Intelligent glass curtain wall cleaning system and method with detection function

Abstract

The invention provides an intelligent glass curtain wall cleaning system and method with a detection function, and the system comprises a power device which is used for fixing a detection sensor and a cleaning device, and flies to a to-be-cleaned position of a glass curtain wall through a control system; the control system is used for controlling the power device to fly to a designated area so as to perform local cleaning; the cleaning device is used for being adsorbed to the surface of the glass curtain wall through a vacuum chuck and cleaning the glass curtain wall through a brush. The detection sensor is used for acquiring image information in a to-be-cleaned range of the glass curtain wall and sending the image information to the data processing system; and the data processing system is used for carrying out denoising, graying and edge detection processing on the image information sent by the detection sensor, carrying out smudginess degree analysis, carrying out cleaning area division and cleaning path planning according to an analysis result, and sending the cleaning area division and the cleaning path planning to the control system.

Description

Technical Field

[0001] The present application relates to the field of glass cleaning technology, and in particular to an intelligent glass curtain wall cleaning system and method with a detection function. Background Art

[0002] With the increasing number of high-rise buildings and the widespread use of glass curtain walls, the glass curtain wall cleaning industry has seen rapid growth. Traditionally, glass curtain wall cleaning relies on manual overhead operations, but this approach carries significant operational risks, high labor costs, and low cleaning efficiency. Consequently, an increasing number of companies are developing and implementing unmanned intelligent glass curtain wall cleaning systems and devices. However, existing unmanned glass curtain wall cleaning technologies lack the ability to detect and determine glass cleaning effectiveness, nor can they re-clean glass that has been poorly cleaned. Consequently, their level of intelligence is limited. Summary of the Invention

[0003] In view of this, the present invention provides an intelligent glass curtain wall cleaning system and method with detection function to solve the technical problems of the existing technology, such as high operation risk, high labor cost, low cleaning efficiency, inability to detect and determine the cleaning effect of the glass, and low degree of intelligence.

[0004] The present invention provides an intelligent glass curtain wall cleaning system with a detection function, the system comprising: a power unit, comprising a fuselage, a frame and blades, for fixing a detection sensor and a cleaning device, and flying to a position on the glass curtain wall to be cleaned through a control system; a control system connected to the power unit, for controlling the power unit to fly to a designated area for local cleaning according to the cleaning area division and cleaning path planning sent by a data processing system; a cleaning device, comprising a water storage bucket, a brush and a vacuum suction cup, for adsorbing the power unit on the surface of the glass curtain wall through the vacuum suction cup and cleaning the glass curtain wall with the brush when the power unit is suspended on the glass curtain wall; a detection sensor and a cleaning device for fixing a detection sensor and a cleaning device, and a control system for controlling the power unit to fly to a designated area for local cleaning according to the cleaning area division and cleaning path planning sent by a data processing system; a detection sensor and a cleaning device for fixing a detection sensor and a cleaning device ... The detection sensor includes a camera, which is arranged on the power device and is used to obtain image information within the range of the glass curtain wall to be cleaned through the camera, and send the image information to the data processing system; the data processing system is connected to the control system and the detection sensor through the network, and is used to perform denoising, grayscale, and edge detection processing on the image information sent by the detection sensor through a dirt detection algorithm, and perform dirt degree analysis. Based on the analysis results, a linear regression algorithm is used to construct a dirt accumulation prediction model, adjust the cleaning strategy, divide the cleaning area and plan the cleaning path, and send the cleaning area division and cleaning path planning to the control system.

[0005] Furthermore, the system also includes: a background display system connected to the data processing system, which is used to generate glass curtain wall information based on the analysis results of the data processing system and display it on the background display screen.

[0006] Furthermore, the cleaning device also includes a robotic arm, which is connected to the brush and is used for mobile cleaning on the surface of the glass curtain wall.

[0007] The present invention also provides a method for an intelligent glass curtain wall cleaning system with a detection function, the method comprising: step 1, a detection sensor acquires image information within a range of the glass curtain wall to be cleaned through a camera, and sends the image information to a data processing system; step 2, using a dirt detection algorithm, denoising, graying, and edge detection processing are performed on the image information sent by the detection sensor, and a dirt degree analysis is performed; then, based on the analysis results, a linear regression algorithm is used to construct a dirt accumulation prediction model, the cleaning strategy is adjusted, and cleaning area division and cleaning path planning are performed, and the cleaning area division and cleaning path planning are sent to a control system; step 3, the control system controls a power device to fly to a designated cleaning area based on the cleaning area division and cleaning path planning sent by the data processing system; step 4, the power device flies to the designated cleaning area and hovers there; step 5, when the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the surface of the glass curtain wall through a vacuum suction cup, and the efficiency of cleaning the glass curtain wall at that position is improved by adjusting the brush rotation speed according to the dirt degree of the glass surface.

[0008] Furthermore, the method further includes: step 0, the background display system sends the set glass curtain wall cleaning range coordinate value to the control system, so that it controls the power device to fly to the glass curtain wall to be cleaned range.

[0009] Furthermore, the method also includes: Step 6, the background display system generates glass curtain wall information according to the analysis result of the data processing system, and displays it on the background display screen.

[0010] Furthermore, the step 1 includes: step 11, flying to any position within the range to be cleaned by a power device according to the received coordinate value of the glass curtain wall cleaning range; step 12, determining the current coordinate point, setting the horizontal and vertical coordinates through the background display system to determine the cleaning range; step 13, moving according to the cleaning range, and collecting image information of the entire glass curtain wall to be cleaned through the camera.

[0011] Furthermore, the method for performing denoising, grayscale conversion, and edge detection on the image information sent by the detection sensor and performing dirtiness analysis includes: step 21, converting the color image into a grayscale image, using the weighted average method, the formula is as follows: Gray = 0.299R + 0.587G + 0.114B, where R, G, and B are the red, green, and blue channel values of the image, respectively, and Gray is the grayscale value; step 22, using the Gaussian kernel function Perform convolution on the image To reduce the influence of noise, where (x, y) is the pixel coordinate, σ is the standard deviation, I′(x, y) is the original image pixel value, and k is the value range determined by the Gaussian kernel size; Step 23, edge detection algorithm is used to extract edge features of the glass surface. Dirty areas usually cause edge features to change; Step 24, contrast is extracted by calculating the gray level co-occurrence matrix. The calculation formula is: When the contrast of a certain area is greater than the threshold, it is determined to be a dirty area. The judgment formula is: Among them, P is the gray level co-occurrence matrix, N is the number of gray levels, C T is the contrast threshold.

[0012] Furthermore, the method of constructing a dirt accumulation prediction model based on the analysis results using a linear regression algorithm and adjusting the cleaning strategy includes: Step 25, calculating the dirt coverage by comparing the dirt area data before and after cleaning, and the calculation formula is: ΔS = S before -S after , Among them, the total dirty area of the glass curtain wall before cleaning is S before , the total dirty area after cleaning is S after , the change in dirt coverage is ΔS, and the decrease rate of dirt coverage is C; Step 26, using the least squares method to fit the relationship between dirt area and time S(t) = αt + β, where α is the dirt accumulation rate and β is the initial dirt area; Step 27, calculate the recommended time for the next cleaning: Among them, α is the dirt accumulation rate, β is the initial dirt area, S trigger The trigger threshold for dirtiness.

[0013] Furthermore, step 5 includes: step 51, when the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the glass curtain wall through the vacuum suction cup; step 52, detecting the contact force between the suction cup and the surface of the glass curtain wall to ensure the cleaning effect and safety; step 53, after the adsorption is completed, the robotic arm drives the brush to clean the glass curtain wall according to the following degree of dirt; step 53, after the cleaning of the glass curtain wall in the current cleaning area is completed, the detection sensor collects the image information after cleaning through the camera and sends it to the background display system; step 54, the background display system judges the cleaning effect according to the image information. If the cleaning effect is not good, it is cleaned again. If the cleaning effect is good, it is moved to the next cleaning area within the cleaning range for cleaning until all the glass curtain walls within the cleaning range are cleaned.

[0014] This invention provides an intelligent glass curtain wall cleaning system and method with detection capabilities. This technical solution collects glass image information and performs image processing to detect and analyze the degree of glass contamination. It then divides cleaning areas and plans cleaning paths based on the degree of contamination. Finally, the designated areas are cleaned according to the cleaning path plan. This technical solution addresses the low efficiency and high labor costs of manual cleaning, as well as the inability to detect and determine glass cleaning results. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a structural diagram of an intelligent glass curtain wall cleaning system with detection function provided by the present invention;

[0016] Figure 2 This is a schematic diagram of a backstage display screen interface provided by the present invention;

[0017] Figure 3 This is a flow chart of an intelligent glass curtain wall cleaning method with detection function provided by the present invention. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0019] Device Item Embodiment

[0020] The present invention provides an intelligent glass curtain wall cleaning system with detection function, such as Figure 1 As shown, the cleaning system includes a power device, a control system, a cleaning device, a detection sensor, a data processing system and a background display system.

[0021] The power unit includes a fuselage 2, a frame 3 and blades 1, which are used to fix the detection sensor 7 and the cleaning device, and fly to the position of the glass curtain wall to be cleaned through the control system;

[0022] A control system is connected to the power unit and is used to control the power unit to fly to a designated area for local cleaning according to the cleaning area division and cleaning path planning sent by the data processing system;

[0023] The cleaning device includes a water storage bucket 4, a brush 6, and a vacuum suction cup 5. When the power device is suspended from the glass curtain wall, the vacuum suction cup 5 is attached to the surface of the glass curtain wall and the brush 6 is used to clean the glass curtain wall. The cleaning device also includes a robotic arm 8 connected to the brush 6 for moving and cleaning the surface of the glass curtain wall. The cleaning device will start cleaning the glass when it receives a start cleaning command or a re-cleaning command from the background display system. This shows that the entire cleaning work is completed automatically by the cleaning device without human intervention, reducing the danger of manual high-altitude work and taking into account various glass curtain wall types.

[0024] The detection sensor 7 includes a camera, which is installed on the power device and is used to obtain image information of the glass curtain wall to be cleaned through the camera and send the image information to the data processing system. The detection sensor 7 can scan the glass curtain wall area and the stain area of a single piece of glass through sensors such as the camera.

[0025] The data processing system, connected to the control system and detection sensor 7 via a network, performs denoising, grayscaling, and edge detection on the image information sent by the detection sensor 7, performing contamination analysis. Based on the analysis results, it then divides cleaning areas and plans cleaning routes, transmitting these areas and routes to the control system. Different data processing systems can detect glass contamination and dust based on the received image information, outputting cleaning results through methods such as before-and-after comparisons to ensure effective cleaning. The data processing system can also detect glass breakage, thereby assisting in routine glass maintenance. Utilizing cloud computing technology, the data processing system boasts powerful data storage and analysis capabilities. It receives and stores data from the cleaning and detection robots and the ground control base station, including glass curtain wall cleaning records, inspection results, and equipment operating status. Data analysis algorithms are used to assess the overall condition of the glass curtain wall, predicting its service life and developing appropriate cleaning and maintenance plans.

[0026] The background display system is connected to the data processing system and is used to generate glass curtain wall information based on the analysis results of the data processing system and display it on the background display screen. Figure 2 As shown, the background display system generates glass curtain wall information based on the information transmitted by the data processing system, and outputs a digital interface to be displayed on the background screen.

[0027] This application provides an intelligent glass curtain wall cleaning system and method with detection capabilities. This technical solution uses a detection sensor 7 to collect glass curtain wall image information. A data processing system then processes and analyzes the image, divides the cleaning area, plans the cleaning path, and transmits the processed data to a background display system. The cleaning system divides the cleaning area and plans the cleaning path to clean the glass curtain wall. The background display system automatically generates a cleaning screen based on the data processed by the data processing system and displays the dust concentration of each glass. This technical solution can effectively solve the problems of low manual cleaning efficiency and high labor costs.

[0028] Method Item Embodiment

[0029] The present invention provides an intelligent glass curtain wall cleaning method with detection function, such as Figure 3 As shown, the operating steps of the method are as follows.

[0030] Step 0: The background display system sends the set glass curtain wall cleaning range coordinate values to the control system so that it controls the power device to fly to the glass curtain wall to be cleaned range.

[0031] Step 1: The detection sensor obtains image information of the glass curtain wall within the range to be cleaned through the camera and sends the image information to the data processing system;

[0032] The steps for the detection sensor to obtain glass curtain wall image information are as follows.

[0033] Step 11: According to the received coordinate values of the glass curtain wall cleaning range, the robot flies to any position within the cleaning range through the power device;

[0034] Step 12: Determine the current coordinate point and set the horizontal and vertical coordinates through the background display system to determine the cleaning range;

[0035] Step 13: Move according to the cleaning range and collect image information of the entire glass curtain wall to be cleaned through the camera.

[0036] Step 2: Using a dirt detection algorithm, the image information sent by the detection sensor is subjected to denoising, grayscale conversion, and edge detection processing to analyze the degree of dirtiness. Based on the analysis results, a linear regression algorithm is used to build a dirt accumulation prediction model, adjust the cleaning strategy, divide the cleaning area, and plan the cleaning path. The cleaning area division and cleaning path planning are then sent to the control system.

[0037] The data processing system has data storage and analysis functions, records the dirt situation and cleaning effect during the cleaning process, analyzes the image through the image processing algorithm, and identifies the type, location and area of dirt on the glass surface. The data processing system performs dirt analysis and adjusts the cleaning strategy in the following steps.

[0038] Step 21: Convert the color image into a grayscale image using the weighted average method. The formula is as follows:

[0039] Gray=0.299R+0.587G+0.114B

[0040] , where R, G, and B are the red, green, and blue channel values of the image respectively, and Gray is the grayscale value;

[0041] Step 22: Use Gaussian kernel function Perform convolution on the image To reduce the influence of noise, (x, y) is the pixel coordinate, σ is the standard deviation, I′(x, y) is the original image pixel value, and k is the value range determined by the Gaussian kernel size;

[0042] Step 23: Using an edge detection algorithm to extract edge features of the glass surface. Dirty areas usually cause changes in edge features.

[0043] Use edge detection algorithms, such as the Canny edge detection algorithm, to extract edge features on the glass surface. Dirty areas usually change edge features. The Canny algorithm steps are: first, use a Gaussian filter to smooth the image; then calculate the image gradient amplitude M(x, y) and direction θ(x, y). The specific calculation formulas are: and Among them, G x (x,y) and G y (x, y) are the gradients in the x and y directions, respectively. Non-maximum suppression is applied, and the pixels with the largest local gradient are retained. Finally, the final edge image is determined through dual-threshold detection and edge connection. By comparing the edge image with a clean glass edge feature library, if the difference in edge features in a certain area exceeds a set threshold, it is determined that the area may be dirty.

[0044] Step 24, extract the contrast by calculating the gray level co-occurrence matrix, the calculation formula is When the contrast of a certain area is greater than the threshold, it is determined to be a dirty area. The judgment formula is: Among them, P is the gray level co-occurrence matrix, N is the number of gray levels, C T is the contrast threshold;

[0045] Contrast extracted based on the gray-level co-occurrence matrix can be used as a texture feature. Therefore, dirty fishing areas can be judged based on the contrast. That is, when the contrast of a certain area is greater than the threshold, it is determined to be a dirty area.

[0046] Step 25: Calculate the dirt coverage by comparing the dirt area data before and after cleaning. The calculation formula is: ΔS = S before -Safter , Among them, the total dirty area of the glass curtain wall before cleaning is S before , the total dirty area after cleaning is S after , the change in dirt coverage is ΔS, and the decrease ratio of dirt coverage is C;

[0047] Step 26: Using the least squares method to fit the relationship between the pollution area and time, S(t) = αt + β, where α is the pollution accumulation rate and β is the initial pollution area.

[0048] Step 27: Calculate the recommended time for the next cleaning: Among them, α is the dirt accumulation rate, β is the initial dirt area, S trigger The trigger threshold for dirtiness.

[0049] Step 3: The control system controls the power unit to fly to the designated cleaning area according to the cleaning area division and cleaning path planning sent by the data processing system;

[0050] Step 4: The power unit flies to the designated clean area and hovers there;

[0051] Step 5: When the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the surface of the glass curtain wall through the vacuum suction cup, and the efficiency of cleaning the glass curtain wall at that position is improved by adjusting the rotation speed of the brush according to the degree of dirt on the glass surface.

[0052] Based on the results of the dirt analysis, such as the degree of dirt on the glass surface, and taking into account the different degrees of dirt in different areas, local cleaning and intelligent path planning can be achieved. The cleaning device includes a water storage bucket, a brush, and a vacuum suction cup. The water storage bucket is used to store cleaning liquid. The cleaning device can spray the cleaning liquid onto the glass surface, and the brush is driven by a motor to brush the glass. The cleaning brush is made of a soft material with strong decontamination ability. It can automatically adjust the brushing force and speed according to the degree of dirt on the glass. The specific adjustment strategy is: when it is detected that the degree of dirt on the glass surface is high, the motor increases the speed and increases the amount of cleaning liquid sprayed; when the degree of dirt is low, the speed and the amount of cleaning liquid sprayed are reduced to save resources and improve cleaning efficiency. The cleaning device is adsorbed on the surface of the glass curtain wall by a vacuum suction cup, and the method of adjusting the rotation speed of the brush according to the degree of dirt on the glass surface includes the following steps.

[0053] Step 51, when the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the glass curtain wall through the vacuum suction cup;

[0054] Step 52: Detecting the contact force between the suction cup and the surface of the glass curtain wall to ensure cleaning effect and safety;

[0055] Step 53: After the adsorption is completed, the robotic arm moves up and down and left and right, driving the brush to clean the glass curtain wall;

[0056] Step 53: After the glass curtain wall in the current cleaning area is cleaned, the detection sensor collects the cleaned image information through the camera and sends it to the background display system;

[0057] Step 54, the background display system determines the cleaning effect based on the image information. If the cleaning effect is not good, it will be cleaned again. If the cleaning effect is good, it will move to the next cleaning area within the cleaning range for cleaning until all the glass curtain walls within the cleaning range are cleaned.

[0058] From the above steps, we can see that the cleaning device automatically controls the movement of the robotic arm and the operation of the brush based on the dirt detection results and the detection sensor feedback. Taking into account the different levels of dirt in different areas, it can achieve local cleaning and intelligent path planning. Its control logic can be expressed as:

[0059] If the dirtiness detection value D is greater than the set threshold value D th1 , then the motor speed R=R max , the amount of cleaning fluid sprayed Q = Q max ;

[0060] II If D is less than or equal to D th1 And greater than the set threshold D th2 , then R=R mid , Q=Q mid ;

[0061] III If D is less than or equal to D th2 , then R=R min , Q=Q min .

[0062] Step 6: The background display system generates glass curtain wall information based on the analysis results of the data processing system and displays it on the background display screen.

[0063] The background display system can provide real-time monitoring interface and reports, making it easy for operators to understand the cleaning progress and effect. Figure 3 As shown, the glass curtain wall information includes: overall glass layout, soiling levels in each cleaning area, and path planning. The backend display system provides real-time monitoring and reporting, allowing operators to easily understand cleaning progress and results. It also provides users with an intuitive interactive interface. Users can log in via computer or mobile phone to view the real-time status and historical data of the glass curtain wall, and remotely issue control commands, enabling intelligent management of the entire system.

[0064] In summary, the embodiments of the present invention provide an intelligent glass curtain wall cleaning system and method with detection capabilities. Compared to traditional manual cleaning, this unmanned cleaning system can achieve rapid cleaning of large-scale glass curtain walls, reducing cleaning costs and improving cleaning efficiency. Furthermore, as an unmanned cleaning system, this system can avoid the risks of cleaning workers working at heights and improve cleaning safety. Through built-in algorithms and sensors, the technical solution can achieve intelligent control of the cleaning process, improving cleaning quality. The cleaning system can use environmentally friendly cleaning fluids to reduce environmental pollution. The cleaning system can be customized based on factors such as the shape, size, and material of the glass curtain wall, demonstrating strong adaptability.

[0065] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent glass curtain wall cleaning system with detection function, characterized in that: The system comprises: The power unit, including a fuselage, a frame, and propeller blades, is used to fix the detection sensor and the cleaning device, and fly to the position of the glass curtain wall to be cleaned through the control system; A control system is connected to the power unit and is used to control the power unit to fly to a designated area for local cleaning according to the cleaning area division and cleaning path planning sent by the data processing system; A cleaning device, comprising a water storage bucket, a brush, and a vacuum suction cup, is used to, when the power device is suspended on the glass curtain wall, adhere to the surface of the glass curtain wall through the vacuum suction cup and clean the glass curtain wall with the brush; The detection sensor, including a camera, is arranged on the power device and is used to obtain image information within the range of the glass curtain wall to be cleaned through the camera and send the image information to the data processing system; The data processing system is connected to the control system and detection sensors through the network. It is used to denoise, grayscale, and detect edges of the image information sent by the detection sensor through the dirt detection algorithm, and analyze the degree of dirt. Based on the analysis results, a linear regression algorithm is used to build a dirt accumulation prediction model, adjust the cleaning strategy, divide the cleaning area and plan the cleaning path, and send the cleaning area division and cleaning path planning to the control system.

2. The intelligent glass curtain wall cleaning system with detection function according to claim 1, characterized in that: The system further comprises: The background display system is connected to the data processing system and is used to generate glass curtain wall information based on the analysis results of the data processing system and display it on the background display screen.

3. The intelligent glass curtain wall cleaning system with detection function according to claim 1, characterized in that: The cleaning device also includes a mechanical arm, which is connected to a brush and is used for mobile cleaning on the surface of the glass curtain wall.

4. A method using the intelligent glass curtain wall cleaning system with detection function according to claims 1-3, characterized in that: The method comprises: Step 1: The detection sensor obtains image information of the glass curtain wall within the range to be cleaned through the camera and sends the image information to the data processing system; Step 2: Using a dirt detection algorithm, the image information sent by the detection sensor is subjected to denoising, grayscale conversion, and edge detection processing to analyze the degree of dirtiness. Based on the analysis results, a linear regression algorithm is used to build a dirt accumulation prediction model, adjust the cleaning strategy, divide the cleaning area, and plan the cleaning path. The cleaning area division and cleaning path planning are then sent to the control system. Step 3: The control system controls the power unit to fly to the designated cleaning area according to the cleaning area division and cleaning path planning sent by the data processing system; Step 4: The power unit flies to the designated clean area and hovers there; Step 5: When the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the surface of the glass curtain wall through the vacuum suction cup, and the efficiency of cleaning the glass curtain wall at that position is improved by adjusting the rotation speed of the brush according to the degree of dirt on the glass surface.

5. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The method further comprises: Step 0: The background display system sends the set glass curtain wall cleaning range coordinate values to the control system so that it controls the power device to fly to the glass curtain wall to be cleaned range.

6. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The method further comprises: Step 6: The background display system generates glass curtain wall information based on the analysis results of the data processing system and displays it on the background display screen.

7. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The step 1 comprises: Step 11: According to the received coordinate values of the glass curtain wall cleaning range, the robot flies to any position within the cleaning range through the power device; Step 12: Determine the current coordinate point and set the horizontal and vertical coordinates through the background display system to determine the cleaning range; Step 13: Move according to the cleaning range and collect image information of the entire glass curtain wall to be cleaned through the camera.

8. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The method of performing denoising, grayscale conversion, and edge detection on the image information sent by the detection sensor and performing dirtiness analysis includes: Step 21: Convert the color image into a grayscale image using the weighted average method. The formula is as follows: Gray = 0.299R + 0.587G + 0.114B , where R, G, and B are the red, green, and blue channel values of the image respectively, and Gray is the grayscale value; Step 22: Use Gaussian kernel function Perform convolution on the image To reduce the influence of noise, (x, y) is the pixel coordinate, σ is the standard deviation, I′(x, y) is the original image pixel value, and k is the value range determined by the Gaussian kernel size; Step 23: Using an edge detection algorithm to extract edge features of the glass surface. Dirty areas usually cause changes in edge features. Step 24, extract the contrast by calculating the gray level co-occurrence matrix, the calculation formula is When the contrast of a certain area is greater than the threshold, it is determined to be a dirty area. The judgment formula is: Among them, P is the gray level co-occurrence matrix, N is the number of gray levels, C T is the contrast threshold.

9. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The method of using a linear regression algorithm to construct a dirt accumulation prediction model based on the analysis results and adjusting the cleaning strategy includes: Step 25: Calculate the dirt coverage by comparing the dirt area data before and after cleaning. The calculation formula is: ΔS = S before -S after , Among them, the total dirty area of the glass curtain wall before cleaning is S before , the total dirty area after cleaning is S after , the change in dirt coverage is ΔS, and the decrease ratio of dirt coverage is C; Step 26: Using the least squares method to fit the relationship between the pollution area and time, S(t) = αt + β, where α is the pollution accumulation rate and β is the initial pollution area. Step 27: Calculate the recommended time for the next cleaning: Among them, α is the dirt accumulation rate, β is the initial dirt area, S trigger The trigger threshold for dirtiness.

10. The intelligent glass curtain wall cleaning method with detection function according to claim 4, characterized in that: The step 5 comprises: Step 51, when the power device is suspended on the glass curtain wall, the cleaning device is adsorbed on the glass curtain wall through the vacuum suction cup; Step 52: Detecting the contact force between the suction cup and the surface of the glass curtain wall to ensure cleaning effect and safety; Step 53: After the adsorption is completed, the robotic arm drives the brush to clean the glass curtain wall according to the following dirtiness: Step 53: After the glass curtain wall in the current cleaning area is cleaned, the detection sensor collects the cleaned image information through the camera and sends it to the background display system; Step 54, the background display system determines the cleaning effect based on the image information. If the cleaning effect is not good, it will be cleaned again. If the cleaning effect is good, it will move to the next cleaning area within the cleaning range for cleaning until all the glass curtain walls within the cleaning range are cleaned.

Images