vehicles entering the cleaning area. Through laser radar or visual positioning technology, the position of the vehicle in the cleaning area is accurately measured to ensure that the cleaning device can accurately align the vehicle for cleaning.

[0156] The second acquisition module uses image segmentation, feature extraction and other technologies to identify the type, distribution area and degree of contamination of stains. These analysis results will serve as the decision-making basis for the intelligent cleaning system.

[0157] Based on the stain recognition results and vehicle model information, the intelligent adjustment module will intelligently adjust the cleaning parameters. The intelligent adjustment module will consider factors such as the type of stains, distribution area, degree of contamination, and vehicle model characteristics, and comprehensively calculate the optimal cleaning solution. These solutions include the settings of parameters such as the type of detergent, concentration, spray volume, and cleaning time.

[0158] The monitoring and processing module monitors the system status in real time and detects abnormal situations. When the system encounters problems such as insufficient water supply and machine failure, it will immediately issue an alarm and take corresponding processing measures to ensure the normal progress of the cleaning operation.

[0159] Furthermore, in the intelligent cleaning system for engineering vehicles provided in this embodiment, the first acquisition module includes an image acquisition unit, a noise removal unit and an image enhancement unit, wherein the image acquisition unit is used to capture the vehicle surface image of the engineering vehicle to be cleaned using a multispectral camera. The noise removal unit is used to apply a filter to remove image noise from the vehicle surface image. The image enhancement unit is used to adjust the contrast and brightness of the vehicle surface image to enhance the stain characteristics.

[0160] Preferably, in the intelligent cleaning system for engineering vehicles provided in this embodiment, the second acquisition module includes a model training unit, a model evaluation unit, a deployment unit, a stain classification unit, a region segmentation unit and a pollution degree assessment unit, wherein the model training unit is used to train the machine learning model using a convolutional neural network. The model evaluation unit is used to evaluate the performance of the machine learning model using a validation set. The deployment unit is used to deploy the trained machine learning model to an actual engineering vehicle cleaning application. The stain classification unit is used to input the extracted features of the stain recognition results and vehicle model information of the engineering vehicle into the classifier of the machine learning model to identify the stain type. The region segmentation unit is used to segment the vehicle surface image into different stain regions based on the stain type or color difference. The pollution degree assessment unit is used to calculate the coverage area, density or intensity index of the stain in each stain region to obtain the stain recognition result of the engineering vehicle.

[0161] Furthermore, the intelligent cleaning system for engineering vehicles provided in this embodiment, the intelligent adjustment module includes a parameter mapping unit, a parameter adjustment unit, an output recommendation unit and a real-time monitoring unit, wherein the parameter mapping unit is used to establish a mapping rule for cleaning parameters according to the type of stain, distribution area and degree of pollution. The parameter adjustment unit is used to dynamically adjust the type, concentration, spraying amount and cleaning time of the cleaning agent based on the mapping rule of the washing parameters. The output recommendation unit is used to output recommended car washing materials and usage suggestions according to the type of stain, distribution area and degree of pollution; the real-time monitoring unit is used to monitor the cleaning effect and water quality in real time, and automatically adjust the cleaning parameters when it is found that the cleaning effect is not good or the water quality has deteriorated.

[0162] Compared with the prior art, the intelligent cleaning method and system for engineering vehicles provided in this embodiment captures the vehicle surface image of the engineering vehicle to be cleaned and pre-processes the vehicle surface image to obtain vehicle image data; obtains the stain recognition result and vehicle model information of the engineering vehicle based on the vehicle image data and the trained machine learning model; comprehensively calculates the optimal cleaning solution based on the stain recognition result and vehicle model information, and intelligently adjusts the cleaning parameters; monitors the cleaning status in real time and detects abnormal conditions. When there is insufficient water or machine failure, an alarm is immediately issued and corresponding treatment measures are taken. The beneficial effects achieved by the intelligent cleaning method and system for engineering vehicles provided in this embodiment are as follows:

[0163] 1. Improve cleaning efficiency: Automated and intelligent cleaning methods significantly improve the cleaning efficiency of engineering vehicles, shorten cleaning time, and reduce labor costs.

[0164] 2. Improve cleaning quality: Through the synergy of stain analysis and intelligent cleaning, targeted cleaning and comprehensive deep cleaning are achieved, improving the cleaning quality.

[0165] 3. Environmental protection and energy saving: The application of environmental protection treatment system reduces the impact of the cleaning process on the environment. At the same time, the use of low VOC (volatile organic compounds) environmentally friendly cleaning agents is more in line with environmental protection requirements.

[0166] 4. Enhance user experience: The intelligent design of the user interaction interface makes the car washing process more convenient and comfortable, improving the user experience.

[0167] Although preferred embodiments of the present invention have been described, additional changes and modifications may be made to these embodiments by those skilled in the art once the basic inventive concepts are known. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention. Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An intelligent cleaning method for an engineering vehicle, characterized in that: The following steps are involved: Capturing a vehicle surface image of a current engineering vehicle to be cleaned, and preprocessing the vehicle surface image to obtain vehicle image data; Obtaining stain recognition results and vehicle model information of the engineering vehicle according to the vehicle image data and the trained machine learning model; Based on the stain recognition results and vehicle model information, the optimal cleaning solution is comprehensively calculated and the cleaning parameters are intelligently adjusted; Monitor the cleaning status in real time and detect abnormal situations. When water shortage or machine failure occurs, an alarm will be issued immediately and corresponding treatment measures will be taken.

2. The intelligent cleaning method for an engineering vehicle according to claim 1, characterized in that: The step of capturing the vehicle surface image of the engineering vehicle to be cleaned and preprocessing the vehicle surface image to obtain vehicle image data includes: Image acquisition: Use a multispectral camera to capture the vehicle surface image of the engineering vehicle to be cleaned; Noise removal: applying a filter to remove image noise on the vehicle surface image; Image enhancement: adjusting the contrast and brightness of the vehicle surface image to enhance the stain features; In the noise removal step, a Gaussian kernel is used to remove noise from the vehicle surface image: Where G(x, y) is the Gaussian kernel, (x, y) is the pixel coordinate, and σ is the standard deviation.

3. The intelligent cleaning method for an engineering vehicle according to claim 1, characterized in that: The step of obtaining the stain recognition result and vehicle model information of the engineering vehicle according to the vehicle image data and the trained machine learning model comprises: Model training: using a convolutional neural network to train the machine learning model; Model evaluation: Use the validation set to evaluate the performance of the machine learning model; Deployment: Deploy the trained machine learning model to the actual engineering vehicle cleaning application; Stain classification: The extracted stain recognition results and features of the vehicle model information of the engineering vehicle are input into the classifier of the machine learning model to identify the type of stain; Region segmentation: segmenting the vehicle surface image into different stain regions based on stain type or color difference; Pollution degree assessment: Calculate the coverage area, density or intensity index of the stains in each stain area to obtain the stain identification results of the engineering vehicle; In the model training step, the machine learning model is trained using a cross entropy loss function, and the cross entropy loss function is: Among them, L is the cross entropy loss, N is the number of samples, M is the number of categories, and y ic is the true label, p ic is the predicted probability.

4. The intelligent cleaning method for an engineering vehicle as claimed in claim 3, characterized in that: The stain recognition result and vehicle model information include vehicle shape and stain distribution. In the step of comprehensively calculating the optimal cleaning solution based on the stain recognition result and vehicle model information and intelligently adjusting the cleaning parameters, the spray angle and intensity are automatically adjusted according to the vehicle shape and stain distribution to perform fine cleaning on the engineering vehicle.

5. The intelligent cleaning method for an engineering vehicle according to claim 3, characterized in that: The stain recognition result and vehicle model information include the stain type, distribution area and pollution degree, the stain recognition result and vehicle model information include the vehicle shape and stain distribution, and the step of comprehensively calculating the optimal cleaning solution based on the stain recognition result and vehicle model information and intelligently adjusting the cleaning parameters includes: Parameter mapping: Establishing the mapping rules of cleaning parameters according to the stain type, distribution area and degree of contamination; Parameter adjustment: based on the mapping rules of the washing parameters, dynamically adjust the type, concentration, spraying amount and cleaning time of the cleaning agent; Output recommendation: Output recommended car washing materials and their usage suggestions based on stain type, distribution area and degree of contamination; Real-time monitoring: Real-time monitoring of cleaning effect and water quality. When it is found that the cleaning effect is poor or the water quality deteriorates, the cleaning parameters are automatically adjusted; In the parameter adjustment step, the cleaning agent concentration C is adjusted according to the contamination level P: C=C base +α·P Where C is the cleaning agent concentration, C base is the basic concentration, α is the adjustment coefficient, and P is the pollution degree index; The spraying volume Q and cleaning time T are adjusted accordingly, taking into account the area of ​​the stain A: T=T base +β·A Where Q is the spraying amount, T is the cleaning time, A is the area of ​​the stain, and Q base and T base are the basic spraying amount and time, A ref is the reference area, β is the time adjustment factor; In the step of outputting the recommendation, the comprehensive score of the outputted car wash materials is: Among them, Score i is the comprehensive score of the i-th car wash material, n is the number of environmental factors considered; w j is the weight of the jth environmental factor; f ij (x j ) is the i-th material under the j-th environmental factor χ j The fitness function under j is the current value of the jth environmental factor.

6. An intelligent cleaning system for engineering vehicles, characterized in that: include: A first acquisition module is used to capture a vehicle surface image of the engineering vehicle to be cleaned, and pre-process the vehicle surface image to acquire vehicle image data; A second acquisition module is used to obtain the stain recognition result and vehicle model information of the engineering vehicle according to the vehicle image data and the trained machine learning model; An intelligent adjustment module, for comprehensively calculating the optimal cleaning solution based on the stain recognition result and vehicle model information, and intelligently adjusting the cleaning parameters; The monitoring and processing module is used to monitor the cleaning status in real time and detect abnormal situations. When water shortage or machine failure occurs, an alarm will be issued immediately and corresponding treatment measures will be taken.

7. The intelligent cleaning system for engineering vehicles according to claim 6, characterized in that: The first acquisition module includes: An image acquisition unit, used for capturing a vehicle surface image of the engineering vehicle to be cleaned by using a multi-spectral camera; a noise removal unit for applying a filter to remove image noise from the vehicle surface image; An image enhancement unit, used to adjust the contrast and brightness of the vehicle surface image and enhance the stain characteristics; In the noise removal unit, a Gaussian kernel is used to remove noise from the image of the vehicle surface: Where G(x, y) is the Gaussian kernel, (x, y) is the pixel coordinate, and σ is the standard deviation.

8. The intelligent cleaning system for engineering vehicles according to claim 6, characterized in that: The second acquisition module includes: A model training unit, used to train the machine learning model using a convolutional neural network; A model evaluation unit, used to evaluate the performance of the machine learning model using a validation set; A deployment unit, used to deploy the trained machine learning model to the actual engineering vehicle cleaning application; A stain classification unit, used to input the extracted stain recognition result and vehicle model information of the engineering vehicle into the classifier of the machine learning model to identify the type of stain; A region segmentation unit, used for segmenting the vehicle surface image into different stain regions based on stain types or color differences; A pollution degree assessment unit, used to calculate the coverage area, density or intensity index of the stains in each stain area, and obtain the stain identification result of the engineering vehicle; In the model training unit, the machine learning model is trained using a cross entropy loss function, and the cross entropy loss function is: Among them, L is the cross entropy loss, N is the number of samples, M is the number of categories, and y ic is the true label, p ic is the predicted probability.

9. The intelligent cleaning system for engineering vehicles according to claim 8, characterized in that: In the intelligent adjustment module, the spray angle and intensity are automatically adjusted according to the vehicle shape and stain distribution, so as to carry out fine cleaning of the engineering vehicle.

10. The intelligent cleaning system for engineering vehicles according to claim 8, characterized in that: The intelligent adjustment module comprises: A parameter mapping unit, for establishing a mapping rule for cleaning parameters according to the type of stain, distribution area and degree of contamination; A parameter adjustment unit, for dynamically adjusting the type, concentration, spraying amount and cleaning time of the cleaning agent based on the mapping rules of the cleaning parameters; An output recommendation unit, used for outputting recommended car washing materials and usage suggestions according to the stain type, distribution area and contamination degree; Real-time monitoring unit, used to monitor the cleaning effect and water quality in real time. When it is found that the cleaning effect is not good or the water quality deteriorates, the cleaning parameters are automatically adjusted; In the parameter adjustment unit, the cleaning agent concentration C is adjusted according to the contamination level P: C=C base +α·P Where C is the cleaning agent concentration, C base is the basic concentration, α is the adjustment coefficient, and P is the pollution degree index; The spraying volume Q and cleaning time T are adjusted accordingly, taking into account the area of ​​the stain A: T=T base +β·A Where Q is the spraying amount, T is the cleaning time, A is the area of ​​the stain, and Q base and T base are the basic spraying amount and time, A ref is the reference area, β is the time adjustment factor; In the output recommendation unit, the comprehensive score of the car wash material output is: Among them, Score i is the comprehensive score of the i-th car wash material, n is the number of environmental factors considered; w j is the weight of the jth environmental factor; f ij (x j ) is the i-th material under the j-th environmental factor χ j The fitness function under j is the current value of the jth environmental factor.

Images