Self-adaptive tiny dust identification method and cleaning device driven by variable-frequency airflow

Abstract

The invention belongs to the technical field of video image processing, and particularly relates to a self-adaptive tiny dust recognition method and a variable-frequency airflow-driven cleaning device.The self-adaptive tiny dust recognition method comprises the following steps that based on self-adaptive graying of a convolutional neural network, through a training network, the self-adaptive graying of the convolutional neural network is carried out; the system learns to extract color gray features which are most beneficial to dust particle identification under different environments and conditions. The invention provides an innovative self-adaptive tiny dust identification method, which is combined with a deep learning technology, multi-scale image processing and a self-adaptive thresholding algorithm, and can realize efficient and accurate tiny dust identification in various complex environments. Meanwhile, a dynamic cleaning device driven by variable-frequency airflow is further designed, the intensity and frequency of the airflow can be automatically adjusted according to the mote coverage degree recognized in real time, it is ensured that the cleaning effect is maximized, and excessive cleaning or insufficient cleaning is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of video image processing, and in particular to an adaptive dust recognition method and a cleaning device driven by variable frequency airflow. Background Art

[0002] With the rapid development of modern industry and electronic equipment, precision equipment such as cameras, sensors, and monitoring equipment have been widely used in outdoor and industrial environments. These devices are usually exposed to harsh environmental conditions for a long time, such as dust, pollutants, and other dust particles, which will accumulate on the surface of the equipment, affecting the performance and reliability of the equipment, and even causing equipment failure or functional failure. Therefore, how to effectively identify and clean the dust on the surface of the equipment has become an important technical problem to improve equipment performance and extend its service life.

[0003] Traditional dust cleaning methods usually rely on manual cleaning or timed cleaning, which is not only inefficient, but also prone to inadequate cleaning or over-cleaning, and cannot achieve automation and precise control. At the same time, most existing dust recognition methods have poor robustness in complex environments, low recognition accuracy, and cannot adapt to the limitations of different dust shapes and sizes, resulting in unsatisfactory dust detection results.

[0004] Therefore, an adaptive dust identification method and a variable frequency airflow driven cleaning device are needed. Summary of the invention

[0005] Based on the existing technical problems, the present invention proposes an adaptive dust identification method and a variable frequency airflow driven cleaning device.

[0006] The present invention proposes an adaptive dust identification method, which includes the following steps: Step 1, adaptive grayscale conversion based on a convolutional neural network. By training the network, the system learns to extract the color grayscale features that are most conducive to dust particle identification under different environments and conditions.

[0007] Step 2: Blur analysis and multi-scale edge detection. Blur measurement can be used to detect changes in image clarity, and multi-scale edge detection can be used to extract dust features at different scales.

[0008] Step 3: Adaptive thresholding and dust coverage calculation: dynamically adjust the threshold according to the local grayscale distribution of the image to accurately segment the dust area.

[0009] Step 4: Dynamic cleaning trigger and variable frequency airflow control, according to the dust coverage , when it exceeds the preset threshold When the system automatically triggers the cleaning device, the cleaning device driven by variable frequency airflow adjusts the intensity and frequency of the airflow according to the real-time dust coverage to ensure the cleaning effect.

[0010] Preferably, the image adaptive grayscale formula in step 1 is expressed as follows: ;

[0011] in: is the image at position Gray value of

[0012] is the input color image at position Where value, Indicates red, Indicates green, Indicates blue;

[0013] It is a trained CNN model that can adaptively extract grayscale images suitable for dust mote recognition;

[0014] Represents the parameters of the network.

[0015] Preferably, the fuzziness measurement formula in step 2 is as follows: ;

[0016] in, is the image position The blurriness measure of , which indicates the clarity of the image;

[0017] is the image at position The gradient at is calculated by the Sobel operator;

[0018] is the radius of the convolution kernel, which is used to define the neighborhood size for blur calculation;

[0019] Indicates the horizontal offset position of the pixel. Indicates the vertical offset position of the pixel;

[0020] Indicates the accumulator. Indicates absolute value;

[0021] Then the convolution kernel slides rightward and downward to form a blurred feature map M;

[0022] Multi-scale edge detection is used to capture dust particles at different scales. The calculation formula is as follows:

[0023] ;

[0024] in, The standard deviation is The Gaussian kernel function is defined as: ;

[0025] represents the Laplace operator, which is used to calculate the second-order derivative of the image; Represented as a feature map after multi-scale edge detection, It is represented as a fuzzy feature map after fuzziness measurement; Indicates The exponential operation with base , where .

[0026] Preferably, the adaptive thresholding formula in step 3 is expressed as follows: ;

[0027] in, It's location Adaptive threshold at ;

[0028] and are the mean and standard deviation of the local area of ​​the image, respectively, which are calculated by sliding window, and this image is formed after blur measurement and multi-scale edge detection;

[0029] is a weighting factor used to adjust the sensitivity of the threshold;

[0030] Once the threshold is determined, it is binarized by comparing it with the image grayscale value to obtain the mask of the dust area: ;

[0031] Finally, calculate the dust coverage index : ;

[0032] in, is the total number of pixels in the image, and are the width and height of the image respectively;

[0033] is the binarization result of the dust area, if the pixel belongs to the dust area, it is 1, otherwise it is 0; Indicates that the binary values ​​in the entire mask area are accumulated.

[0034] Preferably, the variable frequency airflow control formula in step 4 is expressed as follows: ;

[0035] in, For the moment Air flow velocity under

[0036] is the basic airflow speed, indicating the system default minimum airflow speed;

[0037] is the control constant used to adjust the airflow intensity;

[0038] For the current moment The dust coverage under

[0039] is the target dust coverage value. When it is set to 0, it means complete cleaning. Through the variable frequency control formula, the system can adjust the airflow intensity in real time so that it is proportional to the dust coverage, ensuring that the dust on the surface of the equipment is effectively removed.

[0040] The present invention provides a variable frequency airflow driven cleaning device, comprising a fan assembly and an airflow control system, wherein the fan assembly is composed of a variable frequency motor, a variable frequency controller, a high-speed impeller, an air duct, a fan housing, a bracket, a hard nozzle and a nozzle adjustment motor;

[0041] Both side surfaces of the fan housing are hinged to the inner walls at both ends of the U-shape of the bracket and are locked in place by locking screws. A camera is also fixedly installed on the inner wall of the bottom left end of the fan housing, and the inner bottom wall of the right end of the fan housing is fixedly installed on the mounting surface of the variable frequency motor.

[0042] Preferably, the frequency conversion controller is installed on the inner wall of the fan housing and is located on one side of the frequency conversion motor, the output end of the frequency conversion motor is fixedly installed with the rotating shaft of the high-speed impeller, the air duct is installed on the inner top wall of the fan housing, and the air inlet end of the air duct corresponds to the air outlet end of the high-speed impeller, the air outlet end of the air duct is arc-shaped, and a flexible shell is clamped on the inner wall of the air outlet end of the air duct, the air outlet end of the flexible shell is fixedly installed with the surface of the hard nozzle, the outer surface of the air outlet end of the air duct is rotatably connected with a driving shaft through a bearing, one end of the two driving shafts is fixedly installed with a rotating shell, the inner wall of the groove of the rotating shell is clamped with the surface of the hard nozzle, the surface of the flexible shell is provided with a mounting hole, the inner wall of the mounting hole is fixedly installed with the mounting surface of the nozzle adjustment motor, and the output shaft of the nozzle adjustment motor is fixedly installed with the surface of a corresponding one of the driving shafts through a coupling.

[0043] Preferably, the airflow control system includes an airflow intensity adjustment module, a nozzle direction control module and an airflow frequency adjustment module; the airflow control system dynamically adjusts the airflow intensity, frequency, injection direction and airflow concentration according to the real-time feedback data provided by the dust recognition module;

[0044] The airflow intensity adjustment module is based on the real-time dust coverage Data, dynamically adjust the strength of the airflow to ensure that dust particles are effectively removed while avoiding over-cleaning or insufficient cleaning;

[0045] The airflow intensity adjustment module adopts the PID control formula, which is used to smoothly and accurately adjust the airflow intensity. The PID control formula calculates a control quantity based on the error value to drive the adjustment of the fan speed. The PID control formula is: ;

[0046] in, To control the amount; is the proportional error; , and are the proportional, integral and differential gain coefficients of the PID controller respectively; Expressed as the integral error, It is expressed as differential error.

[0047] Preferably, the nozzle direction control module is used to accurately adjust the angle of the nozzle and the injection direction of the airflow. The nozzle is adjusted according to the real-time operating status of the equipment and the distribution of the dust particles. The nozzle direction control changes the angle of the nozzle through a mechanical transmission device. The direction adjustment adopts an angle control formula, which is: ;

[0048] in, For the moment The angle of the nozzle; is the control constant, which indicates the sensitivity between dust coverage and nozzle angle; is the offset angle, indicating the default orientation of the nozzle; It is the real-time dust coverage, which affects the adjustment of the nozzle angle.

[0049] Preferably, the airflow frequency adjustment module intelligently adjusts the frequency of the airflow according to the real-time dust coverage DCI; the airflow frequency adjustment module ensures the high efficiency of the cleaning process by increasing the airflow frequency in the polluted area; and saves energy by reducing the airflow frequency in the lightly polluted area;

[0050] Airflow frequency Dust coverage Proportionally, the airflow frequency is adjusted by the following formula: ;

[0051] in, For the moment The airflow frequency; is the basic airflow frequency, indicating the minimum frequency of the airflow; is the frequency sensitivity coefficient, which indicates the influence of dust coverage on frequency;

[0052] when When it is larger, the airflow frequency will increase, thereby improving the cleaning effect of the airflow;

[0053] when When it is smaller, the frequency will be reduced to reduce energy consumption.

[0054] The beneficial effects of the present invention are:

[0055] The present invention proposes an innovative adaptive dust identification method, which combines deep learning technology, multi-scale image processing and adaptive thresholding algorithm to achieve efficient and accurate dust identification in various complex environments. At the same time, the present invention also designs a dynamic cleaning device driven by variable frequency airflow, which can automatically adjust the intensity and frequency of the airflow according to the dust coverage identified in real time to ensure the maximum cleaning effect and avoid excessive or insufficient cleaning. This technology can not only realize the automation of dust cleaning, but also has strong adaptability and robustness, and can operate stably in complex industrial and outdoor environments, meeting the cleaning and maintenance needs of modern precision equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a structural schematic diagram of an adaptive dust identification method and a cleaning device driven by variable frequency airflow;

[0057] Figure 2 A three-dimensional diagram of the fan housing structure of an adaptive dust identification method and a variable frequency airflow driven cleaning device;

[0058] Figure 3 It is a three-dimensional exploded diagram of a fan assembly of an adaptive dust identification method and a variable frequency airflow driven cleaning device;

[0059] Figure 4 A cross-sectional view of an air duct structure of an adaptive dust identification method and a cleaning device driven by variable frequency airflow;

[0060] Figure 5 A three-dimensional diagram of the rotating shell structure of an adaptive dust identification method and a variable frequency airflow driven cleaning device;

[0061] Figure 6 A three-dimensional diagram of a flexible housing structure of an adaptive dust identification method and a variable frequency airflow driven cleaning device;

[0062] Figure 7 A three-dimensional diagram of the hard nozzle structure of an adaptive dust identification method and a variable frequency airflow driven cleaning device.

[0063] In the figure: 1. Fan assembly; 2. Frequency conversion motor; 3. Frequency conversion controller; 4. High-speed impeller; 5. Air duct; 6. Fan housing; 7. Bracket; 8. Hard nozzle; 9. Nozzle adjustment motor; 10. Camera; 11. Flexible housing; 12. Drive shaft; 13. Rotating housing; 14. Mounting hole. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0065] Embodiment 1

[0066] An adaptive dust particle identification method comprises the following steps: Step 1, adaptive grayscale conversion based on a convolutional neural network, wherein the system learns to extract the color grayscale features that are most conducive to dust particle identification under different environments and conditions by training the network.

[0067] The adaptive grayscale formula of the image in step 1 is expressed as follows: ;

[0068] in: is the image at position Gray value of

[0069] is the input color image at position Where value, Indicates red, Indicates green, Indicates blue;

[0070] It is a trained CNN model that can adaptively extract grayscale images suitable for dust mote recognition;

[0071] Represents the parameters of the network.

[0072] The network learns the optimal image grayscale method through the training set, making the dust particles more prominent in the converted image.

[0073] Step 2: Blur analysis and multi-scale edge detection. Blur measurement can be used to detect changes in image clarity, and multi-scale edge detection can be used to extract dust features at different scales.

[0074] The fuzziness measurement formula in step 2 is as follows: ;

[0075] in, is the image position The blurriness measure of , which indicates the clarity of the image;

[0076] is the image at position The gradient at is calculated by the Sobel operator;

[0077] is the radius of the convolution kernel, which is used to define the neighborhood size for blur calculation;

[0078] Indicates the horizontal offset position of the pixel. Indicates the vertical offset position of the pixel;

[0079] Indicates the accumulator. Indicates absolute value;

[0080] Then the convolution kernel slides rightward and downward to form a blurred feature map M;

[0081] Multi-scale edge detection is used to capture dust particles at different scales. The calculation formula is as follows:

[0082] ;

[0083] in, The standard deviation is The Gaussian kernel function is defined as: ;

[0084] represents the Laplace operator, which is used to calculate the second-order derivative of the image; Represented as a feature map after multi-scale edge detection, It is represented as a fuzzy feature map after fuzziness measurement. Indicates The exponential operation with base , where .

[0085] Through multi-scale calculation, the edge features of dust particles can be detected at different scales, especially for particles of different sizes and shapes, it has strong adaptability.

[0086] Step 3: Adaptive thresholding and dust coverage calculation: dynamically adjust the threshold according to the local grayscale distribution of the image to accurately segment the dust area.

[0087] The adaptive thresholding formula in step 3 is expressed as follows: ;

[0088] in, It's location Adaptive threshold at ;

[0089] and are the mean and standard deviation of the local area of ​​the image, respectively, which are calculated by sliding window, and this image is formed after blur measurement and multi-scale edge detection;

[0090] is a weighting factor used to adjust the sensitivity of the threshold;

[0091] Once the threshold is determined, it is binarized by comparing it with the image grayscale value to obtain the mask of the dust area: ;

[0092] Finally, calculate the dust coverage index : ;

[0093] in, is the total number of pixels in the image, and are the width and height of the image respectively;

[0094] is the binarization result of the dust area. If the pixel belongs to the dust area, it is 1, otherwise it is 0. Indicates that the binary values ​​in the entire mask area are accumulated.

[0095] Step 4: Dynamic cleaning trigger and variable frequency airflow control, according to the dust coverage , when it exceeds the preset threshold When the system automatically triggers the cleaning device, the cleaning device driven by variable frequency airflow adjusts the intensity and frequency of the airflow according to the real-time dust coverage to ensure the cleaning effect.

[0096] The variable frequency airflow control formula in step 4 is as follows: ;

[0097] in, For the moment Air flow velocity under

[0098] is the basic airflow speed, indicating the system default minimum airflow speed;

[0099] is the control constant used to adjust the airflow intensity;

[0100] For the current moment The dust coverage under

[0101] It is the target dust coverage value, which is usually set to 0, that is, complete cleaning is achieved; through the variable frequency control formula, the system can adjust the airflow intensity in real time to make it proportional to the dust coverage, ensuring that the dust on the surface of the equipment is effectively removed.

[0102] The innovative dust identification method of the present invention can accurately identify dust particles of different sizes and shapes by combining deep learning, adaptive thresholding and multi-scale edge detection. Through the variable frequency airflow control mechanism that dynamically adjusts the airflow intensity, efficient and accurate automatic cleaning can be achieved based on the real-time dust identification results. This method shows superior performance and flexibility in dust identification and cleaning applications in complex environments.

[0103] Through specific mathematical calculations and practical examples, it shows how to apply these technologies in the dust particle identification process.

[0104] (1) Adaptive grayscale conversion based on convolutional neural network; Assume that there is an RGB image of size 5×5, and the RGB value of each pixel in the image is as follows:

[0105] ; Now, process this image through a convolutional neural network and perform adaptive grayscale conversion. Assume that the grayscale value output by the trained CNN model is: ; For pixel position , after the output of CNN, the grayscale value obtained is as follows:

[0106] ; This matrix is ​​the result of adaptive grayscale conversion of color images through the CNN model. It not only converts the color information but also highlights the dust particles.

[0107] (2) Fuzzy analysis and multi-scale edge detection; for grayscale images , using blur analysis to calculate the sharpness of the image.

[0108] Assumption Selection The blur is calculated in the neighborhood of and the gradient of the image is calculated using the Sobel operator. , the gradient of the image is calculated by the Sobel operator:

[0109] ;

[0110] The calculation results are: ; ;

[0111] Then calculate the absolute value of the gradient: ; For image position The 3*3 neighborhood is calculated to obtain the image position Fuzziness measure .

[0112] Continue sliding across the entire image and performing similar calculations to obtain the image blur metric map. Next, use the LoG algorithm to perform multi-scale edge detection. Perform convolution operations using Gaussian blur and Laplace operators. Assume that At the scale of , the following edge detection results are obtained: ; This value indicates that there are more obvious edge features of dust particles at this location.

[0113] (3) Adaptive thresholding and dust coverage calculation; Assume that after adaptive thresholding, the local mean is set and standard deviation They are: ; ; ;

[0114] Then the location Adaptive threshold for: ;

[0115] Through binarization processing, we get: ;

[0116] for ,have ; Continue to perform binarization on the entire image to obtain the mask of the dust area , and calculate the dust coverage index DCI.

[0117] Assume that after binarization, the binary image of the dust area is as follows: ; Calculate the dust coverage index DCI: ; , indicating that about 52% of the image area is covered by dust.

[0118] (4) Dynamic cleaning trigger and variable frequency airflow control; based on the calculated dust coverage , assuming the preset cleaning threshold , so when the DCI exceeds the threshold, the system triggers the cleaning device.

[0119] In terms of airflow control, the basic airflow speed is set , control constant ,according to , target value , the air flow velocity is calculated:

[0120] ;

[0121] At this time, the airflow intensity is 11.04m / s. According to the dust coverage, the cleaning device starts and provides a stronger airflow to ensure that the dust on the lens surface is effectively removed.

[0122] Through these examples, we show how to accurately identify dust particles and dynamically adjust the cleaning strategy from image preprocessing, including adaptive grayscale, fuzziness analysis, edge detection, calculation of dust coverage, trigger control of cleaning devices, etc. The application of innovative algorithms enables the system to efficiently cope with different environments and changes in dust particles, ensuring long-term and efficient operation of the equipment.

[0123] Embodiment 2

[0124] Reference Figure 1-Figure 7 A variable frequency airflow driven cleaning device includes a fan assembly 1 and an airflow control system. The fan assembly 1 is composed of a variable frequency motor 2, a variable frequency controller 3, a high-speed impeller 4, an air duct 5, a fan housing 6, a bracket 7, a hard nozzle 8 and a nozzle adjustment motor 9.

[0125] The device has a built-in variable frequency fan, which controls the strength and frequency of the airflow by adjusting the fan speed. The system automatically adjusts the strength of the airflow according to the real-time detected dust coverage and cleaning needs to avoid over-cleaning or insufficient cleaning. The variable frequency fan can adjust the airflow output in real time according to the cleaning needs, making the airflow more flexible and adaptable during the dust cleaning process.

[0126] The fan assembly 1 is one of the core components of the device, responsible for generating and regulating the airflow for effective dust cleaning. The component has a built-in variable frequency fan, and the fan speed is adjusted in real time through the control system to accurately control the intensity, frequency and direction of the airflow, thereby ensuring the maximum cleaning effect while avoiding over-cleaning or under-cleaning. The following is a detailed design of the fan assembly 1, covering the working principle, control method and structural requirements of the fan, providing a more detailed design basis for structural engineers. The variable frequency fan changes the output intensity of the airflow by adjusting its speed. Unlike traditional fans, variable frequency fans do not need to change the airflow through physical resistance or adjusting the blade angle, but adjust the motor speed through frequency conversion, which can finely control the intensity and frequency of the airflow without sacrificing the air volume.

[0127] The speed of the variable frequency fan is controlled by the inverter, which automatically adjusts the fan speed according to the real-time feedback of dust coverage DCI and cleaning requirements.

[0128] The two side surfaces of the fan casing 6 are hinged to the inner walls at both ends of the U-shape of the bracket 7 and are locked in place by locking screws. A camera 10 is also fixedly installed on the inner wall of the bottom left end of the fan casing 6, and the inner bottom wall of the right end of the fan casing 6 is fixedly installed on the mounting surface of the variable frequency motor 2.

[0129] The variable frequency motor 2 is the core component of the fan drive and can run smoothly in a wide range of speeds. According to the design requirements, the motor power selection should be reasonably selected according to the size of the cleaning area and the required air flow. The common power range is 0.5kW to 15kW.

[0130] The fan impeller needs to be designed to meet different airflow and pressure requirements. The impeller type can be centrifugal, axial or mixed flow, and the specific choice depends on the airflow requirements, size restrictions and space configuration of the equipment. Centrifugal fans are suitable for clean environments that require higher air pressure, while axial fans are more suitable for applications with large air volumes.

[0131] The frequency conversion controller 3 is installed on the inner wall of the fan housing 6 and is located on one side of the frequency conversion motor 2. The output end of the frequency conversion motor 2 is fixedly installed with the rotating shaft of the high-speed impeller 4. The air duct 5 is installed on the inner top wall of the fan housing 6, and the air inlet end of the air duct 5 corresponds to the air outlet end of the high-speed impeller 4. The air outlet end of the air duct 5 is arc-shaped, and the inner wall of the air outlet end of the air duct 5 is clamped with a flexible shell 11. The air outlet end of the flexible shell 11 is fixedly installed with the surface of the hard nozzle 8. The outer surface of the air outlet end of the air duct 5 is rotatably connected with a drive shaft 12 through a bearing. One end of the two drive shafts 12 is fixedly installed with a rotating shell 13. The inner wall of the groove of the rotating shell 13 is clamped with the surface of the hard nozzle 8. The surface of the flexible shell 11 is provided with a mounting hole 14. The inner wall of the mounting hole 14 is fixedly installed with the mounting surface of the nozzle adjustment motor 9. The output shaft of the nozzle adjustment motor 9 is fixedly installed with the surface of the corresponding drive shaft 12 through a coupling.

[0132] The fan housing needs to ensure stable airflow transmission and reduce airflow loss. It is usually made of metal or composite materials to ensure sufficient strength and durability. The design of the fan bracket 7 should take into account the vibration characteristics of the equipment to ensure that the fan is installed firmly and vibration-resistant. In the installation position, the bracket 7 should be able to withstand the operating load of the fan and the mechanical stress caused by long-term operation.

[0133] In order to maximize the airflow efficiency, the fan assembly 1 needs to be equipped with a suitable air duct 5. The function of the air duct 5 is to adjust the distribution of the airflow, ensure that the airflow can be concentrated and transmitted to the area that needs to be cleaned, and avoid airflow leakage or diffusion, which wastes energy.

[0134] The adjustable nozzle design can automatically adjust the spray angle, airflow distribution and pressure of the airflow according to different pollution conditions. The nozzle can adjust the direction through the automatic control system to concentrate the airflow for effective cleaning and avoid ineffective or excessive airflow loss.

[0135] The airflow control system includes an airflow intensity adjustment module, a nozzle direction control module and an airflow frequency adjustment module; the airflow control system dynamically adjusts the airflow intensity, frequency, spray direction and airflow concentration according to the real-time feedback data provided by the dust recognition module;

[0136] The airflow intensity adjustment module is based on the real-time dust coverage Data, dynamically adjust the strength of the airflow to ensure that dust particles are effectively removed while avoiding over-cleaning or insufficient cleaning;

[0137] The airflow intensity adjustment module adopts the PID control formula, which is used to smoothly and accurately adjust the airflow intensity. The PID control formula calculates a control quantity based on the error value to drive the adjustment of the fan speed. The PID control formula is: ;

[0138] in, To control the amount; is the error; , and are the proportional, integral and differential gain coefficients of the PID controller respectively; Expressed as the integral error, It is expressed as differential error.

[0139] The nozzle direction control module is used to accurately adjust the nozzle angle and the spray direction of the airflow. The nozzle is adjusted through the real-time operating status of the equipment and the distribution of dust particles. The nozzle direction control changes the nozzle angle through a mechanical transmission device. The direction adjustment adopts the angle control formula, which is: ;

[0140] in, For the moment The angle of the nozzle; is the control constant, which indicates the sensitivity between dust coverage and nozzle angle; is the offset angle, indicating the default orientation of the nozzle; It is the real-time dust coverage, which affects the adjustment of the nozzle angle.

[0141] Nozzle angle control mechanism: According to the working status of the equipment, such as the actual position of the lens, the nozzle will automatically adjust to suit different cleaning needs.

[0142] The airflow frequency adjustment module intelligently adjusts the airflow frequency according to the real-time dust coverage DCI. The airflow frequency adjustment module ensures the efficiency of the cleaning process by increasing the airflow frequency in the polluted area, and saves energy by reducing the airflow frequency in the less polluted area.

[0143] Airflow frequency Dust coverage Proportionally, the airflow frequency is adjusted by the following formula: ;

[0144] in, For the moment The airflow frequency; is the basic airflow frequency, indicating the minimum frequency of the airflow; is the frequency sensitivity coefficient, which indicates the influence of dust coverage on frequency;

[0145] when When it is larger, the airflow frequency will increase, thereby improving the cleaning effect of the airflow.

[0146] when When it is smaller, the frequency will be reduced to reduce unnecessary energy consumption.

[0147] The system detects dust coverage in real time, automatically starts the cleaning function according to the DCI index and cleaning needs, and intelligently selects the airflow mode. If the cleaning process fails to effectively remove dust, the system will make adjustments through a real-time feedback mechanism to ensure that each cleaning can achieve the expected effect. The control system also includes a timed cleaning mode and manual intervention reminder function to ensure that the equipment is continuously in the best working condition.

[0148] Example: Assume , then the airflow frequency is calculated according to the control formula:

[0149] ; At this time, the airflow frequency is 13.5Hz, which means that in areas with heavier pollution, the airflow frequency is higher, which helps to completely remove the dust.

[0150] Step 1: Airflow intensity adjustment module: Dynamically adjust the airflow intensity according to the real-time dust coverage DCI data to ensure that the dust particles are effectively removed while avoiding over-cleaning or insufficient cleaning.

[0151] Assumption: Base air velocity ; Control constant ; Current dust coverage ; The target cleaning intensity is , then completely clean.

[0152] According to the formula: At this time, the airflow intensity is 11.2m / s, and the system automatically increases the airflow intensity according to the higher dust coverage to ensure effective cleaning.

[0153] Control algorithm: Use PID control to smoothly adjust the airflow intensity to ensure that the airflow intensity is adjusted to the optimal state.

[0154] Assumptions: , target value ; , , ;PID constants.

[0155] Calculation error: ;

[0156] Calculate PID control quantity: ;

[0157] Assume that at time When the error is 0.6, the integral part is 0, and the differential part is 0, then the control amount is: ; According to the PID control quantity, the fan speed can be dynamically adjusted to an appropriate level, thereby achieving precise control of the airflow intensity.

[0158] Step 2: Nozzle direction control module: Control the angle of the nozzle and the direction of the airflow to ensure that the airflow can be accurately concentrated on the area on the surface of the equipment that needs to be cleaned.

[0159] Example: Assume , the control constant is that for every 0.1 increase in DCI value, the nozzle angle increases by 30°; Default nozzle direction. Current dust coverage .

[0160] Calculate the nozzle angle according to the formula: ; At this time, the angle of the nozzle is 34°, ensuring that the airflow is concentrated on the surface of the equipment and cleans efficiently.

[0161] Nozzle direction adjustment mechanism: Assuming that the dust coverage changes during the equipment cleaning process; When cleaning, the nozzle angle is 34° to focus on cleaning the most polluted areas.

[0162] When the dust coverage is reduced to , according to the formula: ; At this time, the nozzle angle is 22°, and the airflow concentration is appropriately reduced to avoid excessive cleaning.

[0163] Step 3: Airflow frequency adjustment module, based on real-time dust coverage , intelligently adjusts the frequency of airflow to ensure the cleaning process is efficient and saves energy.

[0164] Assumption: Basic airflow frequency ; Control constant ; Current dust coverage .

[0165] According to the formula: ; At this time, the air flow frequency is 12.5Hz, ensuring efficient cleaning of the cleaning area.

[0166] Frequency adjustment after cleaning: After the equipment is cleaned, the dust coverage , according to the formula: ; Airflow frequency is reduced, saving energy and avoiding over-cleaning.

[0167] The above examples show how each module can ensure the high efficiency and energy efficiency of the cleaning process by dynamically adjusting the airflow intensity, nozzle angle and airflow frequency. The control mechanism of each module clearly describes the impact of dust coverage on the cleaning effect through mathematical formulas, and ensures the flexible adaptation and efficient execution of the system through specific parameter adjustments.

[0168] The airflow control system of the present invention uses three modules: airflow intensity adjustment module, nozzle direction control module and airflow frequency adjustment module to accurately control the clean airflow, ensuring that each link is adaptively adjusted according to the real-time dust coverage DCI. The specific mathematical model and formula provide a theoretical basis for the real-time adjustment of the system, and through methods such as PID control, angle adjustment and frequency adjustment, efficient, accurate and flexible dust cleaning is achieved. These innovative modules effectively improve cleaning efficiency, reduce energy consumption, and ensure the best protection of the equipment during the cleaning process.

[0169] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An adaptive dust identification method, characterized in that: The method comprises the following steps: Step 1: Adaptive grayscale conversion based on a convolutional neural network. By training the network, the system learns to extract the color grayscale features that are most conducive to the identification of dust particles under different environments and conditions; Step 2: fuzziness analysis and multi-scale edge detection. Through fuzziness measurement, the clarity change of the image can be detected, and the dust features at different scales can be extracted in combination with multi-scale edge detection. Step 3: Adaptive thresholding and dust coverage calculation, dynamically adjusting the threshold according to the local grayscale distribution of the image, so as to accurately segment the dust area; Step 4: Dynamic cleaning trigger and variable frequency airflow control, according to the dust coverage , when it exceeds the preset threshold When the system automatically triggers the cleaning device, the cleaning device driven by variable frequency airflow adjusts the intensity and frequency of the airflow according to the real-time dust coverage to ensure the cleaning effect.

2. The adaptive dust identification method according to claim 1, characterized in that: The adaptive grayscale formula in step 1 is as follows: ; in: is the image at position Gray value of is the input color image at position Where value, Indicates red, Indicates green, Indicates blue; It is a trained CNN model that can adaptively extract grayscale images suitable for dust mote recognition; Represents the parameters of the network.

3. The adaptive dust identification method according to claim 1, characterized in that: The fuzziness measurement formula in step 2 is as follows: ; in, is the image position The blurriness measure of , which indicates the clarity of the image; is the image at position The gradient at is calculated by the Sobel operator; is the radius of the convolution kernel, which is used to define the neighborhood size for blur calculation; Indicates the horizontal offset position of the pixel. Indicates the vertical offset position of the pixel; Indicates the accumulator. Indicates absolute value; Then the convolution kernel slides rightward and downward to form a blurred feature map M; Multi-scale edge detection is used to capture dust particles at different scales. The calculation formula is as follows: ； in, The standard deviation is The Gaussian kernel function is defined as: ; represents the Laplace operator, which is used to calculate the second-order derivative of the image; Represented as a feature map after multi-scale edge detection, It is represented as a fuzzy feature map after fuzziness measurement; Indicates The exponential operation with base , where .

4. The adaptive dust identification method according to claim 1, characterized in that: The adaptive thresholding formula in step 3 is expressed as follows: ; in, It's location Adaptive threshold at ; and are the mean and standard deviation of the local area of ​​the image, respectively, which are calculated by sliding window, and this image is formed after blur measurement and multi-scale edge detection; is a weighting factor used to adjust the sensitivity of the threshold; Once the threshold is determined, it is binarized by comparing it with the image grayscale value to obtain the mask of the dust area: ; Finally, calculate the dust coverage index : ; in, is the total number of pixels in the image, and are the width and height of the image respectively; is the binarization result of the dust area, if the pixel belongs to the dust area, it is 1, otherwise it is 0; Indicates that the binary values ​​in the entire mask area are accumulated.

5. The adaptive dust identification method according to claim 1, characterized in that: The variable frequency airflow control formula in step 4 is as follows: ; in, For the moment Air flow velocity under is the basic airflow speed, which indicates the system default minimum airflow speed; is the control constant used to adjust the airflow intensity; For the current moment The dust coverage under is the target dust coverage value. When it is set to 0, it means complete cleaning. Through the variable frequency control formula, the system can adjust the airflow intensity in real time so that it is proportional to the dust coverage, ensuring that the dust on the surface of the equipment is effectively removed.

6. A variable frequency airflow driven cleaning device, using the adaptive dust identification method according to any one of claims 1 to 5, characterized in that: The invention comprises a fan assembly (1) and an airflow control system, wherein the fan assembly (1) is composed of a variable frequency motor (2), a variable frequency controller (3), a high-speed impeller (4), an air duct (5), a fan housing (6), a bracket (7), a hard nozzle (8) and a nozzle adjustment motor (9); Both side surfaces of the fan housing (6) are hingedly connected to the inner walls at both ends of the U-shape of the bracket (7) and are locked in place by means of locking screws. A camera (10) is also fixedly mounted on the inner side wall of the bottom left end of the fan housing (6). The inner bottom wall of the right end of the fan housing (6) is fixedly mounted on the mounting surface of the variable frequency motor (2).

7. A variable frequency airflow driven cleaning device according to claim 6, characterized in that: The variable frequency controller (3) is mounted on the inner side wall of the fan housing (6) and is located on one side of the variable frequency motor (2); the output end of the variable frequency motor (2) is fixedly mounted on the rotating shaft of the high-speed impeller (4); the air duct (5) is mounted on the inner top wall of the fan housing (6); the air inlet end of the air duct (5) corresponds to the air outlet end of the high-speed impeller (4); the air outlet end of the air duct (5) is arc-shaped; and a flexible shell (11) is clamped on the inner wall of the air outlet end of the air duct (5); the air outlet end of the flexible shell (11) is connected to the surface of the hard nozzle (8). The air outlet of the air duct (5) is fixedly installed, the outer surface of the air outlet end of the air duct (5) is rotatably connected to a drive shaft (12) via a bearing, one end of each of the two drive shafts (12) is fixedly installed with a rotating shell (13), the inner wall of the groove of the rotating shell (13) is clamped with the surface of the hard nozzle (8), the surface of the flexible shell (11) is provided with a mounting hole (14), the inner wall of the mounting hole (14) is fixedly installed with the mounting surface of the nozzle adjustment motor (9), and the output shaft of the nozzle adjustment motor (9) is fixedly installed with the surface of a corresponding drive shaft (12) via a coupling.

8. The variable frequency airflow driven cleaning device according to claim 7, characterized in that: The airflow control system includes an airflow intensity adjustment module, a nozzle direction control module and an airflow frequency adjustment module; the airflow control system dynamically adjusts the airflow intensity, frequency, spray direction and airflow concentration according to the real-time feedback data provided by the dust recognition module; The airflow intensity adjustment module is based on the real-time dust coverage Data, dynamically adjust the strength of the airflow to ensure that dust particles are effectively removed while avoiding over-cleaning or insufficient cleaning; The airflow intensity adjustment module adopts PID control formula, which is used to adjust the airflow intensity smoothly and accurately; The PID control formula calculates a control quantity based on the error value to drive the adjustment of the fan speed. The PID control formula is: ; in, To control the amount; is the proportional error; , and are the proportional, integral and differential gain coefficients of the PID controller respectively; Expressed as the integral error, It is expressed as differential error.

9. A variable frequency airflow driven cleaning device according to claim 8, characterized in that: The nozzle direction control module is used to accurately adjust the angle of the nozzle and the spray direction of the airflow. The nozzle is adjusted according to the real-time operating status of the equipment and the distribution of the dust particles. The nozzle direction control changes the angle of the nozzle through a mechanical transmission device. The direction adjustment adopts an angle control formula, which is: ; in, For the moment The angle of the nozzle; is the control constant, which indicates the sensitivity between dust coverage and nozzle angle; is the offset angle, indicating the default orientation of the nozzle.

10. A variable frequency airflow driven cleaning device according to claim 9, characterized in that: The airflow frequency adjustment module intelligently adjusts the airflow frequency according to the real-time dust coverage DCI; the airflow frequency adjustment module ensures the high efficiency of the cleaning process by increasing the airflow frequency in the polluted area; and saves energy by reducing the airflow frequency in the lightly polluted area; Airflow frequency Dust coverage Proportionally, the airflow frequency is adjusted by the following formula: ; in, For the moment The airflow frequency; is the basic airflow frequency, indicating the minimum frequency of the airflow; is the frequency sensitivity coefficient, which indicates the influence of dust coverage on frequency; when When it is larger, the airflow frequency will increase, thereby improving the cleaning effect of the airflow; when When it is smaller, the frequency will be reduced to reduce energy consumption.

Images