Glass cleaning method and related device

Abstract

The present application provides a glass cleaning method and related devices. The method: for a glass sample to be detected, collect glass surface information corresponding to the glass sample; the glass surface information at least includes glass surface images under multiple light source systems; input the glass surface information into an abnormality detection model to perform abnormality detection on the glass sample; if the glass sample has abnormal scratches, and the abnormal scratches exceed the set damage threshold, stop the cleaning process of the glass sample; if the glass sample does not have abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, generate a cleaning strategy corresponding to the glass sample through a cleaning planning model; based on the cleaning strategy, control the wiping head and detergent spraying system of the wiping robot to execute the cleaning process for the glass sample. The embodiment of the present application can automatically control the wiping robot to perform cleaning, operate according to precise strategies, reduce interference from human factors, and ensure cleaning consistency, accuracy, stability and reliability.

Description

Technical Field

[0001] The present application relates to the technical field of curtain wall equipment, and in particular to a glass cleaning method and related devices. Background Art

[0002] At present, in the glass screen printing industry, before printing production, operators need to first check the condition of the glass surface to see if there are obvious flaws, the degree of dust accumulation, and whether there are any residual chemicals, and clean it.

[0003] However, in the related art, factors such as the technical level and work attitude of different operators will lead to uneven glass wiping quality. For example, some operators may miss some areas during the wiping process due to fatigue, or fail to completely remove the detergent on the surface of the glass, resulting in water stains after the glass dries. As long as the program of the automated wiping robot system is correct and the equipment is operating normally, the wiping quality of each piece of glass can be guaranteed to be relatively stable. In addition, long-term glass wiping work will place a great burden on the operator's body. For example, repeated wiping actions will cause fatigue and soreness in the arms, shoulders and other parts. Moreover, when cleaning large glass or a large amount of glass, operators need to frequently bend over, raise their hands and other actions, which can easily cause physical strain. In addition, the chemicals in the detergent may irritate the operator's skin and respiratory tract, and long-term contact may be harmful to health.

[0004] Therefore, it is urgent to design a new technical solution to solve at least one of the above technical problems. Summary of the invention

[0005] In order to solve or improve the above problems, the embodiments of the present application provide a glass cleaning method and related devices.

[0006] In one embodiment of the present application, a glass cleaning method is provided, which is applied to a wiping robot, the wiping robot at least comprising: a robot body, a mechanical arm, a visual detection system, a wiping head, a detergent spraying system, a base, and a main control computer, and the method is executed based on the main control computer; the method comprises:

[0007] For a glass sample to be tested, collecting glass surface information corresponding to the glass sample; the glass surface information at least includes glass surface images under multiple light source systems; the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively;

[0008] Inputting the glass surface information into an anomaly detection model to perform anomaly detection on the glass sample;

[0009] If the glass sample has abnormal scratches and the abnormal scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped;

[0010] If the glass sample has no abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, a cleaning strategy corresponding to the glass sample is generated through a cleaning planning model; wherein the cleaning planning model includes a plurality of planning prediction branches, each planning prediction branch is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample; the cleaning strategy includes at least: the motion trajectory, motion speed, angle, detergent spraying amount, and spraying angle of the wiping head;

[0011] The wiping head and the cleaning agent spraying system of the wiping robot are controlled based on the cleaning strategy to execute a cleaning process for the glass sample.

[0012] In another embodiment of the present application, a glass cleaning device is provided, which is applied to a wiping robot, which at least includes: a robot body, a mechanical arm, a visual detection system, a wiping head, a detergent spraying system, a base, and a main control computer. The glass cleaning device at least includes the following units:

[0013] A collection unit, for collecting glass surface information corresponding to a glass sample to be tested; the glass surface information at least includes glass surface images under multiple light source systems; the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively;

[0014] A detection unit, used for inputting the glass surface information into an abnormality detection model to perform abnormality detection on the glass sample;

[0015] A stopping unit, used for stopping the cleaning process of the glass sample if the glass sample has abnormal scratches and the abnormal scratches exceed a set damage threshold;

[0016] A generation unit, for generating a cleaning strategy corresponding to the glass sample through a cleaning planning model if the glass sample has no abnormal scratches or the abnormal scratches do not exceed a set damage threshold; wherein the cleaning planning model includes a plurality of planning prediction branches, each planning prediction branch is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample; the cleaning strategy includes at least: a motion trajectory, a motion speed, an angle, a spray amount of a cleaning agent, and a spray angle of a wiping head;

[0017] An execution unit is used to control a wiping head and a detergent spraying system of a wiping robot based on the cleaning strategy to execute a cleaning process for the glass sample.

[0018] In one embodiment of the present application, an electronic device includes:

[0019] at least one processor, memory, and input-output unit;

[0020] The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the glass cleaning method in the first embodiment.

[0021] In one embodiment of the present application, a computer-readable storage medium is provided, which includes instructions. When the instructions are executed on a computer, the computer executes the glass cleaning method in the first embodiment described above.

[0022] In the technical solution provided by the embodiment of the present application, for the glass sample to be detected, the glass surface information corresponding to the glass sample is collected; the glass surface information includes at least the glass surface images under multiple light source systems; the glass surface information is input into the abnormality detection model to perform abnormality detection on the glass sample; if the glass sample has abnormal scratches, and the abnormal scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped; if the glass sample does not have abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, the cleaning strategy corresponding to the glass sample is generated through the cleaning planning model; based on the cleaning strategy, the wiping head and the detergent spraying system of the wiping robot are controlled to execute the cleaning process for the glass sample. The embodiment of the present application can automatically control the wiping robot to perform cleaning, operate according to precise strategies, reduce interference from human factors, and ensure cleaning consistency, accuracy, stability and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 A schematic diagram of the structure of a glass cleaning device provided in one embodiment of the present application;

[0025] Figure 2 A schematic diagram of a process of a glass cleaning method provided in one embodiment of the present application;

[0026] Figure 3 A schematic diagram of the structure of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0027] The present application provides the following embodiments to solve or improve at least some of the above problems. In order to enable people in the technical field to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0028] In some features described in the specification, claims and the above-mentioned drawings of this application, descriptions such as "first" and "second" are used to distinguish different components, parts, modules, devices, etc., and do not represent the order of precedence, nor do they limit "first" and "second" to be different types. In addition, the embodiments described below are only some embodiments of the present application, not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0029] In order to solve the above technical problems, an embodiment of the present application provides a glass cleaning device, which includes the following units:

[0030] For a glass sample to be tested, collecting glass surface information corresponding to the glass sample; the glass surface information at least includes glass surface images under multiple light source systems; the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively;

[0031] Inputting the glass surface information into an anomaly detection model to perform anomaly detection on the glass sample;

[0032] If the glass sample has abnormal scratches, and the abnormal scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped;

[0033] If the glass sample has no abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, a cleaning strategy corresponding to the glass sample is generated through a cleaning planning model; wherein the cleaning planning model includes a plurality of planning prediction branches, each planning prediction branch is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample; the cleaning strategy includes at least: the motion trajectory, motion speed, angle, detergent spraying amount, and spraying angle of the wiping head;

[0034] The wiping head and the cleaning agent spraying system of the wiping robot are controlled based on the cleaning strategy to execute a cleaning process for the glass sample.

[0035] In the embodiment of the present application, the weight sensor, ink sensing module and ink supply module work together to achieve precise control and real-time adjustment of the ink volume. This solution not only improves the printing quality and production efficiency of colored glaze glass, but also reduces material and labor costs by reducing ink waste, optimizing screen design and improving operational convenience. At the same time, the environmental adaptability and environmental performance of the equipment have also been significantly improved, providing strong support for the process management level and sustainable development of the screen printing industry.

[0036] Figure 1 FIG. 1 is a schematic diagram showing the structure of a glass cleaning device provided by an embodiment of the present application. The glass cleaning device is mounted on a wiping robot. Figure 1 As shown, an embodiment of the present application proposes a glass cleaning device, which aims to automatically control a wiping robot to perform cleaning, operate according to a precise strategy, reduce interference from human factors, and ensure cleaning consistency, accuracy, stability and reliability. Figure 1 In the embodiment, the wiping robot at least includes: a robot body 1, a robotic arm 2, a wiping head 3, a visual inspection system 4, a detergent spraying system 5, a base 6, and a main control computer 7.

[0037] The embodiment of the present application is a glass cleaning device, the robot body is made of high-strength aluminum alloy, which not only ensures good rigidity and makes its structure stable during operation, but also achieves lightweight and easy to move and operate. The robot arm is a multi-joint structure, and multiple servo motors drive each joint. It can achieve high-precision spatial motion positioning according to the control system instructions and accurately reach the specified position on the glass surface. The joints are equipped with high-precision harmonic reducers, which effectively reduce the errors during the movement process, greatly improve the stability of the robot arm movement, and ensure the accuracy and stability of the wiping action.

[0038] The wiping head is installed at the end of the robot arm. It is made of replaceable flexible material. This material can automatically adjust the contact area according to the shape of the glass surface. Whether it is flat or curved glass, it can fit tightly to ensure the wiping effect. The pressure sensor inside the wiping head monitors the wiping pressure in real time and feeds the data back to the control system. The control system adjusts the movement or wiping parameters of the robot arm according to the pressure data to prevent excessive pressure from damaging the glass or insufficient pressure from causing poor cleaning effect.

[0039] The detergent spraying system is integrated with the wiping head. The liquid storage tank is made of corrosion-resistant materials, and its capacity can be flexibly adjusted according to the operation requirements to store the right amount of detergent. Under the precise instructions of the control system, the micro pump accurately controls the spray flow of the detergent to ensure that the amount of detergent used is just right. The atomizing nozzle sprays the detergent evenly on the glass surface, so that the detergent can better cover the glass and improve the cleaning effect.

[0040] Industrial cameras use high-resolution, high-frame-rate charge-coupled device (CCD) cameras, whose lenses have adjustable focal length and aperture functions. In different working scenarios, when the distance between the glass and the camera changes or the lighting conditions change, the camera can clearly capture the image of the glass surface by adjusting the lens focal length and aperture. The captured image data is quickly transmitted to the image processing unit via a data cable, providing an accurate data basis for the subsequent analysis of the glass surface condition. The light source system consists of multiple LED light sources, whose brightness and angle are adjustable, which can create a uniform and stable lighting environment for the glass surface, reduce problems such as image blur or loss of details caused by uneven or too strong or too weak lighting, thereby ensuring that the image obtained by the camera is clear and complete, so as to accurately judge the stains, scratches, etc. on the glass surface.

[0041] The laser distance sensor uses a high-precision pulsed laser rangefinder, and the specific wavelength laser it emits can effectively reduce ambient light interference. The sensor is installed on an adjustable bracket at a specific position of the robotic arm. For glass of different sizes and shapes, the angle and position of the bracket are adjusted to enable the sensor to fully scan the glass surface. The sensor emits laser and receives reflected light, measures the distance data, and then transmits the data to the data acquisition module, which further transmits the data to the data processing unit. The data processing unit uses the distance data to build a three-dimensional point cloud model of the glass surface. The model can be used to analyze the shape, size and other information of the glass, providing an important basis for the wiping robot to formulate accurate cleaning strategies, such as determining the motion trajectory, angle and other parameters of the wiping head to meet the cleaning needs of different glasses.

[0042] In the embodiment of the present application, the light source system should be set to a certain tilt angle with the glass surface, usually between 30°-60°. 45° can be selected as the initial angle test. This angle can cause the light to produce a certain reflection and refraction on the glass surface, so as to better highlight the texture and defects of the glass surface, including minor scratches, etc., which is convenient for subsequent image analysis. In the case of dim indoor light, the light source intensity can be increased to ensure that the glass surface has sufficient illumination, but excessive light should be avoided to cause excessive reflection on the glass surface, affecting the image clarity. For rectangular glass, at least two vertical directions (such as the long side and the short side direction) are used for shooting; for circular glass, the camera can be shot at equal angle intervals (such as every 45°) around the center of the circle; for special-shaped glass, according to its shape characteristics, select a shooting angle that can cover the glass surface to the greatest extent and highlight the key parts (areas with large curvature changes). In addition, in a strong light environment, add low-angle lighting to reduce the impact of reflected light on camera imaging; in a weak light environment, add auxiliary light sources to fill in light from the side or back to improve the overall brightness uniformity of the glass surface and reduce shadows. The LED light source has the advantages of high luminous efficiency, long life, and strong adjustability. The exposure time is dynamically adjusted according to the intensity of the light source and the reflection of the glass surface. The aperture size of the camera is adjusted according to the shooting distance and the required depth of field.

[0043] It should be added here that the control logic of the control device will be introduced in detail in the method embodiment below.

[0044] An embodiment of the present application provides a glass cleaning method. The method described in this embodiment is implemented on the basis of the structure of the above-mentioned glass cleaning device embodiment. Figure 2 As shown, the method may include the following steps:

[0045] 101. For a glass sample to be tested, collect glass surface information corresponding to the glass sample;

[0046] 102. Inputting the glass surface information into an anomaly detection model to perform anomaly detection on the glass sample;

[0047] 103. If the glass sample has abnormal scratches, and the abnormal scratches exceed a set damage threshold, the cleaning process of the glass sample is stopped;

[0048] 104. If the glass sample has no abnormal scratches, or the abnormal scratches do not exceed a set damage threshold, a cleaning strategy corresponding to the glass sample is generated by a cleaning planning model;

[0049] 105. If the glass sample has no abnormal scratches, or the abnormal scratches do not exceed a set damage threshold, a cleaning strategy corresponding to the glass sample is generated by a cleaning planning model.

[0050] Through the above steps, by collecting the surface images of the glass sample under multiple light source systems, these light source systems have different illumination angles and light intensities, and can comprehensively capture the condition of the glass surface from multiple dimensions, so that the obtained glass surface information is richer and more detailed. By inputting such comprehensive information into the anomaly detection model, it is possible to more accurately detect whether the glass sample has abnormal scratches, effectively avoid missing some subtle damages that are difficult to find, and provide a reliable basis for whether to perform the subsequent cleaning process. When abnormal scratches are detected and the scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped. This measure can prevent the cleaning operation of glass that has been severely damaged, because wiping, spraying and other actions during the cleaning process may further aggravate the degree of damage to the glass, thereby protecting the glass sample and avoiding its damage caused by unreasonable cleaning. Using a cleaning planning model containing multiple planning prediction branches, a cleaning strategy can be adaptively constructed based on the shape of the glass sample. Glass of different shapes often requires different cleaning methods to achieve better cleaning effects. For example, the motion trajectory, motion speed, angle of the wiping head, and the amount and angle of the detergent spray need to be adjusted according to the actual shape of the glass. This adaptively generated cleaning strategy can ensure that the cleaning process is more in line with the characteristics of the glass sample, improve cleaning efficiency and quality, and achieve more targeted and optimized cleaning operations. Based on the generated cleaning strategy, the wiping robot's wiping head and detergent spraying system are controlled to execute the cleaning process, realizing the automation of the entire cleaning process. Compared with manual cleaning, automation can work according to precisely set parameters to ensure the consistency and accuracy of each cleaning action, reduce the occurrence of uneven cleaning, inadequate cleaning, or accidental damage to the glass caused by human factors, and improve the overall stability and reliability of the cleaning work.

[0051] In the embodiment of the present application, the glass surface information at least includes glass surface images under multiple light source systems. Further, the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively.

[0052] When inspecting the condition of the glass surface, relying solely on images from a single light source is often not comprehensive enough. By setting up multiple light source systems to illuminate the glass from different specific angles, and each light source also has a different light intensity, the glass surface can present a variety of different light and shadow effects.

[0053] For example, when a light source of a certain angle and intensity is irradiated, some subtle scratches on the glass surface that are difficult to detect under normal viewing angles and lighting may be highlighted; changing the irradiation angle and intensity can make stains or other abnormalities hidden in specific textures, reflective areas, etc. clearly appear. In this way, multiple sets of glass surface images obtained by light sources of different angles and intensities together constitute richer and more three-dimensional glass surface information, which can capture the characteristics of the glass surface in all aspects to the greatest extent, and provide sufficient and reliable data basis for subsequent accurate analysis of the glass surface condition, such as judging the presence of scratches, the type and distribution of stains, etc.

[0054] When these glass surface image information under multiple light source systems are input into the anomaly detection model, the rich image data helps the model to identify abnormalities on the glass surface from multiple angles. Because the performance characteristics of abnormal parts (such as scratches, pits, spots, etc.) vary under different lighting angles and intensities, the model can more accurately determine whether the glass sample has abnormal scratches and other problems based on these differentiated performances, and then decide whether to stop the cleaning process (when there is an abnormality and exceeds the set damage threshold) or continue to generate the corresponding cleaning strategy, which greatly improves the accuracy and scientificity of the entire detection and decision-making process.

[0055] For example, in some high-precision glass production workshops, such as back glass, even the smallest flaws on the glass surface will affect the final quality and performance of the product. By using this multiple light source system with multiple specific illumination angles and multiple light intensities to collect images of the glass surface, it is possible to promptly detect those subtle defects that may be difficult to detect with the naked eye during the production process, and take corresponding measures in a timely manner to prevent problematic products from flowing into the next production link, thereby ensuring the overall yield and quality level of the product.

[0056] In the embodiment of the present application, the cleaning planning model includes multiple planning prediction branches, each of which is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample. Further, the cleaning strategy includes at least: the motion trajectory, motion speed, angle, detergent spraying amount, and spraying angle of the wiping head.

[0057] The cleaning planning model sets up multiple planning prediction branches, the purpose of which is to be able to flexibly respond to the cleaning needs of glass samples of different shapes. Different shapes of glass, such as square, round, irregular, etc., require different wiping head movement methods and detergent spraying methods when cleaning. Each planning prediction branch is a cleaning plan generator specifically for a specific type or shape of glass. It can call the corresponding algorithms and rules based on the input glass sample shape information, and then adaptively construct the cleaning strategy that best suits the shape of glass.

[0058] It is understandable that the diversity of glass shapes determines that a unified cleaning mode cannot achieve the ideal cleaning effect. For example, for round glass, the wiping head may need to make circular motions along the circumference, and the parameters such as the movement speed and angle should be reasonably set according to factors such as its radius. The spraying angle of the detergent should also match the arc of the circle to ensure comprehensive and uniform cleaning; and for glass with complex and irregular shapes, the movement trajectory of the wiping head should be flexibly adjusted according to its contour characteristics to avoid missing cleaning areas or damaging the glass due to unreasonable movement. By constructing a cleaning strategy based on shape adaptation, it can be ensured that when facing glass samples of various shapes, cleaning work can be carried out in a targeted manner, maximizing cleaning efficiency and cleaning quality, and ensuring that the cleanliness of the glass after cleaning meets the expected standards.

[0059] Among them, the movement trajectory of the wiping head, which clearly defines the route of the wiping head moving on the glass surface, determines which areas can be wiped and the order of wiping, etc., and is the key factor to ensure comprehensive cleaning of the glass. The reasonable trajectories corresponding to glass of different shapes are different, which is related to the integrity of cleaning.

[0060] Movement speed: The appropriate movement speed affects the wiping effect and cleaning efficiency. Too fast speed may lead to incomplete cleaning and missed stains; too slow speed will take too much time and reduce the overall work efficiency. It needs to be determined comprehensively based on the material of the glass, the condition of the stains and the shape.

[0061] Angle: The angle between the wiping head and the glass surface will affect the pressure distribution during wiping and the cleaning coverage. A reasonable angle can make the wiping head fit the glass better and enhance the cleaning power. Angle adjustment is even more important, especially for glass with curved or irregular surfaces.

[0062] The amount of detergent sprayed. If the amount is too little, it may not be able to fully dissolve the stains and affect the cleaning effect; if the amount is too much, it may cause a waste of detergent and may also leave residue, affecting the appearance of the glass. It needs to be precisely controlled based on the glass area, the degree of stains, etc.

[0063] The spraying angle determines the coverage and distribution of the detergent on the glass surface. It cooperates with the movement trajectory of the wiping head to ensure that the detergent is evenly distributed on the glass, so that the subsequent wiping action of the wiping head can better remove stains.

[0064] In summary, the multi-branch structure of the cleaning planning model and the generated cleaning strategy containing these key elements together provide a strong guarantee for achieving efficient, high-quality and targeted glass cleaning work.

[0065] Further optionally, for the glass samples to be tested, before collecting the glass surface information corresponding to the glass samples, for each newly added glass sample, the shape of the glass sample is identified. Then, the shooting mode of the glass sample is set based on the shape of the glass sample, and the shooting parameters in the shooting mode are fine-tuned based on the size information of the glass sample.

[0066] Among them, the shapes of glass samples include rectangular, circular, and irregular; for rectangular glass samples, the shooting angles in the shooting mode are at least two vertical directions, and the at least two vertical directions include the long side direction and the short side direction; for circular glass samples, the shooting positions in the shooting mode include at least a plurality of shooting points distributed at equal angles around the center of the circle; for irregular glass samples, the shooting angles in the shooting mode are angles that cover the glass surface to the greatest extent and include key parts, and the key parts include at least areas with large curvature changes.

[0067] In principle, the shape of the glass sample to be tested is determined through image processing technology or a pre-set geometric shape recognition algorithm. For example, for regular-shaped glass (rectangle and circle), it can be identified by detecting the number of straight line segments and the angle relationship of the boundary (rectangle), or the curvature change law of the boundary (circle). For special-shaped glass, the complexity of its boundary curve and characteristic points, such as the area with large curvature change, are mainly analyzed to determine whether it is a special shape.

[0068] Rectangular glass sample: Shooting from the long side and short side is because these two perpendicular directions can most effectively capture information about the rectangular glass surface. Shooting in the long side direction can cover a longer size range, while shooting in the short side direction can supplement the details in the vertical direction, ensuring that the flatness, scratches, stains, etc. of the entire rectangular surface can be observed.

[0069] Circular glass sample: The setting of the shooting points distributed at equal angles around the center of the circle is based on the geometric symmetry of the circle. This shooting mode can ensure that the image of the glass surface is evenly acquired from all directions, avoid missing some possible defects due to uneven shooting angles, and better reflect the curvature changes of the circular glass surface.

[0070] Shaped glass samples: Shooting at an angle that covers the glass surface to the greatest extent and includes key areas is done to take into account the irregularity of the shape of the shaped glass. Key areas such as areas with large changes in curvature are often prone to stress concentration, manufacturing defects, or more likely to accumulate stains. Focusing on these areas can more accurately detect the quality and surface condition of the glass.

[0071] The shooting parameters are fine-tuned according to the size information of the glass sample because different sizes of glass require different focal lengths, apertures and other parameters to ensure image clarity and detail integrity at the same shooting angle. For example, larger glass may require a smaller aperture to increase the depth of field to ensure that the entire glass surface is within the clear range; at the same time, the focal length may need to be adjusted so that the camera can obtain a suitable field of view to fully capture the glass surface.

[0072] By setting appropriate shooting modes and fine-tuning shooting parameters for glass samples of different shapes, more comprehensive and detailed information about the glass surface can be collected. This method can effectively reduce the detection blind area, and can detect abnormal scratches, stains, bubbles and other problems on the surface to the greatest extent, whether it is regular or irregular shaped glass, thus improving the accuracy of detection.

[0073] Because the shooting mode is set specifically according to the shape, it can ensure the quality of detection while avoiding unnecessary shooting angles and repeated shooting, thereby improving the detection efficiency. For example, for round glass, shooting at shooting points with equal angle intervals can obtain sufficient information within a limited number of shooting times, without the need for a large number of redundant shooting operations like irregular shooting.

[0074] This method can well adapt to glass samples of various shapes and sizes, whether it is common rectangular, round glass, or complex special-shaped glass, it can be properly tested according to its own characteristics. This makes the entire detection system more flexible and universal, and can meet the needs of different glass production or processing scenarios.

[0075] A rectangle has four right angles, and the opposite sides are parallel and equal. This shape of glass is commonly seen in architectural windows, picture frames, etc. The characteristics of its long and short sides make it necessary to shoot from these two perpendicular directions to fully obtain information about the glass surface.

[0076] A circle is a geometric shape where the distances from the center of the circle to the points on its boundary are equal. Circular glass is often used in optical instruments, decorations, etc. When shooting, multiple shooting points spaced at equal angles around the center of the circle can take advantage of its symmetry and evenly capture the image of the glass surface.

[0077] Special-shaped glass refers to irregularly shaped glass with complex boundary curves and no fixed geometric shape. Special-shaped glass may be used to meet special design requirements or in some non-standard application scenarios. When photographing this type of glass, it is necessary to focus on key areas such as areas with large changes in curvature to ensure that possible quality problems or surface abnormalities can be discovered.

[0078] Based on the above introduction, further optionally, calculating the curvature of the boundary of the special-shaped glass is an important method for determining the key parts. The curvature change of the boundary curve is detected by a mathematical algorithm, and the area with a large curvature change is usually the key part. For example, in computer-aided design (CAD) software or specialized image processing software, the second-order derivative of the curve is calculated using the principle of differential geometry to obtain curvature information. When the absolute value of the curvature suddenly increases or decreases, these locations are likely to be places where the glass is subjected to uneven stress, deformation, or easy to accumulate stains during the manufacturing process.

[0079] From a physical perspective, areas with large curvature may experience greater process difficulty during the molding process. For example, in the glass hot bending process, areas with drastic changes in curvature are more likely to have problems such as uneven thickness and uneven surface.

[0080] Due to its irregular shape, special-shaped glass will produce stress concentration in certain parts when subjected to external force or internal stress. These stress concentration points can be detected using finite element analysis (FEA) or photoelastic experiments. In actual operation, although a complete finite element analysis may be complicated, some simplified simulation methods can roughly determine the stress concentration area. For example, according to the shape and support method of the special-shaped glass, considering the stress distribution under the action of gravity, those points where the stress is significantly higher than other areas are the key parts.

[0081] Photoelastic testing is an experimental method that introduces polarized light into a glass model and observes the changes in the interference fringes of the light to determine the stress concentration areas. These stress concentration points are more likely to crack or break in actual use, so they need to be paid special attention to during testing.

[0082] Determine the key parts according to the manufacturing process of the special-shaped glass. If the special-shaped glass is made by cutting, micro cracks or surface roughness changes may occur at the cutting edge, near the incision, and in the area adjacent to the cutting path. These areas are key parts. For special-shaped glass made by splicing, the sealing condition of the splicing, the bonding part between the splicing material and the glass are all key parts that need to be tested.

[0083] For example, in the production of some complex special-shaped artistic glass, colored glass may be spliced ​​or decorative materials may be added to the glass surface. The joints between these decorative materials and the glass are prone to peeling, gaps and other problems, so they are critical parts.

[0084] Determine the key parts according to the actual use function of the shaped glass. If the shaped glass is used as the lens part of an optical device, then the central area where the light passes and the optical correction area on the edge are the key parts, because the quality of these areas directly affects the optical performance. If the shaped glass is used for architectural decoration, such as as part of the building facade, those areas that are directly exposed to the external environment and are susceptible to rain erosion, wind and sand erosion, or direct sunlight are the key parts.

[0085] For example, in some special-shaped glass dome buildings with special lighting functions, the top of the glass and the surrounding edge areas are key areas for detection because they are in direct contact with the outside world and are subject to greater environmental pressure.

[0086] Among the above parameters, calculating the curvature of the boundary of special-shaped glass is an important method to determine the key parts. The curvature change of the boundary curve is detected by mathematical algorithms, and the area with large curvature changes is usually the key part. For example, in computer-aided design (CAD) software or specialized image processing software, the second-order derivative of the curve is calculated to obtain curvature information using the principle of differential geometry. When the absolute value of the curvature suddenly increases or decreases, these locations are likely to be places where the glass is subjected to uneven stress, deformed, or easily accumulates stains during the manufacturing process. From a physical point of view, areas with large curvature may experience greater process difficulties during the molding process. For example, in the glass hot bending process, places where the curvature changes dramatically are more likely to have problems such as uneven thickness and uneven surface.

[0087] Due to its irregular shape, special-shaped glass will produce stress concentration in certain parts when subjected to external force or internal stress. These stress concentration points can be detected by finite element analysis (FEA) or photoelastic experiments. In actual operation, although a complete finite element analysis may be complicated, some simplified simulation methods can roughly determine the stress concentration area. For example, according to the shape and support method of the special-shaped glass, considering the stress distribution under the action of gravity, those points where the stress is significantly higher than other areas are the key parts. Photoelastic experiment is an experimental method that introduces polarized light into the glass model and observes the changes in the interference fringes of the light to determine the stress concentration area. These stress concentration points are more likely to crack or break in actual use, so they need to be paid special attention during detection.

[0088] Determine the key parts according to the manufacturing process of the special-shaped glass. If the special-shaped glass is made by the cutting process, micro cracks or surface roughness changes may occur at the cutting edge, near the incision, and in the area adjacent to the cutting path. These areas are key parts. For special-shaped glass made by the splicing process, the sealing condition of the splicing, the bonding part between the splicing material and the glass, etc. are all key parts that need to be focused on. For example, in the production of some complex special-shaped art glass, colored glass splicing may be used or decorative materials may be added to the glass surface. The bonding parts between these decorative materials and the glass are prone to peeling, gaps and other problems, so they are key parts.

[0089] Determine the key parts according to the actual use function of the shaped glass. If the shaped glass is used as the lens part of an optical device, then the central area where the light passes and the optical correction area on the edge are the key parts, because the quality of these areas directly affects the optical performance. If the shaped glass is used for architectural decoration, such as as part of the building facade, those areas that are directly exposed to the external environment and are susceptible to rain erosion, wind and sand erosion, or direct sunlight are the key parts.

[0090] For example, in some special-shaped glass dome buildings with special lighting functions, the top of the glass and the surrounding edge areas are key areas for detection because they are in direct contact with the outside world and are subject to greater environmental pressure.

[0091] Further optionally, after setting the shooting mode of the glass sample based on the shape of the glass sample, multiple light source systems matching the glass sample can also be set; specific illumination angles between the light sources in the multiple light source systems and the glass surface of the glass sample are configured; wherein the range of the specific illumination angle is 30° to 60°; based on the environment in which the glass sample is located, the illumination intensity is set to match the background light intensity of the environment; wherein the illumination intensity is inversely proportional to the background light intensity.

[0092] Specifically, you first need to use professional light intensity measuring equipment (such as a lux meter) to determine the background light intensity of the environment in which the glass sample is located. This equipment can accurately measure the light level in the environment, usually in lux. For example, in an indoor environment, there may be a higher light intensity near the window, while the light intensity is lower in the corner of the room or in a shaded area. In an outdoor environment, the light intensity on a sunny day may reach more than 100,000 lux, and on a cloudy day it may be between 1,000 and 10,000 lux. The measured background light intensity data provides a basis for subsequent adjustments to the light intensity of the light source system.

[0093] For example, modern light source systems (such as LED light sources) are equipped with controllers that can adjust the brightness. After determining the required light intensity I, the light source system is adjusted through the controller. This adjustment can be manual, where the operator operates the corresponding brightness adjustment knob on the controller according to the calculated light intensity value; or it can be automatic, where the light intensity sensor is connected to the controller of the light source system, and the control program is written so that the light source system can automatically adjust the light intensity according to the ambient light intensity fed back by the sensor to achieve the set matching value. In addition, during the adjustment process, the power limit and light intensity uniformity of the light source system also need to be considered. Ensure that while adjusting the light intensity, the light source can work stably and the light emitted is evenly distributed on the glass surface without local overbrightness or overdarkness. The uniformity of light intensity can be improved by placing a light intensity distribution test board on the glass surface or using a combination of multiple small light sources.

[0094] Furthermore, since rectangular glass has clear long and short sides, in order to fully obtain the glass surface information, the shooting angle should at least include the long side direction and the short side direction. Shooting in the long side direction can cover a longer size range and is used to observe whether there are scratches, stains or flatness problems on the glass surface in this direction. Shooting in the short side direction can supplement the details in the vertical direction to ensure that the entire rectangular surface can be fully inspected. For example, in industrial inspection, for rectangular glass used in large buildings, shooting in the long side direction may find longer scratches, while shooting in the short side direction helps to find defects at the edges. In addition to the two vertical directions, you can also consider adding oblique angle shooting. Oblique angle shooting can highlight some minor undulations or unevenness on the glass surface, because under oblique light, these minor changes will produce shadows or reflection differences, making them easier to be captured by the camera. Generally speaking, the oblique angle can be selected to be about 45° with the long side or short side, so that more details about the glass surface can be obtained without losing the main direction information.

[0095] Circular glass has a high degree of symmetry, and shooting points distributed at equal angles around the center of the circle are an ideal shooting mode. For example, the 360° circumference can be evenly divided into several intervals, such as 8, 12 or 16, and shooting can be performed at these equally divided points. This ensures that images of the glass surface are uniformly obtained from all directions, making full use of the geometric characteristics of the circle and avoiding missing some possible defects due to uneven shooting angles. Moreover, the images captured in this way can be easily spliced ​​into a complete circular glass surface image for subsequent analysis. In addition to shooting at the circumference, it is also important to shoot at the center of the circular glass. Shooting from the center position can obtain information about the central area of ​​the glass, especially for some circular glass with optical uses, the quality of the central area (such as transparency and flatness) is often more critical. Moreover, the image shot at the center position can be used as a reference, combined with the image shot at the circumference position, to more comprehensively evaluate the overall condition of the circular glass.

[0096] For special-shaped glass, its key parts should be determined first, such as areas with large curvature changes, stress concentration points, manufacturing process feature points or functional key areas. The shooting angle should be based on the principle of covering the glass surface to the greatest extent and including these key parts. For example, for special-shaped glass with complex curved shapes, focus on shooting the inner and outer surfaces of the curved parts, because these areas are prone to thickness changes, surface quality problems or stress concentration. Due to the irregular shape of special-shaped glass, it is necessary to shoot from multiple different angles to ensure that there are no blind spots in the detection. You can first determine several main observation directions as a whole, which should be able to roughly outline the shape of the special-shaped glass. Then, in each main direction, adjust the angle in a small range to obtain more detailed surface information. During the shooting process, 3D modeling software or image processing algorithms can be used to mark and splice the captured images to better build a complete surface model of the special-shaped glass.

[0097] Further optionally, the inputting the glass surface information into an anomaly detection model to perform anomaly detection on the glass sample includes:

[0098] Perform feature space projection on the glass surface information to obtain a feature space image corresponding to the glass surface information; extract candidate abnormal features in the feature space image; select a matching standard abnormal database based on the attribute information of the glass sample; the attribute information at least includes: glass material type, shooting environment information, and glass material characteristics; compare the candidate abnormal features with the standard abnormal features preset in the standard abnormal database to obtain a comparison evaluation score corresponding to the candidate abnormal features; if the comparison evaluation score reaches a set threshold, determine that the candidate abnormal feature belongs to an abnormal scratch.

[0099] For thin glass samples (e.g., less than 3mm thick), since thin glass has relatively weak refraction and reflection of light, in order to obtain clear surface details, the camera lens should be as perpendicular to the glass surface as possible and keep a relatively close shooting distance. This can reduce image deformation caused by refraction and ensure that tiny scratches, bubbles and other defects that may exist on the glass surface can be clearly captured. For example, when shooting thin glass samples such as mobile phone screen glass, the distance between the camera and the glass surface can be controlled at about 10-20 cm to obtain high-resolution images.

[0100] Thin glass is easily affected by background light and reflected light, so multiple light sources are needed to illuminate from different angles. The light source angle can be 30° to 60° relative to the glass surface, which can produce appropriate light and shadow effects on the glass surface and highlight surface details. For example, during the shooting process, two or three LED light sources are used to illuminate the glass from different directions. By adjusting the intensity and angle of the light source, defects such as scratches on the glass surface are more obvious under the light and shadow contrast. At the same time, in order to prevent reflected light from interfering with camera imaging, a polarizing filter can also be used to reduce the intensity of reflected light on the glass surface.

[0101] For some thin glass that needs to observe the internal structure, such as glass with multi-layer coating or internal inclusions, the layered shooting method can be used. By adjusting the focal length of the camera, different depth levels of the glass are photographed separately. During each shot, the focus position is precisely controlled so that the camera focuses on a specific level inside the glass, thereby obtaining clear images of each level. This method is similar to the tomography technology in medicine, which can help inspectors find tiny defects or impurities that may exist inside the glass.

[0102] For medium-thick glass samples (such as 3mm to 10mm), in addition to paying attention to surface details, medium-thick glass also needs to consider its side conditions. Multiple angles can be used, including vertical surfaces and angles parallel to the side. Vertical surface shooting is used to observe the quality of the front and back of the glass, such as scratches, stains, etc. Shooting parallel to the side can check whether there are bubbles, inclusions, and uniform thickness inside the glass. For example, when shooting medium-thick glass such as car windshields, shooting from a side angle can reveal the delamination or tiny bubbles inside the glass, which may not be easy to find when shooting vertically. For medium-thick glass, the intensity of the light source can be appropriately increased to ensure that the light can penetrate the glass and form sufficient scattering inside to facilitate observation of the internal structure. The light source angle can be adjusted between 40° and 70°. This angle range helps to produce appropriate light and shadow effects on the surface and inside of the glass. At the same time, be careful to avoid excessive reflection on the glass surface due to excessive light source, which affects the observation of surface details. Reflection can be controlled by using a lens hood or adjusting the distance between the light source and the glass.

[0103] When photographing medium-thick glass, for surface details, you can use macro shooting mode, use a macro lens or add a macro attachment to a normal lens to obtain high-magnification images and clearly show the microscopic features of the surface. At the same time, in order to observe the overall shape of the glass and the surface quality over a larger area, long-distance shooting is also required. Long-distance shooting can provide a full view of the glass, making it easier to find large-area stains, scratches, or irregular shapes.

[0104] For thick glass samples (such as those with a thickness greater than 10mm), it is more important to observe the internal structure of the thick glass, because there may be more bubbles, impurities or stress concentration points inside it. When shooting, the transmitted light shooting mode can be used, with the light source placed on one side of the glass and the camera shooting on the other side. In this way, the propagation path of light inside the glass can be clearly seen, so that internal defects can be found. For example, when shooting thick glass curtain walls for buildings, transmitted light shooting can help inspectors find stones inside the glass (an impurity produced during the glass manufacturing process), which will form obvious shadows under transmitted light. Due to the large thickness of thick glass, it is difficult to fully understand its quality from a single angle. A multi-angle stereo shooting method can be used to set multiple camera positions around the glass to shoot the glass from different directions. These cameras can shoot simultaneously or sequentially, and then these images are spliced ​​into a three-dimensional model through computer image processing technology for comprehensive analysis of the surface and internal conditions of the thick glass. During the shooting process, attention should be paid to the accuracy of the shooting angle and position of each camera to ensure that the three-dimensional structure of the thick glass can be accurately reconstructed. Thick glass will cause significant light scattering and refraction. When shooting, you need to consider these factors and adjust the shooting parameters. For example, appropriately extend the exposure time to compensate for the light loss caused by light scattering; adjust the camera's aperture and focal length to reduce image blur caused by refraction. At the same time, you can use light compensation tools or software algorithms to correct the light scattering and refraction effects of the captured image to improve the image quality so that you can better observe the various characteristics of thick glass.

[0105] For example, the collected glass surface information (including multi-dimensional data such as images under multiple light source systems) is projected into the feature space through specific mathematical transformation methods (such as principal component analysis, linear discriminant analysis, etc.) to form a feature space image. This process aims to reduce the dimension of the data, extract the data information that best represents the key features of the glass surface, remove some redundant and noise information, and make subsequent analysis and processing more efficient and focused on valuable feature performance. For example, after principal component analysis, the originally complex image pixel information is converted into a feature space image composed of a few principal components. These principal components can comprehensively reflect the characteristics of the glass surface texture, light and shadow, etc.

[0106] Based on the feature space image, image recognition and feature extraction algorithms (such as edge detection and texture analysis algorithms) are used to mine features that may indicate abnormalities. For example, if there are scratches on the glass surface, they may appear as texture interruptions, edge mutations, etc. in the feature space image. These unique manifestations can be captured and extracted by the corresponding algorithms as candidate abnormal features, which are the basic clues for subsequent judgment of whether it is a real abnormal scratch.

[0107] Since different glass material types, shooting environments, and glass material properties will affect the manifestation of abnormal features, it is necessary to select the corresponding standard abnormal database based on these property information of the glass sample. For example, under the same scratch condition, the feature performance of tempered glass and ordinary float glass in the image may be different due to different material hardness, transparency and other characteristics; similarly, different shooting environments such as lighting and background will also cause abnormal features to present different states. Only by matching the accurate database can a reasonable and effective comparison be made.

[0108] When the extracted candidate abnormal features are compared with the preset standard abnormal features in the standard abnormal database, a comparison evaluation score will be given through a set quantitative algorithm based on factors such as feature similarity. This score comprehensively considers the matching degree of multiple aspects such as feature shape, position, and intensity, so as to measure the closeness between the candidate abnormal features and the standard abnormal features, and then determine whether it is an abnormal scratch.

[0109] In this way, the multi-faceted attribute information of the glass sample is taken into account and compared with the standard database, avoiding the one-sidedness of relying solely on intuitive judgment of the image, and being able to more accurately identify the real abnormal scratches and reduce misjudgment. For example, some glass surfaces appear to have scratches under a specific angle of illumination due to special material textures, but by comparing with the corresponding standard, this misjudgment can be accurately eliminated. It can adapt to the detection of glass samples of different materials and different shooting environments. Whether it is high-precision glass used for optical instruments or ordinary glass used for construction, no matter the sample information obtained under various shooting environments such as strong light and weak light, it can find a suitable comparison basis based on its attributes to achieve effective abnormal detection, expanding the scope of application of the entire detection process. By comparing and evaluating the scores to determine abnormal scratches, the detection results have a quantitative measurement standard, which is no longer a simple "yes" or "no" judgment, which is convenient for further analysis of the severity of the abnormality according to the score, and is also conducive to optimizing and adjusting the detection system, and improving the algorithm and parameter settings based on the matching degree between the score and the actual situation.

[0110] For example, suppose you want to detect a piece of x glass sample for curtain wall (the material type is determined), the shooting environment is that the indoor lighting is relatively uniform and the background is white (the shooting environment information is clear), and the material has characteristics such as high hardness and high transparency (the material characteristics are known). First, the collected glass surface information is projected into the feature space to obtain the corresponding feature space image, from which candidate abnormal features such as texture interruption and local light and shadow mutation are extracted. Then, according to the above attribute information, a standard abnormal database matching curtain wall glass in indoor environment is selected and the candidate abnormal features are compared with it. For example, if it is found that the extracted texture interruption feature is highly similar to the standard abnormal feature corresponding to the scratch preset in the standard abnormal database in terms of shape, position or relative intensity, and the comparison evaluation score reaches the set threshold, then it can be determined that the candidate abnormal feature here belongs to abnormal scratch, that is, there is scratch on the surface of the glass sample, and it is necessary to further determine the subsequent cleaning or other processing procedures based on the severity of the scratch.

[0111] As an optional embodiment, generating a cleaning strategy corresponding to the glass sample through a cleaning planning model includes:

[0112] The glass sample is comprehensively scanned by a laser scanning module in a visual inspection system to obtain depth feature information of the glass sample; a three-dimensional point cloud model of the glass sample is constructed based on the depth feature information; and a planning prediction branch in a cleaning planning model that matches the three-dimensional point cloud model is used to generate a cleaning strategy corresponding to the glass sample.

[0113] The planning prediction branch matches the glass sample shape and size parameters of the three-dimensional point cloud model. The cleaning planning model includes multiple planning prediction branches, each of which is used to adaptively construct a corresponding cleaning strategy based on the glass sample shape; the cleaning strategy at least includes: the motion trajectory, motion speed, angle, detergent spraying amount, and spraying angle of the wiping head.

[0114] Specifically, the laser scanning module in the visual inspection system works based on the principle of laser ranging. It emits a laser beam to the surface of the glass sample. When the laser beam contacts the glass surface, it will be reflected. The sensor in the laser scanning module receives the reflected laser signal and calculates the distance information from each scanning point to the sensor based on the laser propagation time and known laser propagation speed and other parameters. By scanning the surface of the glass sample point by point, a large amount of distance data at different positions can be obtained. This data reflects the depth of each position on the glass surface, that is, the depth feature information. For example, for a glass surface with a certain curvature or unevenness, the distance measured at different positions is different, and these difference data constitute part of the depth feature information.

[0115] It can accurately obtain information related to the three-dimensional morphology of the glass sample surface. Compared with the information obtained only through two-dimensional images, the depth feature information can more comprehensively display the actual shape of the glass. Whether it is a flat surface, a curved surface or a special-shaped surface with complex undulations, it can be recorded in detail to provide basic data for the subsequent construction of an accurate three-dimensional model, avoiding the omission of surface morphology details that have an important impact on the cleaning strategy due to relying only on a two-dimensional perspective.

[0116] Thus, the numerous depth feature information (three-dimensional coordinate information of each scanning point) acquired by the laser scanning module are integrated to form discrete points, and the collection of many such points in space constitutes a three-dimensional point cloud model. Each point contains its position coordinates in three-dimensional space (usually expressed in X, Y, and Z axis coordinates), and these point clouds present the overall shape and surface details of the glass sample from a macroscopic perspective. For example, for a circular glass dome, point clouds at various positions are obtained through laser scanning, and the arc shape with high middle and low edges, as well as the point cloud distribution corresponding to possible surface unevenness, can be seen in the model.

[0117] In this way, the three-dimensional form of the glass sample is intuitively displayed in a visual and digital form, which facilitates the subsequent accurate extraction and analysis of key parameters such as glass shape and size. At the same time, the three-dimensional point cloud model can be easily combined with computer-aided design (CAD) software, computer vision algorithms, etc., to provide a standard and universal data format basis for further processing and decision-making (such as matching cleaning planning branches, simulating cleaning processes, etc.), which helps to improve the automation and intelligence level of the entire cleaning strategy generation process.

[0118] The cleaning planning model pre-sets multiple planning prediction branches, and each branch sets the corresponding cleaning strategy for glass samples with different shapes and size parameters. After the three-dimensional point cloud model of the glass sample is constructed, the shape of the glass sample (such as round, rectangular, special-shaped, etc.) and the specific size parameters (length, width, height, radius, etc.) are extracted by analyzing the model, and then the matching planning prediction branch is found in the cleaning planning model. For example, if the three-dimensional point cloud model shows that the glass sample is a long strip of special-shaped glass with a certain curvature, the system will find a planning prediction branch specifically for this type of shape and size characteristics. This branch is based on the preset algorithm and rules, considering the specific conditions of the glass surface, to determine the specific parameters of the cleaning strategy such as the motion trajectory of the wiping head (such as the fitting path along the curvature, the key cleaning trajectory of the two ends of the strip, etc.), the movement speed (reasonably set the speed according to the difficulty of surface cleaning and the material conditions), the angle (the fitting angle with the curved surface or special-shaped edge), the amount of detergent sprayed (determined according to the area size, the degree of stains, etc.) and the spraying angle (to ensure that the detergent evenly covers the corresponding area).

[0119] In this way, adaptive generation of cleaning strategies is achieved, and cleaning plans that are highly in line with the actual situation can be formulated for glass samples of different shapes and sizes, greatly improving cleaning efficiency and cleaning quality. Compared with the use of a unified fixed cleaning strategy, this strategy based on matching branch generation can avoid problems such as incomplete cleaning (such as missing some complex corners of special-shaped glass), damage to the glass surface (such as scratching the glass due to improper wiping angles), and waste of cleaning agents due to unreasonable cleaning methods, making the entire cleaning process more scientific, accurate, and efficient, and meeting the cleaning needs of diverse glass samples.

[0120] In summary, through the above series of steps, using the laser scanning module, building a three-dimensional point cloud model, and matching planning prediction branches, a complete and intelligent glass sample cleaning strategy generation system is formed, which helps to improve the overall level of glass cleaning work.

[0121] Further optionally, each planning prediction branch in the cleaning planning model is classified and designed according to the shape (such as circle, rectangle, triangle, polygon and various irregular shapes) and size (including specific parameters such as length, width, height, radius, etc.) of the glass sample. For example, for the planning prediction branch of the circular glass sample, its internal algorithm and rules are built based on the geometric characteristics of the circle. It takes into account the symmetry of the circumference. When generating the cleaning strategy, the motion trajectory of the wiping head may move in a circle around the center of the circle. The motion speed will be set according to the radius of the circle and the expected cleaning time, and the amount of detergent sprayed will also be calculated according to the area of ​​the circle. These branches are usually trained with a large amount of sample data. During the training process, data on successful cleaning cases of glass samples of various shapes and sizes are collected, including the corresponding cleaning strategy parameters (wiping head motion trajectory, speed, angle, detergent spray amount and angle) and the effect evaluation after cleaning. Using machine learning algorithms, such as neural networks, decision trees, etc., these data are learned so that each planning prediction branch can predict a suitable cleaning strategy based on the input glass sample shape and size information.

[0122] In addition to being data-driven, the planning and prediction branch also incorporates knowledge from the physical and chemical aspects. For example, from a physical perspective, the friction between the wiping head and the glass surface, the flow and diffusion characteristics of the detergent on the glass surface, etc. are considered; from a chemical perspective, the appropriate detergent composition and spraying amount are determined based on the glass material and the type of stains (oil stains, dust, water stains, etc.). This knowledge is embedded in the branch's algorithms and rules to ensure that the generated cleaning strategy is both efficient and safe in practical applications.

[0123] The circular glass branch mainly makes circular motion around the center of the circle to ensure that the entire circular surface can be wiped evenly. The radius of the motion trajectory can be adjusted according to the radius of the glass, and multiple layers of circular motion trajectories may be set according to the cleaning requirements, such as first performing a spiral circular motion from the center of the circle to the outside, and then performing a reverse spiral motion from the outside to the center of the circle to ensure comprehensive cleaning. Considering the characteristics of circular motion, the speed is relatively stable. The speed setting is related to the radius of the circle. Generally speaking, the larger the radius, the faster the speed can be, but the material and degree of stains of the glass should also be considered. For example, for optical lens glass with a larger radius, due to its high requirements for surface quality, the movement speed will be moderate to avoid incomplete cleaning or damage to the lens surface due to excessive speed. The angle between the wiping head and the glass surface is kept vertical or nearly vertical to ensure that the wiping pressure is evenly distributed on the glass surface during the circular motion. For circular glass with special optical coatings, precise control of the angle is particularly important to prevent damage to the coating. The spraying amount is calculated based on the area of ​​the circle to ensure that the detergent can evenly cover the entire circular surface. The spraying angle is generally perpendicular to the glass surface, or inclined to a certain extent according to the direction of the circular motion, so that the detergent is evenly distributed under the drive of the circular motion to avoid local accumulation.

[0124] In the rectangular glass branch, the motion trajectory includes linear motion along the long side and short side. It can be that multiple reciprocating wiping is performed along the long side first, and then additional wiping is performed along the short side to ensure that all four corners and edges are cleaned. For large rectangular glass, a diagonal wiping trajectory may also be designed to deal with possible stain accumulation or difficult-to-clean areas. The movement speed in the long side and short side directions can be different. If the long side is longer, the speed may be slightly faster to improve cleaning efficiency, but it will slow down appropriately when approaching edges and corners. The speed in the short side direction is adjusted according to the width of the rectangle and the stain situation. For example, for a large rectangular glass curtain wall used in a building, the wiping speed in the long side direction can be optimized according to the height of the curtain wall and the capacity of the cleaning equipment, while the short side direction will slow down when cleaning near the window frame to ensure the cleaning quality of the junction between the window frame and the glass. During the wiping process, the angle between the wiping head and the glass surface is kept within a suitable range to ensure uniform and effective wiping pressure. At the edges and corners of the rectangle, the angle will be adjusted appropriately so that the wiper head can better fit the corners and avoid missing stains. The spraying amount is calculated based on the area of ​​the rectangle, and considering that the edges and corners may need more detergent to deal with stubborn stains, these areas will be sprayed appropriately. The spraying angle can be perpendicular to the glass surface, or it can be tilted according to the direction of movement of the wiper head, so that the detergent can be evenly distributed along the wiping direction, improving the cleaning effect.

[0125] In the special-shaped glass branch, the motion trajectory of the wiper head can be configured based on the specific shape and key parts of the special-shaped glass (such as areas with large curvature changes, recessed or convex parts). The motion trajectory will focus on multiple circles or fit wipes around the key parts. For irregular edges and complex curved surfaces, flexible and changeable trajectories are adopted, such as moving along the tangent direction of the curve, reciprocating in and out motion in the recessed area, etc., to ensure that all surface areas can be cleaned. The speed varies greatly in different parts of the special-shaped glass. In relatively flat and easy-to-clean areas, the speed can be appropriately increased; while in key areas, especially those with complex shapes and easily damaged areas, the speed will be significantly slowed down. For example, for special-shaped artistic glass with fine carvings or textures, the wiping speed in the carving or texture area will be very slow to ensure the cleaning quality and the integrity of the glass surface. The angle of the wiper head will be continuously adjusted according to the curvature and shape of the special-shaped glass surface. On the convex part, the angle of the wiper head will tilt outward to fit the convex shape; on the concave part, the angle will tilt inward. When switching between areas with different curvature changes, the angle adjustment should be smooth to avoid damage to the glass surface. The spraying amount is allocated according to the area of ​​the special-shaped glass and the cleaning difficulty of different parts. For recessed areas, gaps or complex curved surfaces where stains are easily accumulated, the amount of detergent sprayed will be increased. The spraying angle will be adjusted according to the shape of the special-shaped glass and the movement trajectory of the wiping head to ensure that the detergent can smoothly reach all parts that need to be cleaned, especially some hidden or difficult-to-reach areas.

[0126] In addition, further, in the actual cleaning process, feedback information after the execution of the cleaning strategy is collected, such as cleaning effect evaluation (whether there are stains remaining, whether the glass surface is damaged, etc.), cleaning time, detergent usage and other data. These feedback data are sent back to the corresponding planning prediction branch, and the parameters and rules in the branch are updated and optimized through the machine learning algorithm. For example, if it is found that a certain shape of glass has more stains remaining after cleaning according to the current cleaning strategy, the system will analyze which link has problems, such as the motion trajectory, speed, and detergent spraying of the wiping head, and then adjust the relevant parameters in the planning prediction branch in a targeted manner. When encountering new glass shapes or sizes, or new cleaning requirements and conditions, these new sample data are collected, added to the training data set, and the planning prediction branch of the cleaning planning model is retrained. In this way, the applicability and accuracy of the model can be continuously expanded, so that it can cope with more diverse glass cleaning tasks. For example, as new shaped glass is more and more widely used in architectural decoration or industrial products, these new shaped glass sample data are added to the training, and the rules and parameters of the shaped glass branch are updated to better meet the cleaning needs.

[0127] Further optionally, the adopting a strategy branch in the cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes:

[0128] If the shape of the glass sample of the three-dimensional point cloud model is a rectangle, any corner point in the three-dimensional point cloud model is obtained through the first planning prediction branch; with the corner point as the starting point, the cleaning strategy is set as follows: along the long side direction of the three-dimensional point cloud model, a linear movement is performed at a constant movement speed, and an adaptive wiping pressure and a cleaning agent spraying amount are used for wiping operations; when the movement reaches the end point of the long side, the movement turns and moves in a straight line along the short side direction of the three-dimensional point cloud model, and an adaptive wiping pressure and a cleaning agent spraying amount are used for wiping operations; when the movement reaches the end point of the short side, the steps of moving in a straight line along the long side direction of the three-dimensional point cloud model at a constant movement speed, and an adaptive wiping pressure and a cleaning agent spraying amount are repeated until the cleaning process of the glass sample is completed; wherein the wiping pressure and the cleaning agent spraying amount are adaptively configured based on the degree of dirtiness of the glass surface detected in real time during the movement.

[0129] For rectangular glass samples, any corner point in the 3D point cloud model is selected as the starting point. This is because a rectangle has four right angles, and starting from a corner point, it is convenient to clean regularly along the sides of the rectangle. A corner point is the intersection of two sides. Using this as the starting point can naturally guide the wiping head to clean in sequence along the long and short sides, ensuring that the entire rectangular surface can be covered.

[0130] The linear motion is simple and efficient, with a constant speed along the long side. Since the long side is relatively long, maintaining a constant speed ensures the stability and predictability of the cleaning process, making it easier to control the cleaning time and quality. During the movement, the wiping pressure and the amount of detergent sprayed are adjusted according to the degree of dirt on the glass surface detected in real time. If a certain area is detected to be heavily dirty, the system will automatically increase the wiping pressure and the amount of detergent sprayed to ensure that the stains can be effectively removed. For example, when stubborn oil stains are detected on the glass surface, appropriately increasing the wiping pressure can make the wiping head better contact with the glass surface, while increasing the amount of detergent sprayed will help dissolve and remove the oil stains. This is based on physical and chemical principles, and overcomes the adhesion between the stains and the glass surface by applying appropriate external force and sufficient detergent.

[0131] When the wiping head reaches the end of the long side, it turns and moves in a straight line along the short side. This steering mechanism ensures that the cleaning path can cover the entire surface of the rectangle. The cleaning in the short side direction also uses adaptive wiping pressure and detergent spraying amount. The principle is the same as that in the long side direction, which is to effectively clean according to the dirtiness.

[0132] By repeating the cleaning steps in the long and short sides direction, the cleaning process of the entire glass sample is completed. This cycle ensures that the entire rectangular glass surface is thoroughly and evenly cleaned without missing any area.

[0133] This cleaning strategy ensures that all parts of the rectangular glass surface can be cleaned. Starting from the corner point, the circular cleaning path along the long and short sides can cover the entire rectangular area, including the edges and corners. Compared with random cleaning paths or irregular cleaning methods, this sequential cleaning strategy can effectively reduce cleaning blind spots and improve cleaning integrity.

[0134] The constant movement speed in the long side direction helps improve cleaning efficiency. At the same time, combined with the adaptive wiping pressure and detergent spraying amount, it can be flexibly adjusted according to the actual degree of dirtiness to ensure the cleaning quality. When encountering lightly dirty areas, it will not waste time and detergent by over-cleaning; in heavily dirty areas, effective cleaning measures can be taken to ensure that the stains are completely removed.

[0135] Adaptive wiping pressure can protect the wiping equipment and glass surface. When cleaning does not require high pressure, the wiping head will not be damaged or the glass will be scratched due to excessive pressure; when encountering stubborn stains and needing to increase pressure, it can also be adjusted within a reasonable range to avoid excessive wear of the equipment and damage to the glass surface.

[0136] As an optional embodiment, the adopting of a strategy branch in the cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes:

[0137] If the shape of the glass sample of the three-dimensional point cloud model is a circle, the center position of the three-dimensional point cloud model is determined through the second planning prediction branch; with the center position as the starting point, the cleaning strategy is set as follows: wiping is performed outward along the path of the Archimedean spiral; during the wiping process, the wiping pressure is adjusted according to the radius length, and the amount of cleaning agent sprayed is increased according to a preset functional relationship with the radius length.

[0138] For circular glass samples, determining the center position is a key starting point for the cleaning strategy. This is based on the geometric properties of the circle, where the center is the central symmetric point of the circle. By performing geometric analysis on the 3D point cloud model, an algorithm is used to accurately find this center position. For example, by calculating the distance from each point in the point cloud model to the model boundary, a set of points with equal distances are found to determine the center position. This center position provides a central reference for the subsequent spiral cleaning path.

[0139] The Archimedean spiral is chosen as the wiping path because it can cover the entire circular surface in a uniform and continuous manner. The Archimedean spiral is characterized by starting from the center of the circle, and as the angle increases, the radius increases at a fixed rate, which ensures that the wiping head can pass through each part of the circular surface in sequence during the outward expansion process, avoiding duplication or omission of the cleaning path. In the polar coordinate system, during the cleaning process, the wiping head moves according to this path equation, and as the angle continues to change, the radius gradually increases, thereby achieving comprehensive cleaning from the center of the circle to the circumference.

[0140] Adjust the wiping pressure according to the radius length. This is to take into account that on a circular surface, as the radius increases, the linear velocity of the wiping head increases (assuming the angular velocity of the wiping head is constant). According to the principles of physics, in order to ensure the consistency of the cleaning effect, the wiping pressure needs to be adjusted appropriately. In areas with a small radius, the wiping head moves relatively slowly, and a smaller wiping pressure can be used at this time; when the radius increases and the movement speed increases, the wiping pressure should be appropriately increased to ensure that the friction between the wiping head and the glass surface can effectively remove stains.

[0141] As the radius length increases, the amount of detergent sprayed increases according to a preset functional relationship, because the area of ​​a circle is proportional to the square of the radius. During the cleaning process, as the radius increases, the area to be cleaned also increases. In order to ensure that the detergent can evenly cover the glass surface, the spraying amount needs to be increased accordingly. For example, the preset function can be a linear relationship (such as the spraying amount is proportional to the radius) or a quadratic function relationship (such as the spraying amount is proportional to the square of the radius). The specific functional relationship can be determined based on factors such as the properties of the detergent, the material of the glass surface, and the type of stains.

[0142] In this way, the wiping path of the Archimedean spiral can efficiently cover the entire surface of the circular glass, avoiding cleaning dead corners. Compared with other irregular or discontinuous cleaning paths, this path can complete the cleaning of the circular glass in one continuous movement, greatly improving the cleaning efficiency. By adjusting the wiping pressure according to the radius length, the cleaning effect on the entire circular surface is more consistent. Whether in the area close to the center of the circle or the circumference, it can ensure that the wiping head has appropriate friction with the glass surface, effectively remove stains, and improve the stability of the cleaning quality. Increasing the amount of detergent sprayed with the radius according to the preset functional relationship can ensure the uniform distribution of the detergent on the surface of the circular glass. This not only avoids the waste caused by spraying too much detergent in the central area, but also ensures that there is enough detergent to clean stains in large areas such as the circumference, thereby realizing the rational use of detergent.

[0143] It is understandable that different glass materials have different adsorption and responsiveness to cleaning agents. For example, ordinary soda-lime glass and tempered glass will have different spreading and penetration of cleaning agents on their surfaces due to differences in their chemical composition and microstructure. For glass materials with strong adsorption capacity, in the preset function relationship, the increase in the amount of cleaning agent sprayed with the radius may be relatively small, because the cleaning agent can spread better on the surface; while for glass with weaker adsorption capacity, the spraying amount may need to be increased faster to ensure the cleaning effect.

[0144] The roughness of the glass surface will also affect the amount of detergent used. If the glass surface is rough, stains are easily hidden in pits and gaps, and more detergent is needed to fully contact and dissolve the stains. At this time, the slope of the detergent spray amount with the increase of radius in the preset function relationship may be larger to adapt to the demand for detergent on rough surfaces. For example, frosted glass surfaces require more detergent than ordinary smooth glass surfaces. The function relationship may be adjusted according to the surface roughness test data, such as using a quadratic function relationship with a larger quadratic coefficient.

[0145] Different types of stains require different amounts of detergent to remove. For example, oil stains usually require more detergent to dissolve because the oil has a strong adhesion to the glass surface; while dry stains such as dust are relatively easy to remove and require less detergent. If the glass sample is mainly contaminated with oil stains, the preset function relationship will increase the amount of detergent sprayed with the radius more, and an exponential function relationship may be used to ensure that there is enough detergent to break down the oil stains. For glass contaminated with dust, the function relationship may be smoother, such as a linear relationship with a smaller slope.

[0146] The degree of dirt is also an important factor in determining the preset functional relationship. If the glass surface is heavily stained, more detergent is needed to clean it thoroughly. The functional relationship can be adjusted by evaluating the degree of dirt on the glass surface (optical detection, image analysis, etc. can be used). For example, in areas with high dirt levels, the speed at which the amount of detergent sprayed increases with the radius can be increased, such as increasing the slope of the linear function or adjusting the coefficient of the quadratic function, so that the detergent can fully cope with the dirt.

[0147] The performance of the cleaning equipment's detergent spray system will limit the preset functional relationship. For example, factors such as the spray pressure, nozzle type, and spray pattern of the spray system will affect the actual spray effect of the detergent. If the pressure of the spray system is low and the detergent droplets sprayed from the nozzle are larger, the functional relationship may need to be adjusted so that the detergent spray volume increases more when the radius increases to ensure that the detergent can evenly cover the glass surface.

[0148] Parameters such as the movement speed and angle of the wiping head will also affect the distribution and utilization efficiency of the detergent. If the wiping head moves faster, the detergent needs to cover the glass surface faster. At this time, the speed at which the detergent spraying amount increases with the radius in the preset function relationship may need to be accelerated to avoid the situation where the cleaning area is not covered by the detergent in time. At the same time, the angle between the wiping head and the glass surface will also affect the flow direction and distribution of the detergent, which also needs to be considered when determining the function relationship.

[0149] Before the actual cleaning operation, experiments can be conducted on glass samples of the same material and shape to simulate different stain conditions and record the amount of detergent sprayed to achieve the ideal cleaning effect at different radius positions. Through multiple experiments, data is collected and analyzed to fit a suitable functional relationship. For example, stains of different degrees are manually applied to a circular glass sample, and the sample is cleaned according to the set wiping path and movement speed. The amount of detergent sprayed is adjusted each time, and the cleaning effect is observed. Based on the experimental results, a functional relationship between the amount of detergent sprayed and the radius is constructed.

[0150] Using computer simulation software, a physical model of glass cleaning is established, and information such as glass surface characteristics, stain type and degree, and cleaning equipment parameters are input to simulate the distribution of detergent on the glass surface and the cleaning process. By continuously adjusting the parameters in the preset functional relationship and observing the simulation results, the optimal functional relationship is found so that the detergent can achieve a good cleaning effect on the glass surface in the most economical and efficient way. For example, different functional relationships are set in the simulation software to simulate the spraying and flow process of the detergent, and the cleaning effects and detergent usage under different settings are compared to determine the most appropriate preset functional relationship.

[0151] As an optional embodiment, in the specific manner of increasing the amount of detergent sprayed according to a preset functional relationship as the radius length increases, for cleaning the glass surface, different amounts of detergent are required for different degrees of dirtiness to achieve the best cleaning effect. By dividing the degree of dirtiness into three intervals of light (0-30), medium (30-70) and heavy (70-100), the amount of detergent sprayed can be controlled more finely. In each interval, a linear function is used to describe the relationship between the amount of detergent sprayed and the degree of dirtiness, because within a certain range of dirtiness, the amount of detergent used and the degree of dirtiness are roughly linearly related. For example, in the light dirtiness interval, a small amount of detergent may be sufficient to clean the glass surface, and as the degree of dirtiness increases, the amount of detergent sprayed increases linearly accordingly (Q = k1D + b1). Here, k1 and b1 are coefficients determined experimentally or empirically, k1 represents the rate at which the amount of detergent sprayed increases with the degree of dirtiness, and b1 may be a basic spraying amount used to handle the minimum amount of detergent required even under light dirtiness.

[0152] In the moderately dirty interval (Q = k2D + b2) and the heavily dirty interval (Q = k3D + b3), the effects of k2, k3 and b2, b3 are similar to those in the lightly dirty interval, but the values ​​of the coefficients are different due to the different degrees of dirtiness. k2 and k3 are usually greater than k1 because more detergent is needed to ensure the cleaning effect as the degree of dirtiness increases. The values ​​of b2 and b3 may be adjusted according to actual conditions. For example, considering some characteristics of the glass surface under moderate dirtiness and heavy dirtiness, such as the stronger adhesion of stains, an additional basic spraying amount may be required to start the cleaning process. These coefficients can be determined experimentally. For example, in a laboratory environment, prepare multiple round glass samples of the same material and shape, and create different degrees of dirtiness (different degrees of dirtiness can be simulated by applying known amounts of dust, oil, etc.), and then use a wiping robot to clean along the set path. During the cleaning process, adjust the amount of detergent sprayed, record the spraying amount data when a good cleaning effect is achieved under different degrees of dirtiness, and obtain the coefficients of each interval through data fitting.

[0153] Dirt level sensors can be implemented using a variety of technologies. One common approach is an optical sensor, which determines the degree of dirt on the glass surface by emitting light of a specific wavelength and detecting the intensity and spectral characteristics of the reflected light. When there is dirt on the glass surface, the intensity and spectrum of the reflected light will change, and the sensor quantifies the degree of dirtiness based on these changes. For example, for clear glass, the intensity of reflected light on a clean glass surface is low, while the intensity of reflected light on a stained surface will increase, and the spectrum of reflected light will vary depending on the type of dirt. The degree of dirtiness can be determined by analyzing these parameters.

[0154] Another possible way is to use a capacitive sensor. The presence of stains changes the dielectric constant of the glass surface, thereby affecting the capacitance value of the capacitive sensor. By measuring the change in capacitance value, the degree of dirtiness on the glass surface can be inferred. After receiving the data from the dirtiness sensor, the control system of the wiping robot first determines which range the dirtiness D is in (mild, medium or heavy). Then, the amount of detergent sprayed Q is calculated according to the corresponding piecewise linear function. For example, if the sensor detects that the dirtiness D is 40, the control system determines that it is in the medium dirtiness interval and substitutes D = 40 into Q = k2D + b2 to calculate the required amount of detergent sprayed. The control system adjusts the parameters of the detergent spraying system in real time according to the calculation results, such as the flow control of the micro pump, to achieve accurate detergent spraying. This ensures that during the cleaning process, the amount of detergent sprayed always matches the actual dirtiness of the glass surface.

[0155] In this way, by adjusting the amount of detergent sprayed in real time according to the degree of dirtiness, it is possible to avoid incomplete cleaning due to insufficient detergent use, or waste and residue due to excessive detergent use. In lightly dirty areas, only an appropriate amount of detergent is used, which not only effectively cleans the glass but also saves detergent. In heavily dirty areas, sufficient detergent can be provided to deal with stubborn stains and ensure the cleaning quality. Since the amount of detergent sprayed is closely matched to the degree of dirtiness, the cleaning process is more scientific and reasonable. For areas with different degrees of dirtiness, the most suitable cleaning conditions can be provided, thereby improving the cleaning quality of the entire glass surface, enabling the glass to achieve better cleaning effects and meet different cleaning needs.

[0156] This piecewise linear function control method can maximize the use of detergent. Compared with the fixed spraying amount cleaning method, it can accurately distribute detergent according to actual needs and reduce unnecessary waste. Especially when cleaning large areas of glass or a large number of glass samples, it can significantly reduce the cost of detergent use. Accurate detergent spraying helps to improve cleaning efficiency. Because with the appropriate amount of detergent, stains can be dissolved and removed faster, and the wiping effect of the wiping head is better, thereby reducing cleaning time. At the same time, it avoids the residual processing time after cleaning caused by excessive detergent, further improving the overall cleaning efficiency.

[0157] As an optional embodiment, the adopting of a strategy branch in the cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes:

[0158] If the glass sample shape of the three-dimensional point cloud model is irregular, the contour of the three-dimensional point cloud model is segmented into multiple line segments and / or curve segments through the third planning prediction branch; the cleaning sequence corresponding to the three-dimensional point cloud model is predicted; and based on the cleaning sequence, wiping operations are performed along the multiple line segments and / or curve segments in sequence.

[0159] For special-shaped glass samples, their shapes are complex and irregular. By performing shape analysis on the three-dimensional point cloud model, the contour is divided into multiple line segments and / or curve segments using a geometric algorithm. This process is similar to decomposing a complex figure, splitting the boundary of the special shape into basic geometric elements that are easier to handle. For example, for special-shaped glass with a combination of curved edges and straight edges, by identifying the change points of curvature on the boundary, the continuous contour is divided at the point where the curvature suddenly changes to obtain straight line segments and curve segments. In terms of algorithm implementation, a method based on differential geometry can be used to calculate the curvature of each point on the contour. When the curvature exceeds a certain threshold, it is determined to be the start or end point of the curve segment, thereby dividing different line segments and curve segments. These line segments and curve segments can more accurately describe the contour characteristics of special-shaped glass and provide a basis for the formulation of subsequent cleaning strategies.

[0160] When predicting the cleaning order, the shape characteristics of the special-shaped glass and the convenience of cleaning need to be considered comprehensively. Generally, those parts that are easy to reach and have a larger area are cleaned first to improve cleaning efficiency. For example, for special-shaped glass with concave or convex parts, the convex parts may be cleaned first because these parts are relatively easier to be touched by the wiping head, and pollutants such as sewage after cleaning are not easy to accumulate on them.

[0161] At the same time, the continuity of the cleaning path and the avoidance of repeated cleaning must also be considered. By analyzing the connection relationship between each line segment and curve segment, an optimal cleaning sequence is determined so that the wiping head can naturally transition from one part to another during the movement, reducing unnecessary path turns and jumps. For example, according to the topological structure of the special-shaped glass, the adjacent line segments and curve segments are arranged in a certain order so that the cleaning process can be carried out along a continuous path, similar to planning the most reasonable travel route on a map.

[0162] Based on the determined cleaning sequence, the control system of the wiping robot controls the wiping head to perform wiping operations along multiple line segments and curved segments in sequence. When wiping along a line segment, the motion trajectory of the wiping head is relatively simple, mainly linear motion. By controlling the movement of the robot's mechanical arm joints, the wiping head can maintain stable movement on the straight line segment. For wiping of curved segments, more complex motion control is required.

[0163] The control system adjusts the speed and angle of the wiping head according to the curvature and length of the curve segment. For example, in a curve segment with a large curvature, the speed of the wiping head will be appropriately slowed down to ensure that it can closely follow the curve for cleaning; at the same time, the angle of the wiping head will be continuously adjusted according to the tangent direction of the curve, so that the wiping head and the glass surface always maintain a suitable contact angle, thereby effectively removing stains.

[0164] In this way, by dividing the contour into line segments and curve segments and wiping them in a reasonable cleaning order, it can be ensured that every part of the special-shaped glass can be cleaned. This method effectively overcomes the cleaning difficulties caused by the irregular shape of special-shaped glass. Whether it is glass with complex depressions, protrusions or twisted shapes, this sophisticated strategy can achieve comprehensive cleaning and reduce cleaning dead corners.

[0165] Because the wiping operation can closely follow the contour of the special-shaped glass and is carried out in a reasonable order, the stains can be removed more thoroughly. During the cleaning process, the targeted wiping of each line segment and curve segment, as well as the orderly transition between different parts, can help improve the quality of cleaning and achieve better cleaning results for special-shaped glass.

[0166] Reasonable cleaning sequence prediction can reduce the length of the wiping head's movement path on the surface of special-shaped glass. By avoiding unnecessary path return and repeated cleaning, the cleaning efficiency is improved. For example, in the cleaning of a special-shaped glass with multiple branch shapes, following the optimal cleaning sequence can allow the wiping head to cover most of the area in one continuous movement, rather than shuttling back and forth between the branches, thus saving cleaning time. The optimized cleaning path and motion control method can reduce the frequent start and stop and sharp turns of the wiping head and the robotic arm. This smooth movement process helps to reduce equipment wear, extend the service life of the wiping robot, and reduce equipment maintenance costs.

[0167] As an optional embodiment, wiping operations are performed along multiple line segments and / or curve segments in sequence based on the cleaning order, including: when wiping a line segment, moving in a straight line at a constant motion speed, and using adaptive wiping pressure and detergent spraying amount to perform a straight line wiping operation; and / or, when wiping a curve segment, adjusting the motion trajectory, speed and angle of the wiping head according to the curvature and torsion of the curve segment through a motion control algorithm.

[0168] Moving in a straight line at a constant speed makes the cleaning process predictable and stable. When wiping the line segment, the robot can move steadily at a set speed, which helps to plan the cleaning time. For example, when cleaning a certain number of special-shaped glasses on an industrial production line, the constant speed allows the operator to accurately estimate the cleaning time of each glass, so as to better arrange the production process. This stable speed also helps to improve cleaning efficiency. Because there is no need to adjust the speed frequently, the wiping head can move continuously on the glass surface, reducing the pause time that may be caused by speed changes. Moreover, during the straight-line movement, the continuity of the cleaning path can be guaranteed, avoiding uneven cleaning or missing some areas due to speed fluctuations.

[0169] Adaptive wiping pressure can be flexibly adjusted according to the actual degree of dirt on the glass surface and the material characteristics of the glass. For dirty areas or glass with harder materials, appropriately increasing the wiping pressure can better remove stains. For example, when cleaning some glass segments with stubborn oil stains, the adaptive system will increase the wiping pressure based on the dirt information fed back by the sensor, so that the wiping head can contact the stains more effectively, thereby improving the cleaning effect. At the same time, adaptive wiping pressure can also protect the glass surface. When cleaning cleaner areas or glass with more fragile materials, the pressure will automatically decrease to avoid scratching the glass due to excessive pressure. This flexible adjustment method based on actual conditions ensures high-quality cleaning under different conditions.

[0170] Adaptive detergent spraying ensures that the amount of detergent used matches the degree of dirt on the glass surface during straight-line wiping. In areas with less dirt, reducing the amount of detergent sprayed can avoid waste and reduce residue after cleaning. In areas with heavier dirt, increasing the spraying amount can effectively dissolve and remove stains. For example, for line segments with dust accumulation, spraying an appropriate amount of detergent can achieve the cleaning effect; but if it is an area with sticky stains, the system will automatically increase the spraying amount to ensure that the stains are completely removed.

[0171] By adjusting the motion trajectory of the wiping head according to the curvature and torsion of the curve segment, the wiping head can closely fit the curve shape of the glass surface. Curvature reflects the bending degree of the curve, and torsion describes the degree of distortion of the curve in three-dimensional space. Through the motion control algorithm, the wiping head can accurately move along the actual shape of the curve to ensure that the entire curve segment can be cleaned, avoiding the situation where stains are missed due to the inability to fit the curve. For example, for a curved segment of special-shaped glass with a complex twisted shape, by accurately considering the curvature and torsion, the wiping head can move along the curve as accurately as along a track, and can achieve close contact regardless of whether it is a convex curve or a concave curve, thereby effectively cleaning the stains on the curved surface.

[0172] Adjusting the speed of the wiper head according to the curvature and torsion of the curve segment can ensure that good cleaning results can be achieved in curves of different degrees of curvature. In curves with large curvature or high torsion, the speed is appropriately slowed down so that the wiper head has enough time and suitable conditions to clean the stains. Because in these complex curve areas, stains may be more difficult to remove and the wiper head requires more precise motion control. For example, in a glass curve segment with a sharp bend, slowing down the speed of the wiper head can prevent the wiper head from leaving the curve surface or failing to effectively clean the stains due to excessive speed. This speed adjustment helps to improve the cleaning quality and ensure that the same cleaning standards are achieved in every part of the curve segment. Adjusting the angle of the wiper head can make it better fit the tangent direction of the curve segment, ensuring that the contact angle between the wiper head and the glass surface is always in the best state. On the curve segment, the appropriate contact angle can make the wiping pressure evenly distributed, thereby removing stains more effectively. For example, for a concave curve segment, the angle of the wiper head will be adjusted inward so that the wiper head can go deep into the curve for cleaning; while for a convex curve segment, the angle is adjusted outward. This dynamic adjustment of the angle can avoid incomplete cleaning or damage to the glass surface due to improper angle, and also improves the stability of the wiping head in the curved section.

[0173] Furthermore, in order to make the wiping head move smoothly on the curve segments with different curvatures and torsion, and to consider factors such as cleaning effect and equipment operation safety, a speed scaling factor is introduced. The speed scaling factor of the wiping head is expressed as: .in, , is the weight parameter used to adjust the curvature and torsion The impact on speed. is the minimum speed. Curvature reflects the bending degree of a curve, and torsion describes the degree of distortion of a curve in three-dimensional space. When the curvature or torsion is large, it means that the shape of the curve changes more dramatically, and the wiping head needs to be more careful when moving on such a curve. , Used to adjust the effect of curvature and torsion on speed. When the curvature or torsion is large, or The value of will increase, so that the speed scaling factor This can prevent the wiping head from moving unsteadily or deviating from the curve track due to too fast a curve change. It is used to ensure that the wiping head has a minimum movement speed under any circumstances to ensure that the cleaning process can continue.

[0174] When curvature and torsion are small, the velocity scaling factor When the curvature or torsion is large, the speed scaling factor The wiping head will become smaller and the speed of the wiping head will be reduced accordingly, ensuring that the wiping head can move smoothly along the curve, ensuring the cleaning effect and safe operation of the equipment.

[0175] In the embodiment of the present application, for the glass sample to be detected, the glass surface information corresponding to the glass sample is collected; the glass surface information includes at least the glass surface images under multiple light source systems; the glass surface information is input into the abnormality detection model to perform abnormality detection on the glass sample; if the glass sample has abnormal scratches, and the abnormal scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped; if the glass sample does not have abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, the cleaning strategy corresponding to the glass sample is generated through the cleaning planning model; based on the cleaning strategy, the wiping head and the detergent spraying system of the wiping robot are controlled to execute the cleaning process for the glass sample. The embodiment of the present application can automatically control the wiping robot to perform cleaning, operate according to precise strategies, reduce interference from human factors, and ensure cleaning consistency, accuracy, stability and reliability.

[0176] Based on the same implementation principle, the embodiment of the present application also provides an electronic device, Figure 3 is a structural block diagram of the electronic device 300, such as Figure 3As shown, the electronic device 300 includes: a processor 301, a memory 302, a communication interface 303, a communication bus 304 and a controller 305; wherein the processor 301, the memory 302, and the communication interface 303 communicate with each other via the communication bus 304; the memory 302 is used to store computer programs; the processor 301 is used to execute the programs stored in the memory 302 to implement corresponding processing functions; the communication bus 304 is used for communication between the electronic device 300 and other devices; and the controller 305 is used to implement the glass cleaning method described in the above method embodiment.

[0177] In the embodiment of the present application, the communication bus 304 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 304 may be divided into an address bus, a data bus, a control bus, etc., and the specific form is not limited. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0178] The memory 302 may include a random access memory (RAM) or a non-volatile memory (non-volatile memory), such as at least one disk memory. Optionally, the memory may also be at least one storage device located far away from the processor 301 .

[0179] The processor 301 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other editable logic devices, discrete gate or transistor logic devices, discrete hardware components, and the specific form can be determined according to actual needs.

[0180] Accordingly, an embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed, can implement each step that can be executed by an electronic device in the above method embodiment.

[0181] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A glass cleaning method, characterized in that: The method is applied to a wiping robot, which at least comprises: a robot body, a mechanical arm, a visual detection system, a wiping head, a detergent spraying system, a base, and a main control computer. The visual detection system comprises a laser scanning module, and the method is executed based on the main control computer. The method comprises: For a glass sample to be tested, collecting glass surface information corresponding to the glass sample; the glass surface information at least includes glass surface images under multiple light source systems; the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively; Inputting the glass surface information into an anomaly detection model to perform anomaly detection on the glass sample; If the glass sample has abnormal scratches, and the abnormal scratches exceed the set damage threshold, the cleaning process of the glass sample is stopped; If the glass sample has no abnormal scratches, or the abnormal scratches do not exceed the set damage threshold, a cleaning strategy corresponding to the glass sample is generated through a cleaning planning model; wherein the cleaning planning model includes first, second and third planning prediction branches, each planning prediction branch is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample; the cleaning strategy includes at least: the motion trajectory of the wiping head, the motion speed of the wiping head, the angle between the wiping head and the glass surface, the amount of detergent sprayed, and the spraying angle; Controlling a wiping head and a cleaning agent spraying system of a wiping robot based on the cleaning strategy to execute a cleaning process for the glass sample; The step of generating a cleaning strategy corresponding to the glass sample by using a cleaning planning model includes: Performing a comprehensive scan on the glass sample by using a laser scanning module in a visual inspection system to obtain depth feature information of the glass sample; Based on the depth feature information, construct a three-dimensional point cloud model of the glass sample; The planning prediction branch in the cleaning planning model that matches the three-dimensional point cloud model is used to generate a cleaning strategy corresponding to the glass sample; wherein the planning prediction branch matches the shape and size parameters of the glass sample in the three-dimensional point cloud model; Wherein, the use of the planning prediction branch in the cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes: If the glass sample shape of the three-dimensional point cloud model is irregular, segmenting the contour of the three-dimensional point cloud model into line segments and curve segments through a third planning prediction branch; Predicting a cleaning sequence corresponding to the three-dimensional point cloud model; Performing wiping operations along the line segments and the curve segments in sequence based on the cleaning sequence; Wherein, the wiping operation is performed sequentially along the line segments and the curve segments based on the cleaning sequence, including: When wiping a curve segment, the motion trajectory, speed and angle of the wiping head are adjusted by a motion control algorithm according to the curvature and torsion of the curve segment; wherein the speed scaling factor of the wiping head is expressed as: ;in, , is the weight parameter used to adjust the curvature and torsion The impact on speed, is the minimum speed.

2. The method according to claim 1, characterized in that Before collecting the glass surface information corresponding to the glass sample to be tested, the method further includes: For each newly added glass sample, identifying the shape of the glass sample; Setting a shooting mode of the glass sample based on the shape of the glass sample, and fine-tuning shooting parameters in the shooting mode based on size information of the glass sample; Among them, the shapes of glass samples include rectangular, circular, and irregular; for rectangular glass samples, the shooting angles in the shooting mode are at least two vertical directions, and the at least two vertical directions include the long side direction and the short side direction; for circular glass samples, the shooting positions in the shooting mode include at least a plurality of shooting points distributed at equal angles around the center of the circle; for irregular glass samples, the shooting angles in the shooting mode are angles that cover the glass surface to the greatest extent and include key parts, and the key parts include at least areas with large curvature changes.

3. The method according to claim 2, characterized in that After setting the shooting mode of the glass sample based on the shape of the glass sample, the method further includes: Setting a plurality of light source systems matching the glass samples; Configuring a specific irradiation angle between a light source in a plurality of light source systems and a glass surface of the glass sample; wherein the specific irradiation angle ranges from 30° to 60°; Based on the environment in which the glass sample is located, the illumination intensity is set to match the background light intensity of the environment; wherein the illumination intensity is inversely proportional to the background light intensity.

4. The method according to claim 1, characterized in that: The step of inputting the glass surface information into an abnormality detection model and performing abnormality detection on the glass sample comprises: Performing feature space projection on the glass surface information to obtain a feature space image corresponding to the glass surface information; extracting candidate abnormal features in the feature space image; Based on the attribute information of the glass sample, a matching standard anomaly database is selected; the attribute information at least includes: glass material type, shooting environment information, and glass material characteristics; Comparing the candidate abnormal feature with the standard abnormal feature preset in the standard abnormal database to obtain a comparison evaluation score corresponding to the candidate abnormal feature; If the comparison evaluation score reaches a set threshold, it is determined that the candidate abnormal feature belongs to an abnormal scratch.

5. The method according to claim 1, characterized in that The step of using a planning prediction branch in a cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes: If the shape of the glass sample in the three-dimensional point cloud model is a rectangle, obtaining any corner point in the three-dimensional point cloud model through the first planning prediction branch; Taking the corner point as the starting point, the cleaning strategy is set as follows: moving in a straight line at a constant speed along the long side of the three-dimensional point cloud model, and using adaptive wiping pressure and detergent spraying amount to perform wiping operations; When moving to the end point of the long side, turning and moving in a straight line along the short side direction of the three-dimensional point cloud model, and using adaptive wiping pressure and detergent spraying amount to perform wiping operations; When the movement reaches the end point of the short side, repeat the steps of moving in a straight line at a constant speed along the long side direction of the three-dimensional point cloud model and performing the wiping operation with an adaptive wiping pressure and a spraying amount of the cleaning agent until the cleaning process of the glass sample is completed; The wiping pressure and the amount of detergent sprayed are adaptively configured based on the degree of dirt on the glass surface detected in real time during movement.

6. The method according to claim 1, characterized in that The step of using a planning prediction branch in a cleaning planning model that matches the three-dimensional point cloud model to generate a cleaning strategy corresponding to the glass sample includes: If the shape of the glass sample of the three-dimensional point cloud model is circular, determining the center position of the circle of the three-dimensional point cloud model through the second planning prediction branch; Taking the center of the circle as the starting point, the cleaning strategy is set as follows: wiping is performed outward along the path of the Archimedean spiral; During the wiping process, the wiping pressure is adjusted according to the radius length, and the amount of detergent sprayed is increased according to a preset functional relationship with the radius length.

7. The method according to claim 1, characterized in that The wiping operation is performed sequentially along the line segments and the curve segments based on the cleaning sequence, comprising: When wiping a line segment, it moves in a straight line at a constant speed, and uses adaptive wiping pressure and detergent spraying amount to perform straight line wiping operations.

8. A glass cleaning device, characterized in that: Used to implement the glass cleaning method according to any one of claims 1 to 7, the glass cleaning device is applied to a wiping robot, the wiping robot at least comprises: a robot body, a mechanical arm, a visual detection system, a wiping head, a detergent spraying system, a base, and a main control computer, the visual detection system comprises a laser scanning module, and the glass cleaning device at least comprises the following units, wherein A collection unit, for collecting glass surface information corresponding to a glass sample to be detected; the glass surface information at least includes glass surface images under multiple light source systems; the multiple light source systems correspond to multiple specific illumination angles and multiple illumination intensities, respectively; A detection unit, used for inputting the glass surface information into an abnormality detection model to perform abnormality detection on the glass sample; A stopping unit, used for stopping the cleaning process of the glass sample if the glass sample has abnormal scratches and the abnormal scratches exceed a set damage threshold; A generation unit, for generating a cleaning strategy corresponding to the glass sample through a cleaning planning model if the glass sample has no abnormal scratches or the abnormal scratches do not exceed a set damage threshold; wherein the cleaning planning model includes first, second and third planning prediction branches, each planning prediction branch is used to adaptively construct a corresponding cleaning strategy based on the shape of the glass sample; the cleaning strategy includes at least: a motion trajectory of the wiping head, a motion speed of the wiping head, an angle between the wiping head and the glass surface, a spraying amount of the detergent, and a spraying angle; An execution unit is used to control a wiping head and a detergent spraying system of a wiping robot based on the cleaning strategy to execute a cleaning process for the glass sample.

Images