Air conditioner purification unit filter assembly dust accumulation recognition and replacement early warning system and method

Abstract

The application discloses an air conditioner purification unit filter assembly dust accumulation amount identification and replacement early warning system and method, and relates to the technical field of intelligent identification and early warning.The application comprises an image acquisition module, the image acquisition module at least comprising a 4K automatic focusing high-definition network camera, a high-resolution black-and-white camera and a CCD pinhole type high-definition network camera, and further comprising an image data transmission module, a data storage processing platform, a dust accumulation amount intelligent identification system, a dust accumulation data transmission module and an early warning control module.The early warning control module at least comprises three kinds of prompts in the early warning process, the three kinds of prompts being short message notification, display screen display alarm and indicator light alarm.The application is based on machine learning, and establishes a kind of efficient and fast, low-cost air conditioner filter assembly dust accumulation amount intelligent identification method, which greatly improves the efficiency of air conditioner component dust accumulation amount identification, and the filter assembly replacement early warning system can set corresponding threshold values by itself, thereby performing early warning of different degrees of dust accumulation amount.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent identification and early warning, in particular to an air conditioner purification unit filter assembly dust accumulation identification and replacement early warning system and method. BACKGROUND

[0002] At present, the intelligent identification and replacement early warning system and method for the dust accumulation of the filter assembly of the air conditioner purification unit,

[0003] The main one is the real-time detection system of the air conditioner filter screen dust, which includes the air conditioner indoor unit, the outdoor unit, the filter screen dust detection control chip. By setting the air, the wind speed, the temperature, the humidity and the use time monitoring device, the multi-parameter detection can improve the accuracy of the dust full detection result, and can provide a comfortable and safe air conditioner use experience for the user. The heavy metal lead detector and the dust mite activity detector are both arranged on the air conditioner filter screen, and the air monitoring device includes a mounting plate, a vertical plate, a power component and an air monitor mounted above the air conditioner outdoor unit. The air detector monitors the air quality parameters in the current installation environment of the air conditioner outdoor unit, and returns the obtained air quality parameters to the filter screen dust detection control chip. The air conditioner indoor unit is provided with a wind speed monitoring device on one side, and a wind speed sensor is used for monitoring the average value of the wind speed of the air conditioner indoor unit in the working state in the cycle time, and the wind speed parameter is returned to the filter screen dust detection control chip. The air conditioner system runs in different environments, and the accumulation speed of the dust is different. Through various parameters, the various pollution degrees (light, medium and heavy) of the filter screen dust can be accurately detected. The air conditioner display screen voice broadcaster broadcasts voice according to different pollution degrees, reminding the user to clean the filter screen in time.

[0004] The existing real-time detection system of the air conditioner filter screen dust has the defects of limited detection accuracy, high cost, great maintenance difficulty, data processing lag, low intelligent level and poor environmental adaptability.

[0005] Therefore, a new solution is needed to solve the above problems. SUMMARY

[0006] The purpose of the present application is to provide an air conditioner purification unit filter assembly dust accumulation identification and replacement early warning system and method to solve the technical problems in the background art and achieve accurate, efficient and economic air conditioner filter assembly dust accumulation identification and early warning effect.

[0007] In order to achieve the above object, the present application provides the following technical scheme: the air conditioner purification unit filter assembly dust amount identification and replacement early warning system, including an image acquisition module, the image acquisition module at least includes a 4K automatic focusing high-definition network camera, a high-resolution black-and-white camera and a CCD pinhole type high-definition network camera, also including an image data transmission module, a data storage processing platform, a dust amount intelligent identification system, a dust data transmission module and a warning control module, the warning control module at least includes three kinds of prompts in the early warning process, the three kinds of prompts are short message notification alarm, display screen display alarm and indicator light alarm;

[0008] The image acquisition module transmits the collected image to the data storage processing platform through the image data transmission module, and the dust amount intelligent identification system in the data storage processing platform carries out intelligent identification of the dust amount, and transmits the specific identified data to the warning control module, and carries out alarm processing according to whether the alarm threshold is reached.

[0009] The air conditioner purification unit filter assembly dust amount identification and replacement early warning method at least includes the following steps:

[0010] S1: based on experiment or related method, the image information of filter assembly dust amount and the morphological feature information of filter assembly dust amount are obtained, combined and processed, so as to establish a typical feature database of air conditioner filter assembly dust amount;

[0011] S2: by introducing integrated machine vision and learning algorithm, a basic image recognition model is established, the parameters of the basic image recognition model are set according to the device configuration and image size, the basic image recognition model is used to process the image information of air conditioner filter assembly, and the image information at least includes preliminary identification and classification of dust amount and morphology in the image;

[0012] S3: the morphological feature information and image information of the processed filter assembly dust are input into the basic image recognition model for targeted training, forming an automatic identification basic model of filter assembly dust capable of accurately identifying the images in the typical feature database of filter assembly dust, and having certain identification ability to new images, and the automatic identification basic model of filter assembly dust is used to execute image recognition task and generate recognition result;

[0013] S4: build a filter assembly dust automatic identification model, when the training result of the filter assembly dust automatic identification basic model and the machine vision learning algorithm reaches enough confidence, acquire the filter assembly dust amount image to be processed and input it into the filter assembly dust automatic identification model for identification, and obtain the filter assembly dust amount image to be processed recognition result;

[0014] S5: Randomly extract the recognition result for verification analysis while obtaining the result, eliminate the influence of uncertain factors, improve the accuracy of the automatic recognition result of the filter assembly dust, verify the reliability of the model, and complete iteration;

[0015] S6: The filter assembly dust automatic recognition model combines the deep learning algorithm convolutional neural network to model the change characteristics of the filter assembly dust pollution; a random forest ensemble learning method is used to build a filter assembly dust pollution risk prediction model, further improving the accuracy and stability of the filter assembly dust automatic recognition model;

[0016] S7: Based on the filter assembly dust pollution risk prediction model and the early warning control module, an intelligent replacement early warning system is built to monitor and analyze the filter assembly dust pollution risk in real time. When the filter assembly dust pollution condition reaches a certain threshold, the system will issue a warning through indicator lights or SMS notifications to remind users to take measures to clean the air conditioner filter assembly.

[0017] Further, the S4 obtains the image to be processed and inputs it into the filter assembly dust automatic recognition model for identification before the basic identification model training, that is, before the image to be processed is input into the established filter assembly dust amount automatic recognition model, the model is ensured to have the ability to correctly extract, understand and identify the dust amount related features from the image;

[0018] The basic identification model training at least includes the following steps:

[0019] Obtain the image information of the filter assembly dust amount and combine and form the filter assembly dust amount training;

[0020] Classify the morphological features of the filter assembly dust amount in the training image to obtain multiple sets of filter assembly dust amount training images;

[0021] Based on the morphological features of the filter assembly dust amount in the training image, the basic image recognition model is trained specifically to form a filter assembly dust amount automatic recognition system;

[0022] Based on the filter assembly dust amount typical feature database construction method, the filter assembly dust amount range and morphological features are associated with the image in the form of an electronic tag to form a filter assembly dust amount training image;

[0023] The filter assembly dust amount is divided into a training set, a test set and a verification set in a ratio of 5:4:1 for filter assembly dust amount automatic recognition model training.

[0024] Further, the filter assembly dust automatic recognition model at least includes the following steps:

[0025] The Backbone network structure in the machine vision learning algorithm is used to cut the filter component dust amount image, the Neck network structure is used to extract and fuse the filter component dust amount image, and the Predicton network structure is used to predict and output the filter component dust amount image.

[0026] The machine vision learning algorithm needs to introduce a pre-configured basic training model, which at least includes preset model selection, test data type, prediction result saving path, input picture size, confidence threshold and result display setting.

[0027] The test data type is a picture, and the input picture resolution is set to 640x640 according to the computer parameters, and the confidence threshold is set to 0.80 by default.

[0028] The number of training rounds is set to 200, and the pre-configured basic training model is trained to ensure that the training result reaches a sufficient confidence, thereby forming a reliable filter component dust amount automatic identification model.

[0029] Further, the machine vision learning algorithm is used to construct the filter component dust automatic identification model, which at least includes the following steps:

[0030] Labelme is used to make labels for filter component dust amount training images to form json files, and then the json files are converted into txt files.

[0031] The basic training model is trained by train to form a filter component dust amount automatic identification model, and the filter component dust amount identification is realized by detect.

[0032] Further, after the machine vision learning algorithm is used to construct the filter component dust amount automatic identification model, the filter component dust amount image to be processed needs to be obtained, and the filter component dust amount image to be processed is obtained by at least including the following steps:

[0033] The image acquisition module is used to record the filter component dust amount image.

[0034] The size of the filter component dust amount image is adjusted to the resolution of the input picture to form the filter component dust amount image to be processed for the filter component dust amount automatic identification model to identify.

[0035] Further, the identification result of the image to be processed includes a picture result with confidence information, and an EXCEL table result with filter component dust amount range and morphological feature information.

[0036] Further, the identification result is randomly extracted for verification analysis at the same time of obtaining the result, which at least includes the following steps:

[0037] The dust amount of the filter assembly is analyzed by optical microscopy;

[0038] The optical microscopy analysis result is compared with the identification result of the image to be processed

[0039] The reliability of the automatic identification model is verified, and the model is perfected;

[0040] The model is perfected and added to the filter assembly dust amount typical feature database.

[0041] Further, the machine vision learning algorithm inputs the size of the picture to refer to the computer configuration parameters, so as to improve the training speed of the basic model and the identification speed of the filter assembly dust amount automatic identification model.

[0042] Compared with the prior art, the beneficial effects of the present application are:

[0043] The present application is based on machine learning, and establishes a kind of efficient and fast, low-cost air conditioner filter assembly dust amount intelligent identification method, greatly improves the efficiency of air conditioner component dust amount identification, and filter assembly replacement early warning system can set corresponding threshold value by itself, to carry out early warning of different degrees of dust amount. BRIEF DESCRIPTION OF DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used for the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0045] Figure 1 The basic flow chart of the filter assembly dust amount intelligent identification based on machine vision learning of the present application is shown in the figure;

[0046] Figure 2 The filter assembly dust amount intelligent identification flow chart of the present application is shown in the figure;

[0047] Figure 3 The filter assembly dust amount risk prediction flow chart of the present application is shown in the figure;

[0048] Figure 4 The filter assembly dust amount replacement early warning flow chart of the present application is shown in the figure;

[0049] Figure 5 The hardware system schematic diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application.

[0051] Embodiment one:

[0052] Please refer to Figure 5 , the air conditioning purification unit filter assembly dust accumulation identification and replacement early warning system and method, comprising.

[0053] The air conditioning purification unit filter assembly dust accumulation identification and replacement early warning system comprises an image acquisition module, the image acquisition module at least includes a 4K automatic focusing high-definition network camera, a high-resolution black-and-white camera and a CCD pinhole type high-definition network camera, and further comprises an image data transmission module, a data storage processing platform, a dust accumulation intelligent identification system, a dust data transmission module and an early warning control module. The early warning control module at least includes three kinds of prompts in the early warning process, which are short message notification alarm, display screen display alarm and indicator light alarm.

[0054] The image acquisition module transmits the collected images to the data storage processing platform through the image data transmission module, and the dust accumulation intelligent identification system in the data storage processing platform carries out dust accumulation intelligent identification, and transmits the specific identified data to the early warning control module, and carries out alarm processing according to whether the alarm threshold is reached.

[0055] Embodiment two:

[0056] Based on the above embodiment one, the air conditioning purification unit filter assembly dust accumulation identification and replacement early warning method is further proposed.

[0057] Please refer to Figure 1 - Figure 4 , the air conditioning purification unit filter assembly dust accumulation identification and replacement early warning method, at least comprising the following steps:

[0058] S1: based on experiments or related methods, the image information of the filter assembly dust accumulation and the morphological feature information of the filter assembly dust accumulation are obtained, which are combined and processed, so as to establish a typical feature database of the air conditioning filter assembly dust accumulation;

[0059] S2: by introducing integrated machine vision and learning algorithm, an identification model of the basic image is established, the parameters of the basic image identification model are set according to the device configuration and the image size, and the identification model of the basic image is used to process the image information of the air conditioning filter assembly, the image information at least including the preliminary identification and classification of the dust accumulation and the morphology in the image;

[0060] S3: The morphological feature information and image information of the processed filter assembly dust are input into a basic image recognition model for targeted training to form an automatic recognition basic model of the filter assembly dust capable of accurately identifying images in a filter assembly dust typical feature database, and the automatic recognition basic model of the filter assembly dust has certain recognition ability for new images, and the image recognition task is performed by the automatic recognition basic model of the filter assembly dust to generate a recognition result;

[0061] S4: An automatic filter assembly dust recognition model is built, and when the training result of the automatic filter assembly dust recognition basic model and the machine vision learning algorithm reaches a sufficient confidence level, a to-be-processed filter assembly dust amount image is obtained and input into the automatic filter assembly dust recognition model for recognition to obtain a to-be-processed filter assembly dust amount image recognition result;

[0062] In S4, the basic recognition model is trained before the to-be-processed image is obtained and input into the automatic filter assembly dust recognition model for recognition, that is, before the to-be-processed image is input into the established automatic filter assembly dust amount recognition model, the model is ensured to have the ability to correctly extract, understand and recognize the dust amount related features from the image;

[0063] The basic recognition model training at least includes the following steps:

[0064] The image information of the filter assembly dust amount is obtained and combined to form a filter assembly dust amount training;

[0065] The morphological features of the filter assembly dust amount in the training image are classified to obtain a plurality of groups of filter assembly dust amount training images;

[0066] Based on the morphological features of the filter assembly dust amount in the training image, the basic image recognition model is trained to form an automatic filter assembly dust amount recognition system;

[0067] Based on the filter assembly dust amount typical feature database construction method, the filter assembly dust amount range and morphological features are associated with the image in the form of an electronic tag to form a filter assembly dust amount training image;

[0068] The filter assembly dust amount is divided into a training set, a test set and a verification set according to a ratio of 5:4:1, which is used for training of the automatic filter assembly dust amount recognition model.

[0069] The automatic filter assembly dust recognition model is built at least including the following steps:

[0070] Based on the Backbone network structure in the machine vision learning algorithm, the filter assembly dust amount image is cut, the Neck network structure is used for feature extraction and fusion of the filter assembly dust amount image, and the Predicton network structure is used for image prediction and output of the filter assembly dust amount image;

[0071] The machine vision learning algorithm requires a pre-configured basic training model, which includes at least preset model selection, test data type, prediction result saving path, input picture size, confidence threshold, and result display setting;

[0072] Test the data type as picture, set the input picture resolution as 640x640 according to the computer parameters, and the confidence threshold as 0.80 by default;

[0073] Set the training rounds to 200, and train the pre-configured basic training model to ensure that the training result reaches a sufficient confidence level, thereby forming a reliable filter component dust amount automatic identification model.

[0074] The machine vision learning algorithm for constructing the filter component dust automatic identification model includes at least the following steps:

[0075] Use labelme to make labels for filter component dust training images to form json files, and then convert the json files to txt files;

[0076] Train the basic training model using json files and txt files to form a filter component dust amount automatic identification model, and use detect to realize filter component dust amount identification.

[0077] The model used by the machine vision learning algorithm and the size of the input picture are both referenced to the computer configuration parameters to improve the training speed of the basic model and the identification speed of the filter component dust amount automatic identification model.

[0078] After constructing the filter component dust amount automatic identification model using the machine vision learning algorithm, the filter component dust amount image to be processed needs to be obtained, and the method for obtaining the filter component dust amount image to be processed includes at least the following steps:

[0079] Use the image acquisition module to record the filter component dust amount image;

[0080] Adjust the size of the filter component dust amount image to the resolution of the input picture to form the filter component dust amount image to be processed for the filter component dust amount automatic identification model to identify.

[0081] The identification result of the processed image includes a picture result with confidence information, an EXCEL table result with filter component dust amount range and morphological feature information.

[0082] S5: While obtaining the result, randomly extract the identification result for verification analysis to exclude the influence of uncertain factors, improve the accuracy of the component dust automatic identification result, verify the reliability of the model, and complete iteration;

[0083] The random extraction of the identification result for verification analysis while obtaining the result comprises at least the following steps:

[0084] The dust amount of the filter assembly is analyzed by an optical microscope;

[0085] The optical microscope analysis result is compared with the identification result of the image to be processed

[0086] The reliability of the automatic identification model is verified, and the model is perfected;

[0087] The model is perfected and added to the filter assembly dust amount typical feature database.

[0088] S6: The filter assembly dust automatic identification model combines a deep learning algorithm convolutional neural network to model the change characteristics of the assembly dust pollution; a random forest ensemble learning method is used to construct a filter assembly dust pollution risk prediction model, further improving the accuracy and stability of the filter assembly dust automatic identification model;

[0089] S7: Based on the filter assembly dust pollution risk prediction model and the early warning control module, an intelligent replacement early warning system is built to monitor and analyze the filter assembly dust pollution risk in real time; when the filter assembly dust pollution condition reaches a certain threshold, an early warning is given in the form of an indicator light or a short message to remind the user to take measures to clean the filter assembly of the air conditioner.

[0090] It is apparent to those skilled in the art that the present application is not limited to the details of the foregoing exemplary embodiments, and that the present application can be implemented in other concrete forms without departing from the spirit or essential characteristics of the present application. Therefore, the embodiments should be considered in all respects as illustrative and not restrictive, and the scope of the present application is defined by the appended claims rather than the foregoing description, and it is intended to encompass all changes falling within the meaning and scope of the equivalent elements of the claims. Any reference signs in the claims should not be considered as limiting the claims involved.

Claims

1. A method for identifying and issuing early warnings about dust accumulation in air conditioning purification unit filter components, used in an air conditioning purification unit filter component dust accumulation identification and replacement early warning system. The air conditioning purification unit filter component dust accumulation identification and replacement early warning system includes an image acquisition module, which includes at least a 4K autofocus high-definition network camera, a high-resolution black and white camera, and a CCD pinhole high-definition network camera. It also includes an image data transmission module, a data storage and processing platform, a dust accumulation intelligent identification system, a dust accumulation data transmission module, and an early warning control module. The early warning control module includes at least three types of prompts during the early warning process: SMS notification alarm, display screen alarm, and indicator light alarm. The image acquisition module transmits the acquired images to the data storage and processing platform via the image data transmission module. The data storage and processing platform performs intelligent dust accumulation identification through the dust accumulation intelligent identification system and transmits the specifically identified data to the early warning control module. An alarm is triggered based on whether an alarm threshold has been reached. The method is characterized by: At least the following steps are included: S1: Based on experiments or related methods, obtain image information and morphological feature information of dust accumulation in the filter components, combine and process them to establish a typical feature database of dust accumulation in air conditioning filter components; S2: By introducing integrated machine vision and learning algorithms, a basic image recognition model is established. The parameters of the basic image recognition model are set according to the device configuration and image size. The basic image recognition model is used to process the image information of the air conditioner filter component. The image information includes at least the preliminary identification and classification of the amount and shape of dust accumulation in the image. S3: The morphological features and image information of the dust accumulation in the processed filter components are input into the basic image recognition model for targeted training, forming an automatic recognition basic model for dust accumulation in the filter components that can accurately identify images in the typical feature database of dust accumulation in the filter components, and has a certain recognition ability for new images. The automatic recognition basic model for dust accumulation in the filter components is used to perform image recognition tasks and generate recognition results. S4: Build an automatic dust accumulation recognition model for filter components. When the training results of the automatic dust accumulation recognition model for filter components and the machine vision learning algorithm reach a sufficient confidence level, obtain the image of the amount of dust accumulation on the filter components to be processed and input it into the automatic dust accumulation recognition model for filter components to be processed for recognition, and obtain the recognition result of the image of the amount of dust accumulation on the filter components to be processed. S5: While obtaining the results, randomly sample the identification results for verification and analysis to eliminate the influence of uncertain factors, improve the accuracy of the automatic dust accumulation identification results of components, verify the reliability of the model, and complete the iteration; The step of randomly sampling the identification results for verification and analysis while obtaining the results includes at least the following steps: The amount of dust accumulated in the filter components was analyzed using an optical microscope; Compare the optical microscope analysis results with the image recognition results of the dust accumulation of the filter component to be processed; Verify the reliability of the automatic identification model and improve the model; The model was improved and added to the database of typical characteristics of dust accumulation in filter components; S6: The automatic dust accumulation identification model of the filter component combines the deep learning algorithm convolutional neural network to model the changing characteristics of dust pollution in the component; and uses the random forest ensemble learning method to construct a component dust pollution risk prediction model; S7: Based on the component dust accumulation pollution risk prediction model and early warning control module, an intelligent replacement early warning system is built to monitor and analyze the component dust accumulation pollution risk in real time. When the dust accumulation pollution status of the air conditioning system components reaches a certain threshold, an early warning is issued through indicator lights or SMS notifications to remind users to take measures to clean the air conditioning filter components.

2. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 1, characterized in that: In step S4, before the image to be processed is acquired and input into the automatic dust collection model of the filter component, a basic recognition model is trained. That is, before the image to be processed is input into the established automatic dust collection model of the filter component, the model is ensured to have the ability to correctly extract, understand and recognize the dust collection-related features from the image. The training of the basic recognition model includes at least the following steps: The image information of dust accumulation in the filter components is obtained and combined to form a training function for dust accumulation in the filter components. Classify the morphological features of dust accumulation in the filter components in the training images to obtain multiple sets of training images of dust accumulation in the filter components. Based on the morphological features of dust accumulation on the filter components in the training images, the basic image recognition model is trained in a targeted manner to form an automatic dust accumulation identification system for filter components. Based on the method for constructing a database of typical features of dust accumulation in the filter components, the range and morphological features of dust accumulation in the filter components are associated with images in the form of electronic tags to form training images of dust accumulation in the filter components. The dust accumulation of the filter components is divided into training, testing, and validation sets in a ratio of 5:4:1 for training the automatic dust accumulation identification model of the filter components.

3. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 2, characterized in that: The construction of the automatic dust accumulation identification model for the filter component includes at least the following steps: Based on the machine vision learning algorithm, the Backbone network structure is used to segment the dust accumulation image of the filter component, the Neck network structure is used to extract and fuse features from the dust accumulation image of the filter component, and the Predicton network structure is used to predict and output the dust accumulation image of the filter component. Introducing machine vision learning algorithms requires a pre-configured basic training model. The pre-configuration includes at least the preset model selection, test data type, prediction result storage path, input image size, confidence threshold, and result display settings. The test is conducted using images as the data type. The input image resolution is set to 640×640 according to the computer parameters, and the confidence threshold is set to 0.80 by default. The training rounds are set to 200 to train the pre-configured basic training model and form a reliable automatic dust accumulation identification model for filter components.

4. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 3, characterized in that: Building an automatic dust recognition model for filter components using machine vision learning algorithms includes at least the following steps: Labelme was used to train images for dust accumulation in the filtering component to create labels, forming a JSON file, which was then converted to a TXT file. The basic training model is trained using JSON and TXT files to form an automatic dust accumulation identification model for filter components, and the dust accumulation of filter components is identified using the detect function.

5. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 4, characterized in that: After constructing an automatic dust accumulation recognition model for filter components using machine vision learning algorithms, it is necessary to obtain an image of the dust accumulation on the filter components to be processed. The method for obtaining the image of the dust accumulation on the filter components to be processed includes at least the following steps: The image acquisition module is used to record images of the dust accumulation on the filter components. The size of the dust accumulation image of the filter component is adjusted to the resolution of the input image to form the dust accumulation image of the filter component to be processed.

6. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 1, characterized in that: The image recognition results of the dust accumulation of the filter component to be processed include image results with confidence information, EXCEL table results with the range of dust accumulation of the filter component and morphological feature information.

7. The method for identifying and issuing early warnings about dust accumulation in the filter components of an air conditioning purification unit according to claim 5, characterized in that: The size of the input image for the machine vision learning algorithm is based on the computer's configuration parameters.

Images