Glass curtain wall cleaning system and method based on cooperative targeted cleaning of mother bin and water gun

Abstract

The invention relates to glass curtain wall cleaning, in particular to a glass curtain wall cleaning system and method based on cooperative targeted cleaning of a mother bin and a water gun, the glass curtain wall cleaning system comprises the mother bin, a negative pressure adsorption type cleaning robot and a central control system, and the central control system is integrated in the mother bin. Semantic segmentation is carried out on the image sent by the enhanced global vision module based on a deep learning model, and a stain area, a stain type and a stubborn grade are identified; the targeted cleaning decision-making unit is used for determining targeted cleaning parameters according to the position, the stain type and the stubborn grade of each stain area, and planning a water gun spraying path and a robot cleaning path according to the positions of all the stain areas; the water gun motion control unit is used for calculating a spraying elevation compensation value corresponding to each stain area according to a water gun spraying path and a laser ranging result in combination with the influence of gravity on a trajectory, and calculating the angle of each joint of the mechanical arm; the defects that the cleaning efficiency is low and the cleaning effect is poor can be overcome.

Description

Technical Field

[0001] This invention relates to glass curtain wall cleaning, specifically to a glass curtain wall cleaning system and method based on a mother chamber and water gun for coordinated targeted cleaning. Background Technology

[0002] Stubborn stains on the glass curtain walls of high-rise buildings, such as bird droppings, paint spots, and cement stains, are difficult to clean effectively using traditional methods. Existing glass curtain wall cleaning solutions mainly suffer from the following problems: 1) Relying solely on robots for cleaning requires repeated wiping of stubborn stains that have dried, resulting in extremely low cleaning efficiency and potential damage to the glass; 2) Premixed spray cleaning solutions waste resources and cannot provide intensive treatment for heavily soiled areas; 3) Lack of intelligent identification and classification of stained areas, making it impossible to carry out targeted cleaning work for each stained area, resulting in poor cleaning effect.

[0003] Although cleaning robots with spraying functions have appeared on the market, their spraying behavior is mostly pre-set or simply triggered, failing to be combined with high-precision visual recognition and global planning, and thus unable to achieve an intelligent closed loop of "discovery-identification-decision-precision strike-coordinated execution-effect verification". Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a glass curtain wall cleaning system and method based on the coordinated targeted cleaning of the mother chamber and water gun, which can effectively overcome the defects of low cleaning efficiency and poor cleaning effect of the existing technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A glass curtain wall cleaning system based on coordinated targeted cleaning of a mother chamber and a water gun includes a mother chamber, a negative pressure adsorption cleaning robot, and a central control system. The mother chamber specifically includes: The enhanced global vision module scans the glass curtain wall to be cleaned, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The targeted cleaning module is used for targeted cleaning of stained areas; The central control system is integrated within the mother warehouse and specifically includes: The stain intelligent recognition and classification unit performs semantic segmentation on images sent by the enhanced global vision module based on a deep learning model, and identifies stain areas, stain types and stubbornness levels. The targeted cleaning decision unit determines the targeted cleaning parameters based on the location, type, and stubbornness of each stain area and sends them to the targeted cleaning module. At the same time, it plans the water gun spray path and robot cleaning path based on the location of all stain areas and sends them to the targeted cleaning module and the robot respectively. The water gun motion control unit calculates the spray elevation angle compensation value corresponding to each stain area based on the water gun spray path and laser ranging results, combined with the influence of gravity on the trajectory, and solves the angle of each joint of the robotic arm, and sends it down to the targeted cleaning module. The multi-unit collaborative scheduling unit coordinates the water guns and robots to complete the cleaning work.

[0006] Preferably, the enhanced global vision module includes: A multispectral camera, including a visible light camera and a near-infrared camera, scans the glass curtain wall to be cleaned, enhances the contrast between the stained area and the glass curtain wall, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The laser rangefinder accurately measures the distance between the mother chamber and the stained area and sends the data to the water gun motion control unit.

[0007] Preferably, the targeted cleaning module includes: The multi-degree-of-freedom robotic arm has at least two rotational degrees of freedom. It adjusts its posture according to the water jet path sent by the targeted cleaning decision unit and the angle of each joint of the robotic arm sent by the water jet motion control unit, so as to control the water flow to accurately target the stained area. The high-speed solenoid valve water gun is installed at the end of the robotic arm and controls the spray pressure, spray volume and duration according to the targeted cleaning parameters sent by the targeted cleaning decision unit. The liquid system, which includes a water tank, water pump, water pipes and flow sensor, controls the cleaning solution formula based on the targeted cleaning parameters sent by the targeted cleaning decision unit and delivers the cleaning solution to the water gun.

[0008] Preferably, the stain intelligent recognition and classification unit performs semantic segmentation on the images sent by the enhanced global vision module based on a deep learning model, identifying stain areas, stain types, and stubbornness levels, including: Based on the convolutional neural network (CNN), semantic segmentation is performed on the images sent by the enhanced global vision module to identify all stained areas. The stained areas are classified into stained types and categorized by stubbornness level. The outline and center point coordinates of each stained area are accurately marked in the image. The types of stains include dust, bird droppings, and oil stains, and the stubbornness levels are light, moderate, and heavy.

[0009] Preferably, the targeted cleaning decision unit determines targeted cleaning parameters based on the location, stain type, and stubbornness level of each stained area, and sends these parameters to the targeted cleaning module, including: Based on a preset stain database, appropriate targeted cleaning parameters are matched according to the location, stain type, and stubbornness level of each stain area, and then sent to the targeted cleaning module. The preset stain database includes targeted cleaning parameters corresponding to different stain types and stubbornness levels. These parameters include spray pressure, spray volume, duration, and cleaning solution formula.

[0010] Preferably, the targeted cleaning decision unit plans the water gun spray path and the robot cleaning path based on the location of all stained areas, and sends them to the targeted cleaning module and the robot respectively, including: To maximize efficiency, all stained areas are traversed, the motion sequence of the robotic arm is planned, the water gun spray path is obtained, and the data is sent to the targeted cleaning module. Based on the pre-planned theoretical path, the robot's movement sequence is optimized to obtain the robot's cleaning path, which is then sent to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency.

[0011] Preferably, during the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and the robot's movement to ensure that the robot stays away from the current stained area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stained area after the water gun spraying is completed.

[0012] Preferably, after cleaning all stained areas, the mother chamber performs a second scan of the stained areas using an enhanced global vision module to detect and judge the cleaning effect. If there are residual stains, the central control system decides to spray the stained area with water a second time or mark it as a key cleaning area for the robot, thus completing the cleaning task.

[0013] A glass curtain wall cleaning method based on the coordinated targeted cleaning of the mother chamber and water gun includes the following steps: S1. Intelligent Detection and Recognition: The mother chamber scans the glass curtain wall to be cleaned using an enhanced global vision module, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The stain intelligent recognition and classification unit performs semantic segmentation on the images sent by the enhanced global vision module based on the convolutional neural network (CNN), identifies all stain areas, classifies each stain area into stain types and grades the stubbornness level, and accurately marks the outline and center point coordinates of each stain area in the image. S2. Cleaning Decision and Path Planning: The targeted cleaning decision unit, based on a preset stain database, matches appropriate targeted cleaning parameters according to the location, stain type, and stubbornness level of each stain area, and then sends them to the targeted cleaning module. The targeted cleaning decision unit traverses all stained areas with maximum efficiency, plans the motion sequence of the robotic arm, obtains the water gun spray path, and sends it to the targeted cleaning module. Based on the pre-planned theoretical path, the targeted cleaning decision unit optimizes the robot's movement sequence to obtain the robot's cleaning path and sends it to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency. S3, Precision Strike and Coordinated Execution: The water gun motion control unit calculates the spray elevation angle compensation value corresponding to each stain area based on the water gun spray path and laser range measurement results, combined with the influence of gravity on the trajectory, and solves the angle of each joint of the robotic arm, and sends it to the targeted cleaning module. The mother chamber's targeted cleaning module adjusts its position based on the water gun's spray path and the angles of each joint of the robotic arm to control the water flow to precisely target the soiled area. It also controls the spray pressure, water volume, duration, and cleaning solution formula based on the targeted cleaning parameters. During the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and robot movement to ensure that the robot stays away from the current stain area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stain area after the water gun spraying is completed. S4. Effect Verification and Task Closure: After cleaning all stained areas, the mother compartment uses an enhanced global vision module to perform a second scan of the stained areas to detect and judge the cleaning effect. If there are residual stains, the central control system decides to spray the stained area with water a second time or mark it as a key cleaning area for the robot, thus completing the cleaning task.

[0014] Compared with the prior art, the glass curtain wall cleaning system and method based on the coordinated targeted cleaning of the mother chamber and water gun provided by the present invention has the following beneficial effects: 1) Revolutionary efficiency improvement: Water gun pretreatment for stubborn stains increases the robot's cleaning efficiency several times and significantly reduces the overall operation time; 2) Improved cleaning effect: It can perform targeted cleaning of various stained areas, and the cleaning effect is far superior to traditional single methods; 3) Intelligence and Automation: It achieves full-process automation of "discovery-identification-decision-precision strike-collaborative execution-effect verification", making it a truly intelligent cleaning system; 4) Extreme resource conservation: Changing from overall spraying to "point spraying" greatly saves water and cleaning solution, which is in line with the concept of green environmental protection; 5) High functional scalability: The mother warehouse platform can also be used for external wall fire early warning (identifying smoke and spraying), external wall maintenance (spraying protective agents), etc., with extremely high scalability. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0016] Figure 1 This is a schematic diagram of the system architecture of the present invention; Figure 2 This is a schematic diagram of the system modules of the present invention; Figure 3 This is a schematic diagram of the method flow of the present invention; Figure 4 This is a schematic diagram illustrating the principle of calculating the jet elevation angle compensation value corresponding to each stained area in this invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0018] The following describes the specific architecture and functional modules of the glass curtain wall cleaning system based on the coordinated targeted cleaning of the mother chamber and water gun provided by this invention, using specific examples (such as...). Figure 1 and Figure 2 As shown), it includes a mother chamber, a negative pressure suction cleaning robot, and a central control system. The mother chamber specifically includes: The enhanced global vision module scans the glass curtain wall to be cleaned, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The targeted cleaning module is used for targeted cleaning of stained areas; The central control system is integrated within the mother warehouse and specifically includes: The stain intelligent recognition and classification unit performs semantic segmentation on images sent by the enhanced global vision module based on a deep learning model, and identifies stain areas, stain types and stubbornness levels. The targeted cleaning decision unit determines the targeted cleaning parameters based on the location, type, and stubbornness of each stain area and sends them to the targeted cleaning module. At the same time, it plans the water gun spray path and robot cleaning path based on the location of all stain areas and sends them to the targeted cleaning module and the robot respectively. Water gun motion control unit, such as Figure 4 As shown, based on the water gun spray path and laser ranging results, combined with the influence of gravity on the trajectory, the spray elevation angle compensation value corresponding to each stain area is calculated, and the angle of each joint of the robotic arm is calculated and sent to the targeted cleaning module. The multi-unit collaborative scheduling unit coordinates the water guns and robots to complete the cleaning work.

[0019] I. Enhanced Global Vision Module A multispectral camera, including a visible light camera and a near-infrared camera, scans the glass curtain wall to be cleaned, enhances the contrast between the stained area and the glass curtain wall, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The laser rangefinder accurately measures the distance between the mother chamber and the stained area and sends the data to the water gun motion control unit.

[0020] II. Targeted Cleaning Module The multi-degree-of-freedom robotic arm has at least two rotational degrees of freedom (such as pitch and yaw). It adjusts its posture according to the water jet path sent by the targeted cleaning decision unit and the angle of each joint of the robotic arm sent by the water jet motion control unit, so as to control the water flow to accurately target the stained area. The high-speed solenoid valve water gun is installed at the end of the robotic arm and controls the spray pressure, spray volume and duration according to the targeted cleaning parameters sent by the targeted cleaning decision unit. The liquid system, which includes a water tank, water pump, water pipes and flow sensor, controls the cleaning solution formula based on the targeted cleaning parameters sent by the targeted cleaning decision unit and delivers the cleaning solution to the water gun.

[0021] III. Intelligent Stain Recognition and Classification Unit The stain intelligent recognition and classification unit performs semantic segmentation on images sent by the enhanced global vision module based on a deep learning model, identifying stain regions, stain types, and stubbornness levels, including: Based on the convolutional neural network (CNN), semantic segmentation is performed on the images sent by the enhanced global vision module to identify all stained areas. The stained areas are classified into stained types and categorized by stubbornness level. The outline and center point coordinates of each stained area are accurately marked in the image. The stain types include dust (Class 1), bird droppings (Class 2), and oil (Class 3), and the stubbornness levels include light (Level 1), moderate (Level 2), and heavy (Level 3).

[0022] IV. Targeted Cleaning Decision Unit The targeted cleaning decision unit determines the targeted cleaning parameters based on the location, stain type, and stubbornness level of each stained area, and then sends these parameters to the targeted cleaning module, including: Based on a preset stain database, appropriate targeted cleaning parameters are matched according to the location, stain type, and stubbornness level of each stain area (e.g., for a stain area with stain type Class 2 - bird droppings, use medium spray pressure, mix cleaning solution, and spray for 3 seconds), and then sent to the targeted cleaning module. The preset stain database includes targeted cleaning parameters corresponding to different stain types and stubbornness levels. These parameters include spray pressure, spray volume, duration, and cleaning solution formula.

[0023] The targeted cleaning decision unit plans the water gun spray path and robot cleaning path based on the location of all stained areas, and then distributes these plans to the targeted cleaning module and the robot, including: To maximize efficiency, all stained areas are traversed, the motion sequence of the robotic arm is planned, the water gun spray path is obtained, and the data is sent to the targeted cleaning module. Based on the pre-planned theoretical path, the robot's movement sequence is optimized to obtain the robot's cleaning path, which is then sent to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency.

[0024] V. Multi-body Cooperative Scheduling Unit During the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and robot movement to ensure that the robot stays away from the current stained area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stained area after the water gun spraying is completed.

[0025] In the technical solution of this application, after the cleaning work of all stained areas is completed, the mother chamber performs a second scan of the stained areas through an enhanced global vision module to detect and judge the cleaning effect (such as judging by the degree of reduction in stained area). If there are residual stains, the central control system decides to spray the stained area with water guns for a second time or mark it as a key cleaning area for the robot, and finally completes the cleaning task.

[0026] Based on the aforementioned glass curtain wall cleaning system using a mother chamber and water jet for coordinated targeted cleaning, this application further discloses a glass curtain wall cleaning method based on the same method. Figure 3 As shown, it includes the following steps: S1. Intelligent Detection and Recognition: The mother chamber scans the glass curtain wall to be cleaned using an enhanced global vision module, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The stain intelligent recognition and classification unit performs semantic segmentation on the images sent by the enhanced global vision module based on the convolutional neural network (CNN), identifies all stain areas, classifies each stain area into stain types and grades the stubbornness level, and accurately marks the outline and center point coordinates of each stain area in the image. S2. Cleaning Decision and Path Planning: The targeted cleaning decision unit, based on a preset stain database, matches appropriate targeted cleaning parameters according to the location, stain type, and stubbornness level of each stain area, and then sends them to the targeted cleaning module. The targeted cleaning decision unit traverses all stained areas with maximum efficiency, plans the motion sequence of the robotic arm, obtains the water gun spray path, and sends it to the targeted cleaning module. Based on the pre-planned theoretical path, the targeted cleaning decision unit optimizes the robot's movement sequence to obtain the robot's cleaning path and sends it to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency. S3, Precision Strike and Coordinated Execution: The water gun motion control unit calculates the spray elevation angle compensation value (e.g., based on the water gun spray path and laser rangefinder results, combined with the effect of gravity on the trajectory) for each contaminated area. Figure 4 (as shown), and calculates the angles of each joint of the robotic arm, and sends the results to the targeted cleaning module; The mother chamber's targeted cleaning module adjusts its position based on the water gun's spray path and the angles of each joint of the robotic arm to control the water flow to precisely target the soiled area. It also controls the spray pressure, water volume, duration, and cleaning solution formula based on the targeted cleaning parameters. During the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and robot movement to ensure that the robot stays away from the current stain area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stain area after the water gun spraying is completed. S4. Effect Verification and Task Closure: After cleaning all stained areas, the mother compartment uses an enhanced global vision module to perform a second scan of the stained areas to detect and judge the cleaning effect. If there are residual stains, the central control system decides to spray the stained area with water a second time or mark it as a key cleaning area for the robot, thus completing the cleaning task.

[0027] To better illustrate the technical solution of this application, a specific example will be used for detailed explanation below.

[0028] The mother chamber is equipped with a two-degree-of-freedom (pitch and yaw) gimbal, on which a high-speed solenoid valve-controlled water gun with a nozzle diameter of 0.5mm is mounted, with a maximum range of 20m. The water tank can store different cleaning solutions. The central control system incorporates a lightweight convolutional neural network (CNN) based on TensorFlow Lite to identify three common types of stains (dust, bird droppings, and oil stains).

[0029] Work process: 1) The mother warehouse scanned the glass curtain wall to be cleaned and found a Level 3 (severely stubborn) Class 2 (bird droppings) stain area with a diameter of about 10cm, located 2m away; 2) The targeted cleaning decision unit determined to use "alkaline cleaning solution", with a spray pressure of level 2 and a spray duration of 2.5 seconds; 3) Control the robotic arm to rotate to the corresponding angle, trigger the valve to open, and the water gun will spray for 2.5 seconds; 4) Wait 60 seconds (softening time), then instruct the robot to move to the current stained area and perform regular wiping to easily remove the softened bird droppings; 5) Compare the images of the mother chamber before and after cleaning to confirm that the stained area is clean.

[0030] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A glass curtain wall cleaning system based on coordinated targeted cleaning of a mother chamber and a water gun, characterized in that: It includes a mother chamber, a negative pressure suction cleaning robot, and a central control system. The mother chamber specifically includes: The enhanced global vision module scans the glass curtain wall to be cleaned, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The targeted cleaning module is used for targeted cleaning of stained areas; The central control system is integrated within the mother warehouse and specifically includes: The stain intelligent recognition and classification unit performs semantic segmentation on images sent by the enhanced global vision module based on a deep learning model, and identifies stain areas, stain types and stubbornness levels. The targeted cleaning decision unit determines the targeted cleaning parameters based on the location, type, and stubbornness of each stain area and sends them to the targeted cleaning module. At the same time, it plans the water gun spray path and robot cleaning path based on the location of all stain areas and sends them to the targeted cleaning module and the robot respectively. The water gun motion control unit calculates the spray elevation angle compensation value corresponding to each stain area based on the water gun spray path and laser ranging results, combined with the influence of gravity on the trajectory, and solves the angle of each joint of the robotic arm, and sends it down to the targeted cleaning module. The multi-unit collaborative scheduling unit coordinates the water guns and robots to complete the cleaning work.

2. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 1, characterized in that: The enhanced global vision module includes: A multispectral camera, including a visible light camera and a near-infrared camera, scans the glass curtain wall to be cleaned, enhances the contrast between the stained area and the glass curtain wall, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The laser rangefinder accurately measures the distance between the mother chamber and the stained area and sends the data to the water gun motion control unit.

3. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 2, characterized in that: The targeted cleaning module includes: The multi-degree-of-freedom robotic arm has at least two rotational degrees of freedom. It adjusts its posture according to the water jet path sent by the targeted cleaning decision unit and the angle of each joint of the robotic arm sent by the water jet motion control unit, so as to control the water flow to accurately target the stained area. The high-speed solenoid valve water gun is installed at the end of the robotic arm and controls the spray pressure, spray volume and duration according to the targeted cleaning parameters sent by the targeted cleaning decision unit. The liquid system, which includes a water tank, water pump, water pipes and flow sensor, controls the cleaning solution formula based on the targeted cleaning parameters sent by the targeted cleaning decision unit and delivers the cleaning solution to the water gun.

4. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 3, characterized in that: The stain intelligent recognition and classification unit performs semantic segmentation on images sent by the enhanced global vision module based on a deep learning model, identifying stain regions, stain types, and stubbornness levels, including: Based on the convolutional neural network (CNN), semantic segmentation is performed on the images sent by the enhanced global vision module to identify all stained areas. The stained areas are classified into stained types and categorized by stubbornness level. The outline and center point coordinates of each stained area are accurately marked in the image. The types of stains include dust, bird droppings, and oil stains, and the stubbornness levels are light, moderate, and heavy.

5. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 4, characterized in that: The targeted cleaning decision unit determines the targeted cleaning parameters based on the location, stain type, and stubbornness level of each stained area, and sends them to the targeted cleaning module, including: Based on a preset stain database, appropriate targeted cleaning parameters are matched according to the location, stain type, and stubbornness level of each stain area, and then sent to the targeted cleaning module. The preset stain database includes targeted cleaning parameters corresponding to different stain types and stubbornness levels. These parameters include spray pressure, spray volume, duration, and cleaning solution formula.

6. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 5, characterized in that: The targeted cleaning decision unit plans the water gun spray path and robot cleaning path based on the location of all stained areas, and distributes these plans to the targeted cleaning module and the robot, respectively, including: To maximize efficiency, all stained areas are traversed, the motion sequence of the robotic arm is planned, the water gun spray path is obtained, and the data is sent to the targeted cleaning module. Based on the pre-planned theoretical path, the robot's movement sequence is optimized to obtain the robot's cleaning path, which is then sent to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency.

7. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 6, characterized in that: During the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and robot movement to ensure that the robot stays away from the current stained area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stained area after the water gun spraying is completed.

8. The glass curtain wall cleaning system based on coordinated targeted cleaning of the mother chamber and water gun according to claim 7, characterized in that: After cleaning all stained areas, the mother chamber performs a second scan of the stained areas using an enhanced global vision module to detect and judge the cleaning effect. If there are residual stains, the central control system decides to spray the stained area with water a second time or mark it as a key cleaning area for the robot, thus completing the cleaning task.

9. A glass curtain wall cleaning method based on coordinated targeted cleaning of a mother chamber and a water gun, applied to the glass curtain wall cleaning system based on coordinated targeted cleaning of a mother chamber and a water gun as described in claim 1, characterized in that: Includes the following steps: S1. Intelligent Detection and Recognition: The mother chamber scans the glass curtain wall to be cleaned using an enhanced global vision module, acquires high-resolution images, and sends them to the stain intelligent recognition and classification unit. The stain intelligent recognition and classification unit performs semantic segmentation on the images sent by the enhanced global vision module based on the convolutional neural network (CNN), identifies all stain areas, classifies each stain area into stain types and grades the stubbornness level, and accurately marks the outline and center point coordinates of each stain area in the image. S2. Cleaning Decision and Path Planning: The targeted cleaning decision unit, based on a preset stain database, matches appropriate targeted cleaning parameters according to the location, stain type, and stubbornness level of each stain area, and then sends them to the targeted cleaning module. The targeted cleaning decision unit traverses all stained areas with maximum efficiency, plans the motion sequence of the robotic arm, obtains the water gun spray path, and sends it to the targeted cleaning module. Based on the pre-planned theoretical path, the targeted cleaning decision unit optimizes the robot's movement sequence to obtain the robot's cleaning path and sends it to the robot. This ensures that the robot moves to the stained area to perform cleaning work only after the water gun has finished spraying, thus guaranteeing the softening time of the chemical reaction and improving cleaning efficiency. S3. Precision Strike and Coordinated Execution: The water gun motion control unit calculates the spray elevation angle compensation value corresponding to each stain area based on the water gun spray path and laser range measurement results, combined with the influence of gravity on the trajectory, and solves the angle of each joint of the robotic arm, and sends it to the targeted cleaning module. The mother chamber's targeted cleaning module adjusts its position based on the water gun's spray path and the angles of each joint of the robotic arm to control the water flow to precisely target the soiled area. It also controls the spray pressure, water volume, duration, and cleaning solution formula based on the targeted cleaning parameters. During the water gun spraying process, the multi-body collaborative scheduling unit coordinates the timing of the water gun and robot movement to ensure that the robot stays away from the current stain area, prevents the robot from being accidentally sprayed by the water gun, and notifies the robot to move to the current stain area after the water gun spraying is completed. S4. Results Verification and Task Closure: After cleaning all stained areas, the mother compartment uses an enhanced global vision module to perform a second scan of the stained areas to detect and judge the cleaning effect. If there are residual stains, the central control system decides to spray the stained area with water a second time or mark it as a key cleaning area for the robot, thus completing the cleaning task.

Images