Solar energy dispensing machine automatic control system and method based on image recognition

Abstract

This invention specifically relates to an automated control system and method for solar photovoltaic dispensing machines based on image recognition, belonging to the field of solar photovoltaic manufacturing technology. It includes: an image acquisition and processing module; a motion control module; a dispensing execution and control module; a software and user interface module; and an auxiliary and safety module. In this invention, the image acquisition and processing module employs a high-resolution CCD / CMOS camera, a telecentric lens, and a synchronized LED light source, combined with algorithms such as Gaussian filtering, template matching, and U-Net semantic segmentation to locate subtle features such as Mark points and edges on the lamp housing. Simultaneously, the multi-axis closed-loop motion control system is linked with a height-following system, using laser ranging to correct the dispensing head height in real time, adapting to component warping issues. Combined with precise glue supply and constant temperature control, this ensures uniform glue path width and height, reducing defects such as glue path misalignment, glue breakage, and glue overflow, thereby improving the photovoltaic module dispensing pass rate and product reliability.

Description

Technical Field

[0001] This invention relates to the field of solar photovoltaic manufacturing technology, and in particular to an automated control system and method for solar dispensing machines based on image recognition. Background Technology

[0002] In the production of lamp housings, the dispensing process is a key step in ensuring the sealing performance, structural stability, and electrical performance of the components. It is widely used in scenarios such as frame sealing, busbar fixing, and silicon wafer defect repair. Before dispensing, the leads of solar silicon wafers need to be bent to accurately assemble the bent silicon wafers into the lamp housing. This lays a good material foundation for the subsequent automated dispensing process of solar dispensing machine based on image recognition to carry out precise image positioning and automated dispensing without human teaching. The automated continuous operation of solar silicon wafers, from lead wire bending to adhesive dispensing, significantly reduces manual intervention and improves the overall automation, intelligence, and production efficiency of solar photovoltaic module production. After bending, the wafers are adapted to the corresponding lamps, and then the solar silicon wafers are automatically assembled onto the lamp housing by intelligent mechanical equipment, which facilitates the subsequent automatic adhesive dispensing operation.

[0003] However, the lack of coordination between the motion control and dispensing execution system of existing dispensing equipment, and the absence of adaptive adjustment capabilities, is the core bottleneck restricting dispensing quality. Multi-axis motion control often adopts an open-loop control mode, which is not deeply linked with the vision recognition system. The matching accuracy between the motion trajectory and the area to be glued is low, and it cannot cope with the problems of warping and uneven surfaces of large-sized components, resulting in uneven glue path width and height. The glue supply system lacks precise metering, degassing, and constant temperature control, resulting in large fluctuations in glue viscosity and a tendency for glue breakage and overflow. In addition, the height of the dispensing head cannot keep up with changes in the board surface in real time, further exacerbating dispensing defects and making it difficult to meet the requirements of high-precision dispensing processes.

[0004] Therefore, an automated control system and method for solar dispensing machines based on image recognition are proposed to address the aforementioned problems. Summary of the Invention

[0005] The purpose of this invention is to provide an automated control system and method for solar dispensing machines based on image recognition in order to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: An automated control system for a solar dispensing machine based on image recognition includes: The image acquisition and processing module is configured to transport the lamp housings to be glued in batches to the feeding station of the production line to provide basic materials for subsequent processes, identify the front and back orientation of the lamp housings, screen and remove the reversed housings and put them back in place, acquire images of the area of ​​the lamp housing to be glued, and convert the images into coordinate data and control commands through image preprocessing, feature extraction and positioning and pattern recognition algorithms. The motion control module is configured to receive instructions and precisely control the movement of the dispensing head and the lamp housing through a multi-axis motion control system and a transmission and positioning mechanism. It will precisely arrange the housings after forward and reverse screening at equal intervals on the transmission platform to ensure the accuracy of subsequent photo-taking and hole spacing measurement, plasma processing, and dispensing positioning, and ensure the matching of dispensing trajectory. The dispensing execution and control module is configured to output glue quantitatively and at a constant speed through a dispensing valve, glue supply system, and height tracking system, ensuring consistent glue path width and height. This includes: detecting the hole spacing and hole coordinates in the glue-to-be-dispensed area of ​​the lamp housing, providing a precise physical coordinate reference for subsequent dispensing trajectory planning; performing plasma cleaning on the glue-to-be-dispensed surface of the housing to remove surface impurities; and achieving quantitative, uniform, and high-precision dispensing on the glue-to-be-dispensed surface of the housing to ensure consistent glue path width and height. The software and user interface module is configured as a human-computer interaction window, enabling system integration control, data traceability, and visual operation through visual guidance software, CAD trajectory generation, database, and MES integration.

[0007] Preferably, the image acquisition in the image acquisition and processing module specifically includes: High-resolution CCD or CMOS cameras are used, and the selection should meet the size requirements of the lamp housing and the accuracy requirements of the features to be identified. Telecentric lenses are selected for optical applications; industrial fixed-focus lenses are selected for wide-field-of-view scenarios. High-brightness LED light source is used; the light source style is selected according to the type of feature to be identified. The controller uses a constant current source controller; the light source is installed coaxially with the lens or around the lens, and is adjusted according to the lens focal length and field of view to ensure that the lighting area completely covers the shooting area without shadows or uneven light spots.

[0008] Preferably, the step of converting the image into coordinate data and control commands through image preprocessing, feature extraction and localization, and pattern recognition algorithms specifically includes: The original image is subjected to noise reduction, contrast enhancement, and geometric correction to eliminate interference and ensure the clarity of the features to be identified. A threshold segmentation algorithm is used to divide the image into foreground and background; In the preprocessed image, the reference features and the features to be applied to the lamp housing are extracted, and their actual physical coordinates are calculated to provide a position reference for the motion control module. Using the Mark point coordinates as a reference, the detected edge coordinates are converted into actual physical coordinates, and the relative position between the area to be dispensed and the edge is calculated to ensure that the deviation between the dispensing trajectory and the edge meets the preset requirements. Identify the specific area to be glued and judge the glue dispensing quality.

[0009] Preferably, the motion control module specifically includes: An AC servo motor is selected, and the X / Y axes use ball screw drive, with the screw connected to the servo motor via a coupling; the Z axis uses a ball screw or linear guide rail to accommodate the lifting and lowering motion of the dispensing head. It adopts an industrial-grade multi-axis motion controller with an interface to receive coordinate commands from the image acquisition and processing module in real time and output control signals to the servo motor. The conveying and positioning mechanism is responsible for transporting the lamp housing from the feeding station to the dispensing station, completing the positioning and fixing, preventing the lamp housing from shifting during the dispensing process, and ensuring dispensing accuracy.

[0010] Preferably, the dispensing valve and dispensing system in the dispensing execution and control module specifically include: The dispensing valve is selected based on the type and viscosity of the adhesive and the requirements of the dispensing process. It controls the opening, closing, and output of the adhesive to ensure dispensing accuracy. Specific parameters are as follows: Screw valve: suitable for high viscosity adhesives; spray valve: suitable for medium and low viscosity adhesives; impact valve: suitable for low viscosity adhesives; The glue supply system is responsible for the storage, transportation, degassing, and temperature control of the glue to ensure stable glue performance and continuous glue dispensing. Its components include glue tanks, pressure tanks, pipelines, glue volume controllers, degassing mechanisms, and temperature control mechanisms. The adhesive enters the pressure tank from the glue bucket. After degassing and constant temperature treatment, it is measured by the glue quantity controller and delivered to the dispensing valve through pipeline. The glue supply pressure and dispensing volume are fed back to the control system in real time and dynamically adjusted according to the dispensing process requirements.

[0011] Preferably, the height-following system includes: A laser rangefinder sensor is used and installed on the side of the dispensing head, with the laser direction perpendicular to the surface of the lamp housing. The laser rangefinder sensor collects the distance between the needle tip and the lamp housing in real time, transmits the data to the motion controller, compares it with the preset distance, calculates the height deviation, controls the Z-axis servo motor to drive the dispensing head to rise and fall, and corrects the height deviation in real time.

[0012] Preferably, the operations that need to be performed after the dispensing is completed include: Pre-curing treatment is performed on the adhesive layer after dispensing the adhesive onto the housing to shape the surface of the adhesive layer, thereby preventing deformation and sticking of the adhesive path during subsequent silicon wafer assembly and improving the initial bonding strength of the adhesive layer. The solar silicon wafers are precisely arranged in batches and transported to the wafer assembly station, where they are precisely docked with the housings that have been glued and cured, thus achieving automated assembly of the silicon wafers and the housings. Based on the silicon wafer model and the dimensions of the housing assembly position, adjust the position and gripping angle of the silicon wafer picking robot module, and at the same time measure the actual length of the silicon wafer leads to determine whether it meets the assembly process requirements. The qualified solar silicon wafers are precisely picked up and transferred to the adhesive application point on the pre-cured casing to achieve initial bonding between the silicon wafers and the casing.

[0013] Preferably, the method further includes: Based on the dimensions of the housing perforation and assembly requirements, the silicon wafer leads are precisely bent multiple times to ensure that the formed leads fit the housing perforation and subsequent electrical connection requirements. The silicon wafer leads, after being bent multiple times, are precisely inserted into the pre-set holes in the lamp housing to achieve mechanical fixation between the silicon wafer and the housing; The housing, after silicon wafer assembly and lead wire perforation, is precisely clamped and flipped to a preset angle so that the back of the housing faces upward, providing an operating surface for subsequent hole plugging and lead wire bending processes. The perforations in the housing are sealed with adhesive, and the leads that penetrate the back of the housing are bent again to make the lead shape adapt to the subsequent electrical assembly requirements, thus completing the initial forming of the photovoltaic module.

[0014] An automated control method for solar dispensing machines based on image recognition includes: The image of the area to be glued on the lamp housing is acquired, processed, and converted into coordinate data and control commands to complete pre-glue positioning and post-glue detection. Through a multi-axis motion control system and a transmission and positioning mechanism, instructions are received and the movement of the dispensing head and the lamp housing are controlled to achieve dispensing trajectory matching. By utilizing a dispensing valve, a glue supply system, and a height tracking system, the glue can be dispensed quantitatively and at a constant speed, maintaining the width and height of the glue path. By integrating visual guidance software, CAD trajectory generation, database and MES, the system achieves integrated control, data traceability and visual operation; By leveraging fault alarm self-diagnosis and cleaning / curing mechanisms, the system ensures stable operation, equipment and personnel safety, and reduces downtime due to malfunctions.

[0015] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: 1. This invention utilizes an image acquisition and processing module, employing a high-resolution CCD / CMOS camera, a telecentric lens, and a synchronized LED light source. Combined with algorithms such as Gaussian filtering, template matching, and U-Net semantic segmentation, it achieves the localization of subtle features such as Mark points and edges on the lamp housing. Simultaneously, a multi-axis closed-loop motion control system and a height-following system work together to correct the dispensing head height in real time via laser ranging, adapting to component warping issues. Combined with precise glue supply and constant temperature control, this ensures uniform glue path width and height, reducing defects such as glue path misalignment, glue breakage, and glue overflow, thereby improving the glue dispensing qualification rate and product reliability of photovoltaic modules.

[0016] 2. This invention utilizes visual guidance software and CAD trajectory import functionality to automatically extract and correct the dispensing trajectory, eliminating the need for manual instruction. Combined with MES system integration and database management, it enables full traceability of production process parameters and test results, reducing manual intervention. Simultaneously, it integrates comprehensive fault self-diagnosis, automatic needle cleaning, curing treatment, and safety protection mechanisms. This allows for real-time monitoring of equipment operation, automatic fault diagnosis, and solution suggestions, minimizing downtime and maintenance costs. The visual human-machine interface simplifies the operation process, lowers the barrier to entry for operators, and achieves closed-loop control throughout the entire process—from pre-dispensing positioning and adaptive control during dispensing to post-dispensing quality inspection—improving production efficiency and facilitating the intelligent and large-scale upgrading of the photovoltaic industry. Attached Figure Description

[0017] Further details, features, and advantages of this application are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which: Figure 1 This is a system structure diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention; Figure 3 This is a product image of a solar lamp housing processed by the dispensing machine in this invention. Detailed Implementation

[0018] Several embodiments of this application will now be described in more detail with reference to the accompanying drawings to enable those skilled in the art to implement this application. This application may be embodied in many different forms and for various purposes and should not be limited to the embodiments set forth herein. These embodiments are provided to make this application thorough and complete, and to fully convey the scope of this application to those skilled in the art. The embodiments described do not limit this application.

[0019] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It will be further understood that terms such as those defined in commonly used dictionaries shall be interpreted as having a meaning consistent with their meaning in the relevant field and / or the context of this specification, and shall not be interpreted in an idealized or overly formal sense unless expressly defined herein.

[0020] Example 1

[0021] Its specific implementation method is combined with the appendix Figure 1 and attached Figure 3 Please provide a detailed explanation.

[0022] Appendix Figure 1 The block diagram of the automated control system for a solar dispensing machine based on image recognition provided in this embodiment of the invention shows the connection relationship between the image acquisition and processing module and the auxiliary and safety module, and marks the main functional interaction flow of each module.

[0023] Appendix Figure 2 The flowchart of the automated control method for a solar dispensing machine based on image recognition provided in this embodiment of the invention shows the complete steps from acquiring an image of the area to be dispensed on the lamp housing to achieving quantitative and uniform glue output while maintaining consistent glue path width and height.

[0024] In this embodiment, it includes: The image acquisition and processing module is configured as the system's visual perception unit. It uses an industrial camera, optical lens, LED light source, and controller to transport the lamp housings to be glued in batches to the production line's feeding station, providing basic materials for subsequent processes. It identifies the front and back orientation of the lamp housings, filters and removes reversed housings, and returns them to their original positions. It acquires high-quality images of the area of ​​the lamp housing to be glued, providing a basis for subsequent image processing and recognition. Through image preprocessing, feature extraction and localization, and pattern recognition algorithms, it converts the images into coordinate data and control commands, realizing pre-glue positioning and post-glue detection. Specifically, it includes: Industrial cameras: High-resolution CCD (charge-coupled device) or CMOS (complementary metal-oxide-semiconductor) cameras are used, and the selection should meet the size requirements of the lamp housing and the accuracy requirements of the features to be identified. The specific parameters and selection criteria are as follows: Resolution: Determined based on the minimum feature size of the area to be glued. Typically, 12-24 megapixels are selected, with a pixel size ≤3.2μm, to ensure that fine features such as the edge of the battery cell, grid lines, and Mark points (reference points) can be clearly captured. In the case of glue dispensing to repair fine defects, a camera with more than 24 megapixels is required. Frame rate: ≥30fps, ensuring real-time capture of dynamic images of the lamp housing during the conveying process, avoiding image blurring or missed shots due to excessive conveying speed, and adapting to the conveying speed requirements of production lines of 2~5m / min; Interface type: GigEVision or USB3Vision interface, with a transmission rate of ≥1000Mbps, to achieve high-speed, zero-latency transmission of image data and ensure real-time collaboration with the image processing module; Installation method: Use a gantry beam for suspension or side mounting. The camera optical axis should be perpendicular to the lamp housing panel. The installation height should be adjusted according to the lens focal length to ensure that the shooting field of view completely covers the area to be glued. Redundant configuration: For large-size photovoltaic modules (such as 166mm, 182mm, and 210mm modules), 2-4 cameras are configured for stitching and shooting, with a stitching accuracy of ≤0.01mm. The image stitching algorithm achieves full coverage of the panel, avoiding the problem of insufficient field of view of a single camera; the variable zoom lens is adapted to different sizes of lamp housings, with a focal length adjustment range of 8~50mm, supporting manual or automatic focus, and a focus accuracy of ≤0.005mm.

[0025] Optical lens: The selection of the model is based on the system's working distance, shooting field of view, and accuracy requirements. Its core function is to clearly image the features to be identified on the lamp housing onto the camera's image sensor, reducing imaging errors. Specific parameters are as follows: Lens type: Telecentric lenses (object-side telecentric or double telecentric) are preferred. Their object-side telecentric design can effectively reduce perspective error and ensure that the imaging ratio of the features to be identified (such as the edge of the battery cell, Mark point) at different positions is consistent, with a measurement accuracy of ≤0.003mm. For wide field of view scenes, industrial fixed-focus lenses with a focal length of 12~25mm and a field of view of 30°~60° are selected. Aperture range: F1.4~F16, which can be automatically adjusted according to the lighting intensity to ensure image sharpness while avoiding overexposure or underexposure; Distortion rate: ≤0.1%, reducing image geometric distortion caused by lens distortion and avoiding deviations in subsequent feature extraction and localization; Interface compatibility: The interface must be compatible with industrial cameras (such as USB-C or USB-C), ensuring a secure installation and preventing image shift caused by looseness.

[0026] Light source and controller: To provide stable and uniform lighting for the shooting area, highlighting the contrast between the features to be identified (such as adhesive trails, cell edges, and adhesive mark points) and the background, and eliminating interference from ambient light (such as workshop fluorescent lights and natural light), the specific selection and control methods are as follows, taking into account the high reflectivity of solar silicon wafers and the transparency of some component surfaces: Light source type: High-brightness LED light source with a lifespan of ≥50,000 hours and power consumption ≤50W, avoiding heat generation that could affect the lamp housing or system components; select the light source style according to the type of feature to be identified. Ring light: used for dispensing mark identification and battery cell edge detection, providing uniform illumination and adjustable angle (0°-45°) to adapt to lamp housing surfaces with different reflective properties; Coaxial light: used for detecting adhesive path traces in transparent or semi-transparent components, reducing surface reflection interference and ensuring clear adhesive path edges; Strip light: Used for identifying linear features such as busbars and grid lines. The direction of illumination can be adjusted to highlight the outline of the linear features.

[0027] Light source parameters: wavelength 450~650nm, brightness adjustment range 0~100%, supports continuous dimming, adjusts brightness according to the surface material of the lamp housing (such as film, bare silicon wafer), and ensures feature-to-background contrast ≥3:1; Controller: It adopts a constant current source controller with an output current of 0~2A and an accuracy of ±1%. It supports linkage with the image acquisition module to realize "camera shooting - light source synchronous turn-on", avoiding energy waste and heat generation caused by the light source being constantly on. It also has overcurrent and overvoltage protection functions to prevent damage to the light source. Installation location: The light source should be installed coaxially with the lens or around the lens, with a distance of 150~300mm from the surface of the lamp housing. Adjust according to the lens focal length and field of view to ensure that the lighting area completely covers the shooting area without shadows, uneven light spots, or other problems.

[0028] Image preprocessing: The original image undergoes noise reduction, contrast enhancement, and geometric correction to eliminate interference from image noise, uneven illumination, and geometric distortion, ensuring the clarity of the features to be identified and laying the foundation for subsequent feature extraction. The specific processing steps and algorithms are as follows: Noise reduction: Gaussian filtering algorithm (Gaussian kernel size 3×3 or 5×5) is used to remove Gaussian noise from the image; for salt-and-pepper noise, median filtering algorithm (filter window 3×3) is used, and the filtering intensity can be automatically adjusted according to the noise level to ensure that the image edges are not blurred after noise reduction. Contrast Enhancement: A histogram equalization algorithm is used to adjust the gray value distribution of the image, mapping the gray value range to 0-255, thereby enhancing the contrast between the features to be identified (such as Mark points and edges) and the background and solving the feature blurring problem caused by uneven lighting. For highly reflective areas, an adaptive contrast enhancement algorithm is used to avoid local overexposure. Geometric correction: Camera calibration is performed based on a calibration board (chessboard calibration board, grid size 10×10mm) to obtain camera intrinsic parameters (focal length, principal point coordinates) and extrinsic parameters (camera attitude). The geometric distortion of the image is corrected by perspective transformation algorithm to ensure that the mapping error between the coordinates of each point in the image and the actual physical coordinates is ≤0.005mm. At the same time, rotation and translation corrections are performed on the image to eliminate image offset caused by the tilt of the lamp housing. Image segmentation: A threshold segmentation algorithm (adaptive threshold or fixed threshold) is used to divide the image into foreground (features to be identified) and background. The threshold range is automatically calculated based on the gray value distribution of the image to ensure complete extraction of foreground features without omission or misextraction.

[0029] Feature extraction and localization: In the preprocessed image, the reference features and adhesive application features of the lamp housing are accurately extracted, and their actual physical coordinates are calculated to provide a position reference for the motion control module. The specific implementation method is as follows: Mark point recognition and localization: Mark point specifications: Pre-set reference points on solar modules, usually circular (diameter 2~5mm) or square (side length 2~5mm), with colors that contrast sharply with the module surface (e.g., black Mark points, white background). Recognition Algorithm: A template matching algorithm is adopted. Standard templates of Mark points are pre-stored (the template size is consistent with the actual Mark points). Matching is performed by normalized cross-correlation coefficient (NCC) with a matching threshold ≥0.8 to ensure accurate recognition even under conditions of light variation and slight contamination. Positioning accuracy: The positioning accuracy of the center coordinate of the Mark point is ≤0.003mm. By identifying 2-4 Mark points, the positional offset of the lamp housing (X, Y direction offset and rotation angle) is calculated. This is used to correct the physical positional deviation of the lamp housing caused by conveyor belt vibration and errors in the previous process. The correction range is ±5mm, and the rotation angle correction range is ±3°.

[0030] Edge detection and localization: Inspection targets: edges of battery cells, grid lines, busbar edges, and edge of frame mounting positions, with an edge width ≥ 0.1mm; Detection algorithm: The Canny edge detection algorithm is adopted, the threshold range is automatically adjusted according to the image contrast, the edge detection accuracy is ≤0.002mm, and the linear edges are extracted by the Hough linear transform algorithm to determine the coordinates and direction of the edges. Positioning logic: Based on the coordinates of the Mark point, the detected edge coordinates are converted into actual physical coordinates, and the relative position of the area to be dispensed and the edge is calculated to ensure that the deviation between the dispensing trajectory and the edge meets the preset requirements (deviation ≤ 0.01mm). Pattern recognition algorithm: Mature and reproducible pattern recognition technology is employed to identify the specific area to be dispensed and determine the dispensing quality, achieving closed-loop control of "pre-dispensing positioning and post-dispensing detection." The specific algorithm selection and implementation are as follows: Identification of the area to be dispensed: Template matching: Suitable for dispensing at fixed locations (such as junction box mounting positions, frame sealant tracks), pre-stores standard templates for the area to be dispensed, matching accuracy ≤0.005mm, recognition speed ≥30 frames / second; Blob analysis: Applicable to irregular glue application areas (such as scratch repair areas). By extracting connected regions (Blob) in the image, analyzing their area, shape, perimeter and other features, and comparing them with preset thresholds, the area to be glued is determined. Deep learning semantic segmentation: suitable for complex scenarios (such as simultaneous repair of multiple defects and dispensing of adhesive to irregularly shaped components). It adopts the U-Net network model, and the training dataset contains images of lamp housings of different types and specifications (sample size ≥ 10,000 images). The segmentation accuracy is ≥ 99%, the recognition speed is ≥ 20 frames / second, and it can automatically identify multiple areas to be dispensed without the need for manual template preset.

[0031] Post-application quality inspection: Adhesive path inspection: By comparing images, the system detects whether there are defects such as glue breakage, misalignment, glue overflow, or insufficient glue amount in the adhesive path. The detection threshold for glue path misalignment is ≤0.01mm, and the minimum length for glue breakage detection is ≥0.5mm. Detection logic: The image after dispensing is compared with the standard glue path image, and the grayscale difference and glue path width deviation are calculated. If they exceed the preset threshold, they are judged as unqualified, an alarm signal is generated and fed back to the control system, and the defect location and type are recorded for subsequent traceability.

[0032] Algorithm operating environment: Uses an industrial-grade processor (such as an Intel Core i7 or an embedded processor with equivalent performance), ≥16GB of memory, ≥128GB of storage to ensure real-time operation of the algorithm without lag or delay, and the processing time for a single frame image is ≤30ms.

[0033] The motion control module is configured as the system execution drive unit. Through the multi-axis motion control system and the transmission and positioning mechanism, it receives instructions and precisely controls the movement of the dispensing head and the lamp housing. It arranges the housings after forward and reverse screening at precise equal intervals on the transmission platform to ensure the positioning accuracy of subsequent processes such as photographing and measuring hole spacing, plasma processing, and dispensing, and to ensure the matching of dispensing trajectory. Specifically, it includes: Multi-axis motion control system: Multi-axis linkage control is adopted to realize the movement of the dispensing head in the X, Y, and Z directions and the rotation in the R direction, adapting to the dispensing needs of lamp housings of different sizes and placement angles. The specific hardware and algorithm are as follows: Hardware configuration: Servo motor: An AC servo motor with a power of 1.5~3kW, a rated speed of 3000r / min, an encoder resolution of ≥2500 lines, a positioning accuracy of ≤0.001mm, and a torque of ≥5N·m is selected to ensure smooth and vibration-free movement; the X / Y axis adopts a linear motor with a positioning accuracy of ≤0.0005mm, which is suitable for high-speed dispensing requirements (movement speed ≥1m / s). Transmission mechanism: The X / Y axes adopt ball screw drive with a lead of 5~10mm, accuracy grade C3, and repeatability ≤0.002mm. The screw and servo motor are connected by a coupling to ensure backlash-free power transmission. The Z axis adopts ball screw or linear guide rail to adapt to the lifting motion of the dispensing head, with a lifting stroke of 50~100mm and a lifting speed of 0.1~0.5m / s. Motion controller: It adopts an industrial-grade multi-axis motion controller (supporting 3~4 axis linkage), with interface type supporting EtherCAT or Profinet, communication rate ≥1000Mbps, real-time reception of coordinate commands from the image acquisition and processing module, output of control signals to the servo motor, and control cycle ≤1ms; Platform structure: adopts a gantry structure with a crossbeam rigidity of ≥1000N / mm to ensure no deformation during movement. The X / Y axis movement range is determined according to the size of the lamp housing (e.g., 2000mm×1500mm), and is compatible with photovoltaic modules of 166mm-210mm specifications.

[0034] Interpolation algorithm: Algorithm type: Linear interpolation and circular interpolation algorithms are adopted, supporting continuous interpolation of multi-segment trajectories with an interpolation accuracy of ≤0.001mm, ensuring smooth dispensing trajectory without inflection points or stuttering; For complex curve dispensing (such as irregular borders), spline interpolation algorithm is adopted with a fitting error of ≤0.002mm. Speed ​​control: S-curve acceleration and deceleration control is adopted, with an acceleration time of 0.1-0.3s to avoid dispensing head jitter caused by sudden changes in movement speed and ensure uniform glue path; the movement speed can be adjusted (0.1-1m / s) according to the dispensing accuracy requirements. 0.1-0.3m / s is selected for high-precision dispensing scenarios, and 0.5-1m / s is selected for high-speed mass production scenarios. Closed-loop control: The motion controller collects the encoder feedback signal of the servo motor in real time, compares it with the target coordinates sent by the image processing module, calculates the motion deviation, and corrects the motor speed and position in real time to ensure that the motion positioning accuracy is ≤0.002mm and the deviation correction response time is ≤10ms.

[0035] Transmission and positioning mechanism: This unit is responsible for transporting the lamp housing from the feeding station to the dispensing station, achieving precise positioning and fixation to prevent displacement of the lamp housing during dispensing and ensuring dispensing accuracy. The specific structure and implementation are as follows: Transmission mechanism: Type: Belt conveyor or roller conveyor. The belt is made of wear-resistant and anti-static rubber with a thickness of 2-3mm and a width that matches the size of the lamp housing (e.g., 1200-2000mm). The roller conveyor uses stainless steel rollers with a diameter of 50-80mm and a spacing of 100-150mm to ensure smooth conveying. Conveying speed: Adjustable, range 0.5-5m / min, linked to dispensing speed to ensure uniform movement of the lamp housing during dispensing (conveying can be stopped in static dispensing scenarios). Positioning sensors: Photoelectric sensors (diffuse reflection or through-beam type) are installed on both sides of the dispensing station to detect whether the lamp housing has reached the designated position. The detection accuracy is ≤0.1mm. The trigger signal is transmitted to the motion controller to control the conveying mechanism to stop running. The positioning response time is ≤50ms.

[0036] Fixed mechanism: Type: Vacuum adsorption or clamping is used, with vacuum adsorption being preferred (to avoid damaging the surface of the lamp housing); the vacuum adsorption mechanism consists of a vacuum suction cup, a vacuum pump, and a vacuum pressure sensor. The suction cup is made of silicone material, with a diameter of 10~20mm, and the number is determined according to the size of the lamp housing (e.g., 10~20), and is evenly distributed on the conveyor platform. Vacuum parameters: Vacuum degree ≤ -0.08MPa, adsorption force ≥ 50N / suction cup, ensuring that the lamp housing is firmly adsorbed and there is no displacement; the vacuum pressure sensor monitors the vacuum degree in real time, and when it is lower than the preset threshold (such as -0.06MPa), an alarm signal is issued to prevent adsorption failure; Clamping: For severely warped lamp housings, pneumatic clamps are provided on both sides with adjustable clamping force (50~100N). The clamping contact surfaces are made of soft materials (such as sponge) to avoid damaging the edges of the lamp housing. The clamping position does not interfere with the dispensing trajectory.

[0037] After the glue dispensing is completed, the following steps are required: Pre-curing treatment is applied to the adhesive layer after dispensing onto the casing to shape the adhesive surface, preventing deformation and sticking of the adhesive path during subsequent silicon wafer assembly, and simultaneously improving the initial bonding strength of the adhesive layer. The system utilizes the original system's curing mechanism, automatically switching between curing methods (UV curing / hot air curing) based on the adhesive type. After dispensing, the housing is conveyed at a uniform speed to the curing station by a transport mechanism, with the curing mechanism linked to the transport speed. For UV adhesives, a 365~405nm wavelength UV curing lamp with a power of 100~300W and a preset curing time of 10~30s is used. For thermosetting adhesives, a hot air gun is used with a temperature of 50~150℃ (accuracy ±1℃), a wind speed of 0.5~2m / s, and a preset curing time of 30~60s. The auxiliary and safety module monitors the curing temperature / power in real time, immediately stopping curing and issuing an alarm when the threshold is exceeded to prevent over-curing and cracking of the adhesive layer.

[0038] The solar silicon wafers are precisely arranged in batches and transported to the wafer assembly station, where they are precisely docked with the housings that have been glued and cured, thus achieving automated assembly of the silicon wafers and the housings. The silicon wafers are transported using an anti-static vacuum adsorption conveyor belt (silicone suction cups, 5~10mm in diameter, vacuum degree ≤-0.07MPa, to avoid damaging the silicon wafer surface). At the arrangement station, the camera (8 megapixels, adapted to the fine features of the silicon wafer) of the image acquisition and processing module identifies the edge and grid line features of the silicon wafer to achieve precise positioning of the silicon wafer (positioning accuracy ≤0.01mm). The motion control module drives the pusher block to arrange the silicon wafers at equal intervals (the spacing matches the dispensing position on the housing). After the arrangement is completed, the silicon wafers are picked up and transported by the robotic arm. The silicon wafer conveying speed is linked to the housing conveying speed (0.5~2m / min) to ensure that the silicon wafers and the housing arrive at the assembly station synchronously.

[0039] Based on the silicon wafer model and the dimensions of the housing assembly position, the position and gripping angle of the silicon wafer picking robot module are precisely adjusted. At the same time, the actual length of the silicon wafer lead is measured to determine whether it meets the assembly process requirements, providing parameter basis for subsequent lead bending and perforation. The software and user interface module automatically generates the gripping coordinates and angle parameters of the silicon wafer picking robot based on the imported silicon wafer and housing CAD drawings. The motion control module drives the robot module to complete automatic adjustment (adjustment accuracy ≤0.002mm), and manual fine-tuning is also possible on the interface. The image acquisition and processing module uses a strip LED light source combined with an industrial fixed-focus lens to photograph the silicon wafer leads. The lead contour is extracted through the Blob analysis algorithm, the actual lead length is calculated, and compared with the process preset length. Silicon wafers with a length deviation exceeding ±0.5mm are automatically rejected.

[0040] The qualified solar silicon wafers are precisely picked up and transferred to the adhesive application point on the housing that has been glued and cured, so as to achieve the initial bonding between the silicon wafers and the housing. A four-axis vacuum adsorption robot (AC servo motor driven, encoder resolution ≥2500 lines, positioning accuracy ≤0.001mm) is used. Silicone suction cups are evenly distributed at the end of the robot (to avoid damaging the silicon wafer), with an adsorption force ≥20N / cube. The image acquisition and processing module simultaneously acquires the coordinates of the dispensing position on the housing and the silicon wafer, calculates the assembly offset, and transmits it to the motion control module. The robot accurately picks up the silicon wafer and moves it to the dispensing position on the housing according to the coordinate instructions, achieving coaxial assembly of the silicon wafer and the housing with an assembly deviation ≤0.01mm. The auxiliary and safety module monitors the vacuum adsorption pressure in real time. If the pressure is insufficient, the silicon wafer is immediately released and an alarm is triggered to prevent the silicon wafer from falling and breaking.

[0041] Based on the dimensions of the perforation in the housing and the assembly requirements, the silicon wafer leads are precisely bent multiple times to ensure that the formed leads are compatible with the housing perforations and subsequent electrical connection requirements, thus replacing manual bending and improving bending accuracy and consistency. The motion control module is equipped with a dedicated lead wire bending module (composed of a servo motor and a precision bending fixture, with a bending angle adjustment accuracy of ≤0.5°). The image acquisition and processing module identifies the starting coordinates of the silicon wafer lead wire bending and transmits the data to the bending module. The module automatically completes the bending according to the number of bends, bending angle, and bending position preset in the software interface (2 to 5 bends can be preset, with an angle of 0 to 90°). During the bending process, the laser sensor detects the lead wire bending formation status in real time, and automatically re-bends if the standard is not met.

[0042] The silicon wafer leads, after being bent multiple times, are precisely inserted into the pre-set holes in the lamp housing to achieve mechanical fixation between the silicon wafer and the housing, laying the foundation for subsequent secondary bending of the leads and electrical connection. The image acquisition and processing module simultaneously acquires the coordinates of the housing perforation and the bent lead wire using a camera, calculates the offset between the perforation center and the lead wire, and transmits the corrected coordinate commands to the motion control module. The lead wire perforation module is driven by a servo motor to push the lead wire precisely into the housing perforation (push pin positioning accuracy ≤ 0.002mm). For multi-lead silicon wafers, a multi-axis perforation module is used for synchronous operation, and the perforation speed is matched with the lead wire specifications. After perforation is completed, the image acquisition and processing module detects the perforation status of the lead wire. If it is not perforated, the operation is automatically restarted. If it still does not meet the standard, it is rejected.

[0043] The housing, after silicon wafer assembly and lead wire perforation, is precisely clamped and flipped to a preset angle (usually 180°) so that the back of the housing faces upward, providing an operating surface for subsequent hole plugging and lead wire re-bending processes.

[0044] A five-axis heavy-duty vacuum adsorption robot (adapted to the overall weight of the housing and silicon wafer, with an adsorption force ≥100N) is used to grip the non-adhesive / non-assembly areas of the housing, avoiding damage to the adhesive layer and silicon wafer. The motion control module drives the robot to complete the rotation according to the preset rotation angle (0~180° adjustable, accuracy ≤0.1°). During the rotation, S-curve acceleration and deceleration control is used to avoid sudden changes in rotation speed that could cause the silicon wafer to fall off or the adhesive layer to crack. The auxiliary and safety module monitors the gripping pressure and rotation angle of the robot in real time. If any deviation occurs, the operation will stop immediately and an alarm will be triggered.

[0045] The perforations in the housing are sealed with glue to improve the waterproof and dustproof performance of the module. At the same time, the leads that penetrate into the back of the housing are bent again to make the lead shape adapt to the subsequent electrical assembly requirements, thus completing the initial forming of the photovoltaic module. The plugging process utilizes the original system's dispensing execution and control module. Depending on the plugging process requirements, either a needle valve (for low-viscosity plugging adhesive) or a screw valve (for high-viscosity sealant) is selected. After the adhesive supply system completes degassing and temperature control, the image acquisition and processing module precisely locates the perforation position. The dispensing head then performs circular dispensing along the edge of the perforation position, with the adhesive layer height and width preset according to the process (accuracy ≤ ±5%). After plugging, a miniature curing lamp rapidly pre-cures the plugging adhesive layer (curing time 5~10s) to prevent adhesive flow.

[0046] The lead wire is bent again using the lead wire bending module. According to the electrical assembly requirements on the back of the housing, the bending angle (0~180°) and bending position are preset in the software interface. The motion control module drives the bending module to complete the lead wire bending again. During the bending process, the laser sensor detects the forming status in real time to ensure that the lead wire is accurately connected to the electrical assembly position on the back of the housing. The dispensing execution and control module, configured as the core unit of the system process, delivers quantitative and uniform glue output through a dispensing valve, glue supply system, and height tracking system, ensuring consistent glue path width and height. It detects the hole spacing and coordinates in the glue-dispensing area of ​​the lamp housing, providing a precise physical coordinate reference for subsequent dispensing trajectory planning, replacing manual measurement and improving dispensing positioning accuracy. It also performs plasma cleaning on the glue-dispensing surface of the housing to remove surface oil, dust, and other impurities, improving the adhesion strength between the glue and the housing and preventing defects such as glue detachment and glue layer cracking after dispensing. It achieves quantitative, uniform, and high-precision dispensing on the glue-dispensing surface of the housing, ensuring consistent glue path width and height, providing a reliable glue layer foundation for subsequent silicon wafer bonding. Specifically, it includes: Dispensing valve: The selection of the model should be based on the type of adhesive, viscosity, and dispensing process requirements. Its core function is to control the opening, closing, and output of the adhesive to ensure dispensing accuracy. Specific parameters and selection criteria are as follows: Type selection: Screw valve: Suitable for high viscosity adhesives (viscosity 10000~100000mPa·s), such as sealing silicone and thermal conductive adhesive. It uses screw rotation to push the adhesive, with dispensing accuracy ≤±1%, adhesive width adjustment range 0.5-5mm, adhesive height adjustment range 0.1~2mm, suitable for long strip dispensing (such as frame sealing). Spray valve: Suitable for medium and low viscosity adhesives (viscosity 1000~10000mPa·s), such as manifold fixing adhesive. It uses high pressure spraying to dispense adhesive, with a dispensing speed ≥100 dots / second. It is non-contact dispensing to avoid damage to the surface of the lamp housing. The dispensing accuracy is ≤±2%, which is suitable for high-speed and high-precision dispensing. Impact valve: Suitable for low viscosity adhesives (viscosity 100~1000mPa·s), such as scratch repair adhesive. It has a simple structure, fast response speed (≤10ms), dispensing accuracy ≤±3%, and adjustable glue dot diameter (0.1-1mm).

[0047] Key parameters: Response time: ≤10ms, ensuring linkage with the motion control module to achieve "move to the designated position - immediately dispense glue", avoiding glue path deviation; Dispensing pressure: 0.1~0.8MPa, adjustable according to adhesive viscosity, with an adjustment accuracy of ±0.01MPa; Needle specifications: Stainless steel needles are used, with an inner diameter of 0.1-1mm and a length of 10-50mm. The needles are threaded to the dispensing valve and can be quickly replaced to accommodate different glue width and height requirements. Temperature resistance: -10℃ to 80℃, suitable for the constant temperature control requirements of adhesives, avoiding valve blockage caused by adhesive curing.

[0048] Adhesive supply system: Responsible for the storage, transportation, degassing, and temperature control of the adhesive to ensure stable adhesive performance, continuous and uniform dispensing, and absence of bubbles and lumps. Specific structure and parameters are as follows: The components include: Glue bucket: Made of stainless steel or corrosion-resistant plastic, with a capacity of 5-20L, equipped with a sealing cap to prevent glue oxidation and contamination; the inner wall of the glue bucket is smooth for easy cleaning, and there is a discharge port at the bottom for connection to pipelines; Pressure tank: Used to provide stable pressure for glue delivery, with a pressure adjustment range of 0.1-0.8MPa, and a volume matching the glue tank (e.g., 5-20L). Equipped with a pressure sensor to monitor the pressure inside the tank in real time, with a pressure deviation ≤ ±0.01MPa. Piping: Corrosion-resistant and high-pressure resistant hoses (such as PTFE hoses) are used, with an inner diameter of 4-10mm and a length of 1-3m. The pipe joints are sealed to prevent glue leakage. The pipes are equipped with filters (filtration accuracy 10-20μm) to remove glue impurities and prevent clogging of the dispensing valve. Adhesive volume controller: It adopts a metering pump driven by a servo motor with a metering accuracy of ≤±1%. It can adjust the dispensing volume (0.01-1ml / s) according to the dispensing speed and adhesive width requirements. It is linked with the motion control module to realize the synchronous adjustment of "motion speed-dispensing volume". Degassing mechanism: Vacuum degassing is adopted, with a vacuum degree ≤-0.09MPa and a degassing time of 10-30min. It removes air bubbles inside the glue and avoids air bubbles and glue breakage during the dispensing process. The degassing mechanism is linked with the glue tank and can automatically degas at timed intervals or be manually started according to the glue bubble situation. Temperature control mechanism: The glue bucket is wrapped with a heating jacket, with a heating power of 500-1000W, a temperature adjustment range of 20-50℃, and a temperature accuracy of ±1℃. The glue temperature is monitored in real time by a temperature sensor to ensure the glue viscosity is stable (the glue viscosity changes with temperature by ≤5% / ℃) and avoid fluctuations in glue output due to temperature changes.

[0049] Adhesive supply logic: The adhesive enters the pressure tank from the glue bucket, and after degassing and constant temperature treatment, it is measured by the adhesive quantity controller and delivered to the dispensing valve through the pipeline. The adhesive supply pressure and dispensing volume are fed back to the control system in real time and dynamically adjusted according to the dispensing process requirements to ensure continuous and uniform dispensing.

[0050] The high-speed tracking system includes: To address potential warping and uneven surface issues in lamp housings (especially large components), the dispensing head height is adjusted in real-time to maintain a constant distance between the nozzle and the lamp housing surface, ensuring consistent glue width and height. This is achieved as follows: Detection component: A laser rangefinder sensor is used, with a measurement range of 0~100mm, a measurement accuracy of ≤0.001mm, and a response speed of ≥100Hz. It is installed on the side of the dispensing head, with the laser direction perpendicular to the surface of the lamp housing. The distance between the measurement point and the needle is ≤10mm to ensure measurement accuracy. Control logic: The laser rangefinder sensor collects the distance between the needle and the lamp housing in real time, transmits the data to the motion controller, compares it with the preset distance (0.5~2mm, determined according to the glue viscosity and needle specifications), calculates the height deviation, controls the Z-axis servo motor to drive the dispensing head to rise and fall, corrects the height deviation in real time, with a correction response time ≤10ms, and ensures that the distance deviation between the needle and the board surface is ≤0.01mm. Adaptive adjustment: A preset height deviation threshold (e.g., ±0.02mm) is provided. The system automatically adjusts when the deviation exceeds the threshold and remains stable when it does not exceed the threshold, thus avoiding frequent adjustments that could cause the dispensing head to vibrate. At the same time, it supports preset height compensation parameters based on the warp conditions of different areas to improve dispensing consistency.

[0051] The software and user interface module is configured as a human-computer interaction window, and through visual guidance software, CAD trajectory generation, database and MES integration, it realizes system integration control, data traceability and visual operation. Specifically, it includes: Visual guidance software: By deeply integrating image processing algorithms, motion control algorithms, and dispensing control algorithms, the system achieves image acquisition, recognition, motion, and dispensing control. The software architecture and core functions are as follows: Software architecture: It adopts a modular design, which is divided into an image acquisition and processing module, a motion control module, and a dispensing execution and control module. Each module communicates through a bus protocol, and the data transmission delay is ≤10ms. Core functions: Image acquisition and control: Controls the industrial camera and light source to turn on and off, sets camera parameters (resolution, frame rate, exposure time), receives and stores image data, and supports single-frame shooting and continuous shooting modes; Algorithm call: Integrates the aforementioned image preprocessing, feature extraction, and pattern recognition algorithms. The corresponding algorithm can be selected according to the dispensing scenario (such as edge dispensing and defect repair). It supports manual adjustment and automatic optimization of algorithm parameters (such as filter kernel size and matching threshold). Linkage control: The coordinate data and dispensing instructions obtained from image processing are transmitted to the motion control module and dispensing execution module in real time to realize synchronous linkage of "see-move-dispense". It supports two modes: manual control (jog, single-step execution) and automatic control (continuous execution). Accuracy calibration: Supports camera calibration, motion axis calibration, and dispensing accuracy calibration. The calibration process is visualized, and calibration data is automatically stored and can be recalled periodically to ensure stable system accuracy.

[0052] Operating environment: Supports Windows 10 Industrial or Linux embedded systems, CPU ≥ Intel Core i7, memory ≥ 16GB, storage ≥ 128GB, graphics card ≥ NVIDIA GTX 1050 (for deep learning algorithm acceleration), software startup time ≤ 30s, no crashes or lag during operation.

[0053] Importing CAD drawings and generating trajectories: It supports direct import of CAD drawings of solar modules, automatically extracts dispensing paths, and achieves accurate correspondence between drawings and actual products, reducing manual programming workload. The specific implementation is as follows: Drawing format support: Supports mainstream CAD formats such as DXF and DWG. After importing, the drawings can be scaled, rotated, and translated to fit the actual lamp housing size, with a scaling accuracy of ≤0.001mm. Trajectory Extraction: Automatically identifies the dispensing path (such as straight line, arc, polygon) in CAD drawings, extracts the path coordinates, fits them with the actual product coordinates identified by the image processing module, calculates the path offset, and automatically corrects the dispensing trajectory with a fitting error ≤0.005mm. Track editing: Supports manual editing of dispensing tracks, allowing users to add, delete, and modify track points, adjust track speed and dispensing volume, and save tracks (in custom format or CSV format) for repeated use, adapting to dispensing needs of different specifications of lamp housings; Preview function: Supports dispensing trajectory preview, displaying the overlay effect of the trajectory and the actual product image in the software interface, which facilitates manual inspection of the trajectory accuracy and avoids dispensing defects caused by trajectory deviation.

[0054] Database integration with MES: It enables the recording, storage, and traceability of production data, and interfaces with the factory's MES (Manufacturing Execution System) to achieve intelligent management of the production process. Specific functions and interfaces are as follows: Database Management: Database type: Uses SQL Server or MySQL database, supports local storage and cloud storage, storage capacity ≥1TB, can store at least 1 year of production data; Record content: process parameters for each dispensing (adhesive volume, dispensing speed, dispensing pressure, temperature, height), test results (pass / fail, defect type, defect location), timestamp, equipment number, operator number, product batch number, etc. Data query and traceability: Supports querying production data by product batch, time, operator, etc., and can export to Excel format for easy data analysis and quality traceability; supports anomaly data marking and automatically records process parameters when a fault occurs for easy troubleshooting.

[0055] MES Integration: Interface protocol: Supports OPCUA or ModbusTCP protocol, communication rate ≥1000Mbps, ensuring real-time data transmission; Data interaction: Receive production tasks (such as product specifications, dispensing process parameters, and production quantity) from the MES system, and report production progress (quantity produced, qualified quantity, and unqualified quantity), equipment operating status (normal / faulty), process parameters, and other data to the MES system. Formula management: Supports receiving dispensing formulas from the MES system, and also allows editing formulas locally. Formulas are stored in the database and can be automatically recalled according to production tasks, avoiding manual input errors.

[0056] user interface: It adopts a visual user interface that is simple, easy to understand, and convenient for operators. Specific functions are as follows: Main interface: Displays equipment operating status (normal / fault), production progress, current process parameters, image acquisition screen, dispensing trajectory preview, etc. Parameter setting interface: Camera parameters, light source parameters, motion parameters, dispensing parameters, constant temperature parameters, etc. can be set. Parameters can be saved and recalled, and the setting range is clearly defined (e.g., dispensing speed 0.1-1m / s). Fault alarm interface: Displays fault type, fault location, and fault cause; supports fault reset and fault record query; some faults (such as insufficient glue or insufficient air pressure) can automatically prompt solutions. Access control: Supports multi-user access settings (administrator, operator). Administrators can modify parameters and view all data, while operators can only operate the equipment and view production progress to prevent accidental operation.

[0057] The auxiliary and safety module is configured to ensure stable system operation, equipment and operator safety, and reduce downtime through fault alarm self-diagnosis, cleaning and curing mechanism and other auxiliary functions; Specifically, it includes: The system monitors the operational status of each module in real time, promptly detects anomalies and issues alarms, and performs preliminary self-diagnosis to facilitate troubleshooting. The specific implementation is as follows: Monitoring objects include: air pressure (adhesive supply system, pneumatic clamps), adhesive level (adhesive tank), camera status (focus, image transmission), servo motor status (speed, position), temperature (adhesive temperature control, equipment temperature), vacuum degree (adsorption mechanism), and dispensing quality (adhesive breakage, misalignment), etc. Alarm method: It adopts audible and visual alarm, and the alarm light color distinguishes the fault level (red: serious fault, the machine needs to be stopped; yellow: general fault, it can continue to operate but needs attention). The alarm sound volume is ≥80dB and can be manually turned off. At the same time, the fault information is displayed on the user interface, and the time, type and location of the fault are recorded. Self-diagnostic function: The system performs self-checks on each module periodically (such as camera self-check, motor self-check, and dispensing valve self-check), and the self-check cycle can be set (such as 10 minutes / time); for common faults (such as insufficient glue, insufficient air pressure, and camera defocusing), it automatically diagnoses the cause of the fault and provides solutions on the interface (such as adding glue, checking the air circuit, and refocusing). Troubleshooting: When a serious fault (such as motor failure or dispensing valve blockage) occurs, the system will automatically shut down and cut off the power to the relevant modules to prevent equipment damage; when a general fault occurs, the system will continue to run and issue an alarm to remind the operator to handle it in time.

[0058] Cleaning and curing facilities: Used to keep the dispensing valve needle clean, prevent glue from curing and clogging, and pre-cure certain types of glue to improve dispensing quality. Its specific structure and function are as follows: Cleaning services: Structure: Integrated at the end of the dispensing station, it consists of a cleaning cup, cleaning solution, and air blowing device; the cleaning cup is made of corrosion-resistant material, has a capacity of 500ml, and contains cleaning solution (such as acetone or special cleaning solution). Cleaning logic: After dispensing, the dispensing head moves above the cleaning cup, the needle is inserted into the cleaning solution and soaked for 1~5 seconds (adjustable). Then, the needle is dried by the air blowing device (air pressure 0.3~0.5MPa) to remove residual glue. The cleaning frequency can be set (e.g., clean once for every 100 dispensing products, or clean once before each shutdown). Additional function: The cleaning fluid level sensor monitors the remaining cleaning fluid level in real time and issues an alarm when the level falls below a preset threshold, reminding the operator to replace the cleaning fluid.

[0059] Curing mechanism: Type: Select a UV curing lamp or a hot air gun according to the type of adhesive; UV curing lamp is used for UV adhesives, with a wavelength of 365~405nm, power of 100~300W, and curing time of 10~30s; hot air gun is used for thermosetting adhesives, with a temperature adjustment range of 50-150℃, a wind speed adjustment range of 0.5-2m / s, and a curing time of 30-60s. Installation location: Located behind the dispensing station, linked with the conveying mechanism. After the lamp housing is dispensed, it is conveyed to the bottom of the curing mechanism for pre-curing. After pre-curing, the surface of the adhesive is cured to prevent deformation of the adhesive path in subsequent processes. Control: Curing time, temperature / power can be set according to the type of adhesive and linked to the dispensing speed to ensure curing effect while avoiding over-curing that could cause the adhesive to crack.

[0060] Other auxiliary functions: Dust protection: The entire system is equipped with a dust cover to prevent workshop dust from entering the equipment and affecting the operation of precision components such as cameras and servo motors; the dust cover is made of transparent material, making it easy to observe the equipment's operating status; Emergency shutdown: The equipment is equipped with an emergency shutdown button, which is installed in an easily accessible location for operators. Pressing the button will immediately cut off the system power supply to ensure the safety of operators. Maintenance prompts: The system automatically prompts operators to perform equipment maintenance (such as replacing needles, cleaning tubing, and checking lead screw lubrication) based on the running time (e.g., 1000 hours). Maintenance records are automatically stored in the database.

[0061] Example 2

[0062] Please see Figure 2An automated control method for solar dispensing machines based on image recognition includes the following parts: The image of the area to be glued on the lamp housing is acquired, and after preprocessing, feature extraction and localization and pattern recognition algorithms, it is converted into coordinate data and control commands to complete the pre-glue positioning and post-glue detection. Through a multi-axis motion control system and a transmission and positioning mechanism, instructions are received and the movement of the dispensing head and the lamp housing is precisely controlled to achieve precise matching of the dispensing trajectory. By utilizing a dispensing valve, a glue supply system, and a height tracking system, the glue can be dispensed quantitatively and at a constant speed, maintaining the width and height of the glue path. By integrating visual guidance software, CAD trajectory generation, database and MES, the system achieves integrated control, data traceability and visual operation; With the help of fault alarm self-diagnosis, cleaning and curing mechanism and other auxiliary functions, the system can ensure stable operation, equipment and personnel safety, and reduce downtime due to faults.

[0063] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0064] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0065] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0066] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0067] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0068] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0069] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0070] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0071] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0072] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An automated control system for a solar-powered dispensing machine based on image recognition, characterized in that: include: The image acquisition and processing module is configured to transport the lamp housings to be glued in batches to the feeding station of the production line to provide basic materials for subsequent processes, identify the front and back orientation of the lamp housings, screen and remove the reversed housings and put them back in place, acquire images of the area of ​​the lamp housing to be glued, and convert the images into coordinate data and control commands through image preprocessing, feature extraction and positioning and pattern recognition algorithms. The motion control module is configured to receive instructions and precisely control the movement of the dispensing head and the lamp housing through a multi-axis motion control system and a transmission and positioning mechanism. It will precisely arrange the housings after forward and reverse screening at equal intervals on the transmission platform to ensure the accuracy of subsequent photo-taking and hole spacing measurement, plasma processing, and dispensing positioning, and ensure the matching of the dispensing trajectory. The dispensing execution and control module is configured to output glue quantitatively and at a constant speed through a dispensing valve, glue supply system, and height tracking system, ensuring consistent glue path width and height. This includes: detecting the hole spacing and hole coordinates in the glue-to-be-dispensed area of ​​the lamp housing, providing a precise physical coordinate reference for subsequent dispensing trajectory planning; performing plasma cleaning on the glue-to-be-dispensed surface of the housing to remove surface impurities; and achieving quantitative, uniform, and high-precision dispensing on the glue-to-be-dispensed surface of the housing to ensure consistent glue path width and height. The software and user interface module is configured as a human-computer interaction window, enabling system integration control, data traceability, and visual operation through visual guidance software, CAD trajectory generation, database, and MES integration.

2. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 1, characterized in that, Image acquisition in the image acquisition and processing module specifically includes: High-resolution CCD or CMOS cameras are used, and the selection should meet the size requirements of the lamp housing and the accuracy requirements of the features to be identified. Telecentric lenses are selected for optical applications; industrial fixed-focus lenses are selected for wide-field-of-view scenarios. High-brightness LED light source is used; the light source style is selected according to the type of feature to be identified. The controller uses a constant current source controller; the light source is installed coaxially with the lens or around the lens, and is adjusted according to the lens focal length and field of view to ensure that the lighting area completely covers the shooting area without shadows or uneven light spots.

3. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 2, characterized in that, Image preprocessing, feature extraction and localization, and pattern recognition algorithms are used to convert images into coordinate data and control commands, specifically including: The original image is subjected to noise reduction, contrast enhancement, and geometric correction to eliminate interference and ensure the clarity of the features to be identified. A threshold segmentation algorithm is used to divide the image into foreground and background; In the preprocessed image, the reference features and the features to be applied to the lamp housing are extracted, and their actual physical coordinates are calculated to provide a position reference for the motion control module. Using the Mark point coordinates as a reference, the detected edge coordinates are converted into actual physical coordinates, and the relative position between the area to be dispensed and the edge is calculated to ensure that the deviation between the dispensing trajectory and the edge meets the preset requirements. Identify the specific area to be glued and judge the glue dispensing quality.

4. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 1, characterized in that, The motion control module specifically includes: An AC servo motor is selected, and the X / Y axes use ball screw drive, with the screw connected to the servo motor via a coupling; the Z axis uses a ball screw or linear guide rail to accommodate the lifting and lowering motion of the dispensing head. It adopts an industrial-grade multi-axis motion controller with an interface to receive coordinate commands from the image acquisition and processing module in real time and output control signals to the servo motor. The conveying and positioning mechanism is responsible for transporting the lamp housing from the feeding station to the dispensing station, completing the positioning and fixing, preventing the lamp housing from shifting during the dispensing process, and ensuring dispensing accuracy.

5. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 1, characterized in that, The dispensing valve and dispensing system in the dispensing execution and control module specifically include: The dispensing valve is selected based on the type and viscosity of the adhesive and the requirements of the dispensing process. It controls the opening, closing, and output of the adhesive to ensure dispensing accuracy. Specific parameters are as follows: Screw valve: suitable for high viscosity adhesives; spray valve: suitable for medium and low viscosity adhesives; impact valve: suitable for low viscosity adhesives; The glue supply system is responsible for the storage, transportation, degassing, and temperature control of the glue to ensure stable glue performance and continuous glue dispensing. Its components include glue tanks, pressure tanks, pipelines, glue volume controllers, degassing mechanisms, and temperature control mechanisms. The adhesive enters the pressure tank from the glue bucket. After degassing and constant temperature treatment, it is measured by the glue quantity controller and delivered to the dispensing valve through pipeline. The glue supply pressure and dispensing volume are fed back to the control system in real time and dynamically adjusted according to the dispensing process requirements.

6. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 5, characterized in that, The high-speed tracking system includes: A laser rangefinder sensor is used and installed on the side of the dispensing head, with the laser direction perpendicular to the surface of the lamp housing. The laser rangefinder sensor collects the distance between the needle tip and the lamp housing in real time, transmits the data to the motion controller, compares it with the preset distance, calculates the height deviation, controls the Z-axis servo motor to drive the dispensing head to rise and fall, and corrects the height deviation in real time.

7. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 1, characterized in that, After the glue dispensing is completed, the following steps are required: Pre-curing treatment is performed on the adhesive layer after dispensing the adhesive onto the housing to shape the surface of the adhesive layer, thereby preventing deformation and sticking of the adhesive path during subsequent silicon wafer assembly and improving the initial bonding strength of the adhesive layer. The solar silicon wafers are precisely arranged in batches and transported to the wafer assembly station, where they are precisely docked with the housings that have been glued and cured, thus achieving automated assembly of the silicon wafers and the housings. Based on the silicon wafer model and the dimensions of the housing assembly position, adjust the position and gripping angle of the silicon wafer picking robot module, and at the same time measure the actual length of the silicon wafer leads to determine whether it meets the assembly process requirements. The qualified solar silicon wafers are precisely picked up and transferred to the adhesive application point on the pre-cured casing to achieve initial bonding between the silicon wafers and the casing.

8. The automated control system for a solar-powered dispensing machine based on image recognition according to claim 7, characterized in that, Also includes: Based on the dimensions of the housing perforation and assembly requirements, the silicon wafer leads are precisely bent multiple times to ensure that the formed leads fit the housing perforation and subsequent electrical connection requirements. The silicon wafer leads, after being bent multiple times, are precisely inserted into the pre-set holes in the lamp housing to achieve mechanical fixation between the silicon wafer and the housing; The housing, after silicon wafer assembly and lead wire perforation, is precisely clamped and flipped to a preset angle so that the back of the housing faces upward, providing an operating surface for subsequent hole plugging and lead wire bending processes. The perforations in the housing are sealed with adhesive, and the leads that penetrate the back of the housing are bent again to make the lead shape adapt to the subsequent electrical assembly requirements, thus completing the initial forming of the photovoltaic module.

9. An automated control method for a solar-powered dispensing machine based on image recognition, wherein the automated control system for a solar-powered dispensing machine based on image recognition according to any one of claims 1-8 is characterized in that, include: The image of the area to be glued on the lamp housing is acquired, processed, and converted into coordinate data and control commands to complete pre-glue positioning and post-glue detection. Through a multi-axis motion control system and a transmission and positioning mechanism, instructions are received and the movement of the dispensing head and the lamp housing are controlled to achieve dispensing trajectory matching. By utilizing a dispensing valve, a glue supply system, and a height tracking system, the glue can be dispensed quantitatively and at a constant speed, maintaining the width and height of the glue path. By integrating visual guidance software, CAD trajectory generation, database and MES, the system achieves integrated control, data traceability and visual operation; By leveraging fault alarm self-diagnosis and cleaning / curing mechanisms, the system ensures stable operation, equipment and personnel safety, and reduces downtime due to malfunctions.

Images