Enamel product and intelligent spraying system thereof

Abstract

The invention discloses an enamel product and an intelligent spraying system thereof, and belongs to the technical field of enamel product spraying, and the enamel product comprises a high-precision visual guidance module, an environment intelligent control module and an auxiliary function module, the intelligent environment control module is used for monitoring and dynamically regulating and controlling parameters such as temperature and humidity and dust concentration of a spraying environment in real time, ensuring stable coating quality and reducing energy consumption, and the auxiliary function module is used for realizing accurate spraying track planning and quality control through process connection, safety guarantee and data collaboration. And high efficiency and reliability of the spraying process are guaranteed. On the basis that spraying of the enamel product is achieved, the quality of the whole process is improved, and comprehensive benefits can be optimized.

Description

Technical Field

[0001] This invention relates to the field of enamel product spraying technology, and more specifically, to an enamel product and its intelligent spraying system. Background Technology

[0002] Enameled products are composite materials made by coating an inorganic vitreous porcelain enamel (enamel) onto a metal substrate and firing it at high temperatures. They combine the mechanical strength of metal with the corrosion resistance and easy-to-clean properties of porcelain enamel. The manufacturing process includes metal forming, surface treatment, enamel coating, and high-temperature firing. Based on their uses, they can be divided into seven categories: daily-use, sanitary, industrial, architectural, medical, artistic, and special enamels, and are widely used in household utensils, chemical equipment, building materials, and other fields.

[0003] Enamelware originated in ancient Egypt, with cloisonné enamelware from the Ming Dynasty in China being a branch of its craft. After 1916, with the application of steel stamping and cast iron casting technologies, enamelware products transitioned from art pieces to everyday items, becoming common household utensils in China by the mid-20th century. After the 1990s, due to issues such as the susceptibility of enamel to chipping and the tendency for metal to rust, it was gradually replaced by stainless steel and plastic, leading to the decline of the traditional everyday enamelware industry. Modern enamelware products are shifting towards high-end and functional applications, used in industrial corrosion protection, medical equipment, and aerospace, while also expanding their artistic and design value through enamel techniques.

[0004] Existing spraying systems rely on manual spraying, which depends on worker experience. Curved areas are prone to missed spraying, and environmental interference cannot be compensated for in real time. In addition, the path planning of the robotic arm depends on offline programming and cannot dynamically respond to workpiece pose drift. Furthermore, defect detection relies on manual visual inspection, which increases the rate of missed bubble detection and thus increases the rework cost as a percentage of the total spraying cost. To address this, we propose an intelligent spraying system for enamel products. Summary of the Invention

[0005] 1. Technical problems to be solved In view of the problems existing in the prior art, the purpose of this invention is to provide an enamel product and its intelligent spraying system. This invention not only realizes the spraying of enamel products, but also improves the quality of the entire process and optimizes the overall benefits.

[0006] 2. Technical Solution

[0007] To solve the above problems, the present invention adopts the following technical solution: An intelligent spraying system for enamel products includes: a high-precision vision guidance module, an intelligent environmental control module, and an auxiliary function module; The high-precision vision guidance module is used to achieve precise spraying trajectory planning and quality control by sensing the geometric features of the workpiece and the environmental conditions in real time. The intelligent environmental control module is used to monitor and dynamically adjust parameters such as temperature, humidity, and dust concentration in the spraying environment in real time to ensure stable coating quality and reduce energy consumption. The auxiliary function module is used to ensure the efficiency and reliability of the spraying process through process connection, safety assurance, and data collaboration.

[0008] As a preferred embodiment of the present invention, the high-precision visual guidance module includes a three-dimensional perception and modeling submodule, a defect detection and compensation submodule, a dynamic trajectory planning submodule, and a real-time positioning and attitude calibration submodule. The 3D perception and modeling submodule is used to collect point cloud data of the workpiece surface through multi-sensor fusion, and stitch the point cloud to generate a 3D model to fit the features of the irregular curved surface. The defect detection and compensation submodule is used to capture spraying defects in real time using multispectral imaging technology and to distinguish defect types using AI classification algorithms. The dynamic trajectory planning submodule is used to divide the high-precision area into the ordinary area, and dynamically adjust the spray gun speed and distance, while training the obstacle avoidance logic of the robotic arm according to the reinforcement learning algorithm. The real-time positioning and attitude calibration submodule solves the problems of positioning drift and attitude deviation under complex working conditions through multi-sensor fusion and dynamic compensation algorithms.

[0009] As a preferred embodiment of the present invention, the three-dimensional perception and modeling submodule includes a data acquisition unit, a high-precision calibration unit, a data preprocessing unit, and a three-dimensional reconstruction optimization unit; The data acquisition unit includes a depth camera array, an active infrared supplementary lighting subunit, and a lidar scanner. The depth camera array consists of multiple structured light cameras and a line laser profilometer deployed at the processing position. The depth camera array is used to synchronously acquire point cloud data from multiple angles around the workpiece, covering the entire surface. The active infrared supplementary lighting subunit is used to project infrared structured light in low-light or high-reflectivity environments and calculate surface depth information by analyzing the deformation of the light spot. The lidar scanner is used to acquire continuous contour data by sliding a line lidar along a guide rail. The high-precision calibration unit is used to establish the mapping relationship between the coordinate system of each camera and the world coordinate system based on the checkerboard calibration plate and the pinhole imaging principle, thereby eliminating lens distortion. It also uses PTP protocol hardware synchronization signal to ensure the synchronization of data acquisition time of multiple cameras. The data preprocessing unit is used to remove noise points using a radius filtering statistical algorithm, retain effective surface data, and then synthesize a multi-frequency phase map by projecting phase shift patterns of different frequencies. The synthesis adopts point cloud stitching and registration by extracting SIFT key points in the point cloud and using the ICP iterative nearest point algorithm to achieve accurate stitching of multi-view point clouds. After stitching, the divergence-free wavelet technology is applied to process the original scan data to reconstruct a high-fidelity depth map. The 3D reconstruction optimization unit includes a triangular patch meshing subunit and an adaptive mesh subdivision subunit. The triangular patch meshing subunit is used to convert point cloud data into a triangular mesh model, retaining geometric features, and optimizing mesh noise through an edge-preserving smoothing algorithm while sharpening edge features. The adaptive mesh subdivision subunit is used to automatically refine the mesh in high curvature areas and simplify the mesh in low curvature areas, balancing data volume and accuracy.

[0010] As a preferred embodiment of the present invention, the defect detection and compensation submodule includes a thermal imaging and thickness sensing unit, a defect identification unit, a classification and evaluation unit, and a compensation decision unit. The thermal imaging and thickness sensing unit is used to monitor the temperature field distribution and thickness change of the coating in real time and identify areas of uneven spraying. The thermal imaging and thickness sensing unit consists of an infrared thermal imager and a laser thickness gauge. The infrared thermal imager detects the coating distribution based on the Seebeck effect, and the laser thickness gauge dynamically scans the film thickness through the triangulation principle. The defect identification unit is used to extract defect features from multi-source data, where the defect features are scratches, bubbles and color differences. The defect identification unit locates the defect area by combining threshold segmentation, morphological operations and grayscale histogram analysis. At the same time, it uses a U-Net network to perform semantic segmentation of the defect and outputs a pixel-level mask image. The classification and evaluation unit distinguishes defect categories based on LBP texture features and gradient features, classifies defect types, and evaluates mild or severe severity levels. The compensation decision unit is used to adjust the spraying flow rate, distance, and angle according to the defect type, and to automatically switch faulty nozzles. The compensation decision unit performs compensation in collaboration with AI through a rule base, increasing the paint flow rate when the coating is too thin, and adjusting the nozzle angle and reducing the spraying speed when there is a sagging defect. The flow rate formula is as follows: ,in This is for thickness deviation.

[0011] As a preferred embodiment of the present invention, the dynamic trajectory planning submodule includes a three-dimensional model processing unit, an initial path planning unit, a multi-objective optimization unit, a real-time trajectory correction unit, an environmental coupling compensation unit, a process parameter matching unit, a zoned spraying strategy unit, and an execution control unit. The 3D model processing unit includes a geometric topology analysis subunit and a feature semantic segmentation unit. The geometric topology analysis subunit calculates vertex normal vectors based on the area weighted interpolation of adjacent triangular facets. Simultaneously, it combines curvature analysis to analyze the STL format 3D point cloud model generated by the vision module and extracts the topological relationships of the triangular mesh. The topological relationships include vertex, edge, and facet set. The feature semantic segmentation unit uses the U-Net deep learning network to segment the surface. At the same time, it uses multi-scale Gaussian filtering combined with principal curvature analysis to distinguish edges, corners, and planar regions, and divides the workpiece surface into functional regions, including inner walls, outer edges, and grooves, and outputs a spraying priority mask map. The initial path planning unit generates intersection trajectories based on the triangular mesh domain slicing algorithm, generates tool paths using the surface method, and supports hovering and detouring around marked obstacle areas, thereby generating a basic spraying path. The multi-objective optimization unit uses quadratic curves to model the thickness of adjacent trajectory superposition, dynamically adjusts the spray width, and optimizes the transition path between groups, thereby balancing spray uniformity, efficiency, and paint consumption. The real-time trajectory correction unit trains the robotic arm to avoid protruding parts, compensates for accumulated errors by combining IMU data, and dynamically adjusts the spray gun speed and distance according to the position deviation, thereby reducing the pose drift caused by mechanical vibration and temperature and humidity fluctuations. The environmental coupling compensation unit is used to monitor humidity changes, dynamically increase the spray gun voltage, and adjust the atomization pressure based on aerosol particle counter data, thereby offsetting the effect of high humidity environment on electrostatic adsorption force. The process parameter matching unit is used to associate the paint characteristic database with the spraying parameters. The process parameter matching unit optimizes the parameters through a weighted negative binomial distribution to adapt to different glaze characteristics, and simulates 100,000 working conditions on the digital twin platform to preview the parameter effects. The partitioned spraying strategy unit is used to dynamically adjust the spraying mode according to the surface characteristics, adjusting the spraying mode to a wide fan-shaped spray for flat workpieces and a narrow cone-shaped spray for curved workpieces; The execution control unit is used to convert the planned trajectory into six-axis robotic arm joint commands. The execution control unit calculates joint variables based on the target pose of the end effector and fine-tunes the contour spray gun to ensure continuous changes in speed and flow rate at corners and avoid coating accumulation. The execution control unit also controls the switching of heterogeneous nozzles through relays and automatically cleans the clogged unit after triggering.

[0012] As a preferred embodiment of the present invention, the real-time positioning and attitude calibration submodule includes a multimodal sensor array and a multi-coordinate system synchronization engine; The multimodal sensor array is used to collect workpiece spatial pose data in real time, covering six degrees of freedom information of position and angle. The multi-coordinate system synchronization engine is used to unify the visual coordinate system, the robot arm base coordinate system, and the workpiece coordinate system.

[0013] As a preferred embodiment of the present invention, the multimodal sensor array includes an IMU (Inertial Measurement Unit), a visual target unit, and a laser rangefinder. The IMU includes a three-axis gyroscope and a three-axis accelerometer. The IMU is used to output pitch, roll, and yaw angles in real time. The visual target unit is a 4×4 dot matrix reflective target and an infrared camera. The visual target unit identifies the physical coordinates of the target. The laser rangefinder is used to assist in calibrating the Z-axis depth and compensate for blind spots. The multi-coordinate system synchronization engine solves the camera extrinsic parameters through the PnP algorithm, maps pixel coordinates to physical coordinates, and refreshes the pose data every 120Hz to ensure that the robot arm trajectory is synchronized with the actual pose of the workpiece.

[0014] As a preferred embodiment of the present invention, the intelligent environmental control module includes an environmental monitoring submodule, a temperature and humidity control submodule, a dehumidification unit, an air purification submodule, a real-time monitoring submodule, and a linkage control submodule. The environmental monitoring submodule is used to synchronously collect temperature and humidity data using distributed temperature and humidity sensors, detect dust concentration using the β-ray method, and monitor VOCs data using an electrochemical sensor. The temperature and humidity control submodule is used to dynamically adjust the spraying parameters based on environmental data to counteract the effects of temperature and humidity fluctuations.

[0015] The dehumidification unit is used to maintain a constant temperature and humidity in the spraying isolation room; The air purification submodule is used to capture spraying dust and degrade VOCs to ensure environmental safety. The real-time monitoring submodule is used to display visual environmental parameter curves and trigger audible and visual alarms when abnormalities occur. The linkage control submodule is used to link the vision guidance module and the spraying execution unit to achieve dynamic matching of environment and process.

[0016] As a preferred embodiment of the present invention, the auxiliary function module includes a human-computer interaction submodule, a mobile collaboration platform, a full lifecycle database, and an intelligent analysis engine; The human-computer interaction submodule is a touch screen used to display real-time process parameter curves, defect heat maps and environmental monitoring data. The mobile collaboration platform is used to push alarm information and supports remote start / stop and policy adjustment. The full lifecycle database is used to store workpiece IDs, spraying parameters, defect records, and environmental data, and supports SQL queries and cloud synchronization. The intelligent analysis engine is used to calculate the mean and standard deviation of parameters using regression analysis to evaluate process stability, and to aggregate data from multiple production lines through a federated learning framework to optimize the generalization ability of the AI ​​model.

[0017] An enamel product, adapted to the processing methods for enamel products. Beneficial effects

[0018] Compared with the prior art, the advantages of this invention are: (1) This invention uses multi-sensor fusion positioning to compress pose error, curvature-driven partition spraying strategy to reduce thickness deviation in groove area, AI defect classification to improve bubble detection rate and reduce missed detection rate, temperature and humidity coupling control model to dynamically adjust voltage and atomization pressure to ensure humidity and adsorption force fluctuation during operation, and digital twin to improve parameter matching accuracy for various working conditions.

[0019] (2) The present invention collects workpiece point cloud data in real time and generates spraying trajectory through a high-precision vision guidance module. The environmental intelligent control module dynamically adjusts the spray gun voltage and atomization pressure based on distributed temperature and humidity sensors and dust monitoring unit to offset the electrostatic attenuation caused by high humidity. The auxiliary function module realizes remote control of process parameters through human-machine interaction interface, solving the problem of coating bubbles and uneven thickness caused by environmental disturbance in traditional spraying. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of a module of an intelligent spraying system for enamel products according to the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0022] Example:

[0023] Please see Figure 1 An intelligent spraying system for enamel products includes: a high-precision vision guidance module, an intelligent environmental control module, and an auxiliary function module; The high-precision vision guidance module is used to achieve precise spraying trajectory planning and quality control by sensing the geometric features of the workpiece and the environmental conditions in real time. The intelligent environmental control module is used to monitor and dynamically adjust parameters such as temperature, humidity and dust concentration in the spraying environment in real time to ensure stable coating quality and reduce energy consumption. The auxiliary function module is used to ensure the efficiency and reliability of the spraying process through process connection, safety assurance and data collaboration.

[0024] In a specific embodiment of the present invention, a high-precision vision guidance module collects workpiece point cloud data in real time and generates a spraying trajectory. An environmental intelligent control module dynamically adjusts the spray gun voltage and atomization pressure based on a distributed temperature and humidity sensor and a dust monitoring unit to counteract electrostatic attenuation caused by high humidity. An auxiliary function module enables remote control of process parameters through a human-machine interface, solving the problems of coating bubbles and uneven thickness caused by environmental disturbances in traditional spraying.

[0025] Specifically, the high-precision vision guidance module includes a 3D perception and modeling submodule, a defect detection and compensation submodule, a dynamic trajectory planning submodule, and a real-time positioning and attitude calibration submodule; The 3D perception and modeling submodule is used to collect point cloud data of the workpiece surface through multi-sensor fusion, and stitch the point cloud to generate a 3D model and fit the features of the irregular curved surface. The defect detection and compensation submodule is used to capture spraying defects in real time using multispectral imaging technology and to distinguish defect types using AI classification algorithms. The dynamic trajectory planning submodule is used to divide the high-precision area into the normal area and dynamically adjust the spray gun speed and distance, while training the obstacle avoidance logic of the robotic arm according to the reinforcement learning algorithm. The real-time positioning and attitude calibration submodule solves the problems of positioning drift and attitude deviation under complex working conditions through multi-sensor fusion and dynamic compensation algorithms.

[0026] In a specific embodiment of the present invention, the 3D perception and modeling submodule synchronously acquires point clouds through multiple structured light cameras, combines the ICP algorithm to stitch the point clouds and generate an STL format mesh model, the defect detection submodule uses multispectral imaging to identify bubbles and scratches in real time, the dynamic trajectory planning submodule trains the obstacle avoidance logic of the robotic arm based on reinforcement learning algorithm, and the positioning calibration submodule compensates for pose drift by fusing data from IMU and laser rangefinder, thereby achieving complex curved surface coverage spraying, reducing the missed spraying rate, and suppressing positioning drift.

[0027] Specifically, the 3D perception and modeling submodule includes a data acquisition unit, a high-precision calibration unit, a data preprocessing unit, and a 3D reconstruction optimization unit; The data acquisition unit includes a depth camera array, an active infrared supplementary lighting subunit, and a lidar scanner. The depth camera array consists of multiple structured light cameras and a line laser profilometer deployed at the processing position. The depth camera array is used to synchronously acquire point cloud data from multiple angles around the workpiece, covering the entire surface. The active infrared supplementary lighting subunit is used to project infrared structured light in low light or high reflectivity environments and calculate surface depth information by analyzing the deformation of the light spot. The lidar scanner is used to acquire continuous contour data by sliding a line lidar along a guide rail. The high-precision calibration unit is used to establish the mapping relationship between the coordinate system of each camera and the world coordinate system based on the checkerboard calibration plate and the pinhole imaging principle, thereby eliminating lens distortion. It also uses PTP protocol hardware synchronization signal to ensure the synchronization of data acquisition time of multiple cameras. The data preprocessing unit is used to remove noise points using a radius filtering statistical algorithm and retain effective surface data. Then, phase shift patterns projected at different frequencies are synthesized into a multi-frequency phase map. The synthesis adopts point cloud stitching and registration. By extracting SIFT key points from the point cloud, the ICP iterative nearest point algorithm is used to achieve accurate stitching of multi-view point clouds. After stitching, the divergence-free wavelet technology is applied to process the original scan data and reconstruct a high-fidelity depth map. The 3D reconstruction optimization unit includes a triangular patch meshing subunit and an adaptive mesh subdivision subunit. The triangular patch meshing subunit is used to convert point cloud data into a triangular mesh model, preserving geometric features and optimizing mesh noise through an edge-preserving smoothing algorithm, while sharpening edge features. The adaptive mesh subdivision subunit is used to automatically refine the mesh in high curvature areas and simplify the mesh in low curvature areas, balancing data volume and accuracy.

[0028] In a specific embodiment of the present invention, the data acquisition unit scans the workpiece contour using a line laser profilometer, the active infrared illumination unit projects structured light onto the reflective surface and analyzes the light spot deformation, the high-precision calibration unit uses a checkerboard calibration plate and PTP protocol to achieve time synchronization of multiple cameras, the data preprocessing unit applies radius filtering to denoise and registers multi-view point clouds using the ICP algorithm, the 3D reconstruction unit uses an edge-preserving smoothing algorithm to optimize the mesh, and the mesh in high curvature areas is densified to 0.5mm resolution to improve the model reconstruction speed and support adaptive spraying of irregular curved surfaces.

[0029] Specifically, the defect detection and compensation submodule includes a thermal imaging and thickness sensing unit, a defect identification unit, a classification and evaluation unit, and a compensation decision unit; The thermal imaging and thickness sensing unit is used to monitor the temperature field distribution and thickness change of the coating in real time and identify areas of uneven spraying. The thermal imaging and thickness sensing unit consists of an infrared thermal imager and a laser thickness gauge. The infrared thermal imager detects the coating distribution based on the Seebeck effect, and the laser thickness gauge dynamically scans the film thickness through the triangulation principle. The defect identification unit is used to extract defect features from multi-source data, including scratches, bubbles, and color differences. The defect identification unit locates the defect area by combining threshold segmentation, morphological operations, and grayscale histogram analysis. At the same time, it uses the U-Net network to perform semantic segmentation of the defects and outputs a pixel-level mask image. The classification and assessment unit distinguishes defect categories based on LBP texture features and gradient features, classifies defect types, and assesses mild or severe severity levels. The compensation decision unit adjusts the spray flow rate, distance, and angle based on the defect type, and automatically switches to the faulty nozzle. The compensation decision unit collaborates with AI through a rule base to perform compensation, increasing the paint flow rate when the coating is too thin and adjusting the nozzle angle and reducing the spraying speed when there is a sagging defect. The flow rate formula is as follows: ,in This is for thickness deviation.

[0030] In a specific embodiment of the present invention, the thermal imaging unit detects the temperature field distribution based on the Seebeck effect, the laser thickness gauge dynamically scans the film thickness through triangulation, the defect identification unit uses LBP texture features and U-Net network to output pixel-level defect mask images, the classification unit evaluates the defect level through the SVM algorithm, and the compensation decision unit dynamically adjusts the flow rate according to the formula and switches the faulty nozzle.

[0031] Specifically, the dynamic trajectory planning submodule includes a 3D model processing unit, an initial path planning unit, a multi-objective optimization unit, a real-time trajectory correction unit, an environmental coupling compensation unit, a process parameter matching unit, a zoned spraying strategy unit, and an execution control unit. The 3D model processing unit includes a geometric topology analysis subunit and a feature semantic segmentation unit. The geometric topology analysis subunit calculates vertex normal vectors based on the area weighted interpolation of adjacent triangular facets. At the same time, it combines curvature analysis to analyze the STL format 3D point cloud model generated by the vision module and extracts the topological relationships of the triangular mesh. The topological relationships include the set of vertices, edges, and regions. The feature semantic segmentation unit uses the U-Net deep learning network to segment the surface. At the same time, it uses multi-scale Gaussian filtering combined with principal curvature analysis to distinguish edges, corners, and planar regions, and divides the workpiece surface into functional regions, including inner walls, outer edges, and grooves. It outputs a spraying priority mask map. The initial path planning unit generates the intersection trajectory based on the triangular mesh domain slicing algorithm, generates the tool path using the surface method, and supports hovering and detouring around the marked obstacle areas, thereby generating the basic spraying path; The multi-objective optimization unit uses quadratic curves to model the thickness of adjacent trajectories, dynamically adjusts the spray width, and optimizes the transition path between groups, thereby balancing spray uniformity, efficiency, and paint consumption. The real-time trajectory correction unit trains the robotic arm to avoid protruding parts, combines IMU data to compensate for accumulated errors, and dynamically adjusts the spray gun speed and distance according to the position deviation, thereby reducing the pose drift caused by mechanical vibration and temperature and humidity fluctuations. The environmental coupling compensation unit is used to monitor humidity changes, dynamically increase the spray gun voltage, and adjust the atomization pressure based on aerosol particle counter data, thereby offsetting the effect of high humidity environment on electrostatic adsorption force. The process parameter matching unit is used to associate the paint characteristic database with the spraying parameters. The process parameter matching unit optimizes the parameters through a weighted negative binomial distribution to adapt to different glaze characteristics, and simulates 100,000 working conditions on the digital twin platform to preview the parameter effects. The zoned spraying strategy unit is used to dynamically adjust the spraying mode according to the surface characteristics, adjusting it to a wide fan-shaped spray for flat workpieces and a narrow cone-shaped spray for curved workpieces; The execution control unit is used to convert the planned trajectory into joint commands of the six-axis robotic arm. The execution control unit calculates joint variables based on the target pose of the end effector and fine-tunes the contour spray gun to ensure continuous changes in speed and flow at corners and avoid coating buildup. The execution control unit also controls the switching of heterogeneous nozzles through relays and automatically cleans the clogged unit after triggering.

[0032] In a specific embodiment of the present invention, the geometric topology analysis unit extracts the vertex normal vectors of the triangular mesh, the initial path unit uses the domain slicing algorithm to generate a collision-free path, the multi-objective optimization unit uses quadratic curves to model the coating film superposition thickness and dynamically adjusts the spray width, and the environmental coupling unit pre-simulates 100,000 working conditions on the digital twin platform according to the process matching unit.

[0033] Specifically, the real-time positioning and attitude calibration submodule includes a multimodal sensor array and a multi-coordinate system synchronization engine; A multimodal sensor array is used to acquire workpiece spatial pose data in real time, covering six degrees of freedom information of position and angle. A multi-coordinate system synchronization engine is used to unify the vision coordinate system, the robot arm base coordinate system, and the workpiece coordinate system.

[0034] In a specific embodiment of the present invention, a multimodal sensor array acquires six-degree-of-freedom pose data, and a multi-coordinate system synchronization engine solves the camera extrinsic parameters through the PnP algorithm, mapping pixel coordinates to the physical coordinate system at a 120Hz refresh rate to reduce pose alignment errors.

[0035] Specifically, the multimodal sensor array includes an IMU (Inertial Measurement Unit), a visual target unit, and a laser rangefinder. The IMU includes a three-axis gyroscope and a three-axis accelerometer. The IMU is used to output pitch, roll, and yaw angles in real time. The visual target unit consists of a 4×4 dot matrix reflective target and an infrared camera. The visual target unit identifies the physical coordinates of the target. The laser rangefinder is used to assist in calibrating the Z-axis depth and compensate for blind spots. The multi-coordinate system synchronization engine solves the camera extrinsic parameters through the PnP algorithm, maps pixel coordinates to physical coordinates, and refreshes pose data every 120Hz to ensure that the robot arm trajectory is synchronized with the actual pose of the workpiece.

[0036] In a specific embodiment of the present invention, the IMU unit outputs pitch or roll angle, the visual target unit locates the physical coordinates through a 4×4 dot matrix reflective target, the laser rangefinder compensates for the Z-axis blind zone, and the PnP algorithm updates the pose data every 120Hz and synchronizes it to the robotic arm base coordinate system, solving the positioning deviation caused by the swaying of the basket, which is suitable for high-altitude operation scenarios.

[0037] Specifically, the intelligent environmental control module includes an environmental monitoring submodule, a temperature and humidity control submodule, a dehumidification unit, an air purification submodule, a real-time monitoring submodule, and a linkage control submodule; The environmental monitoring submodule is used to synchronously collect temperature and humidity data using distributed temperature and humidity sensors, detect dust concentration using the β-ray method, and monitor VOCs data using electrochemical sensors. The temperature and humidity control submodule is used to dynamically adjust spraying parameters based on environmental data to counteract the effects of temperature and humidity fluctuations.

[0038] The dehumidification unit is used to maintain a constant temperature and humidity in the spraying isolation chamber; The air purification submodule is used to capture spray dust and degrade VOCs to ensure environmental safety; The real-time monitoring submodule is used to display visual environmental parameter curves and trigger audible and visual alarms when anomalies occur; The linkage control submodule is used to link the vision guidance module and the spraying execution unit to achieve dynamic matching of environment and process.

[0039] In a specific embodiment of the present invention, the environmental monitoring submodule collects temperature, humidity and VOCs data through a distributed sensor network, the temperature and humidity control unit links the variable frequency air conditioner and the rotary dehumidifier, the air purification unit uses a cyclone separator + catalytic combustion device to degrade VOCs, and the linkage control submodule dynamically adjusts the spray gun distance according to the humidity data.

[0040] Specifically, the auxiliary function modules include a human-computer interaction sub-module, a mobile collaboration platform, a full lifecycle database, and an intelligent analysis engine; The human-machine interaction submodule is a touch screen used to display real-time process parameter curves, defect heat maps, and environmental monitoring data; The mobile collaboration platform is used to push alarm information and supports remote start / stop and policy adjustment; The full lifecycle database is used to store workpiece IDs, spraying parameters, defect records, and environmental data, and supports SQL queries and cloud synchronization. The intelligent analysis engine is used to calculate the mean and standard deviation of parameters using regression analysis to assess process stability, and to aggregate data from multiple production lines through a federated learning framework to optimize the generalization ability of AI models.

[0041] In a specific embodiment of the present invention, the human-machine interaction submodule displays real-time process curves through a 10-inch touchscreen, pushes humidity change alarms through a mobile platform, stores 128GB of operating condition data in a full lifecycle database, and the intelligent analysis engine aggregates data from multiple production lines through federated learning to optimize the AI ​​model, reduce remote control response, and improve the accuracy of process stability analysis.

[0042] An enamel product, adapted to the processing methods for enamel products.

[0043] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and its improved concept, should be covered within the scope of protection of the present invention.

Claims

1. An intelligent spraying system for enamel products, characterized in that, include: High-precision vision guidance module, intelligent environmental control module, and auxiliary function module; The high-precision vision guidance module is used to achieve precise spraying trajectory planning and quality control by sensing the geometric features of the workpiece and the environmental conditions in real time. The intelligent environmental control module is used to monitor and dynamically adjust parameters such as temperature, humidity, and dust concentration in the spraying environment in real time to ensure stable coating quality and reduce energy consumption. The auxiliary function module is used to ensure the efficiency and reliability of the spraying process through process connection, safety assurance, and data collaboration.

2. The intelligent spraying system for enamel products according to claim 1, characterized in that, The high-precision vision guidance module includes a 3D perception and modeling submodule, a defect detection and compensation submodule, a dynamic trajectory planning submodule, and a real-time positioning and attitude calibration submodule. The 3D perception and modeling submodule is used to collect point cloud data of the workpiece surface through multi-sensor fusion, and stitch the point cloud to generate a 3D model to fit the features of the irregular curved surface. The defect detection and compensation submodule is used to capture spraying defects in real time using multispectral imaging technology and to distinguish defect types using AI classification algorithms. The dynamic trajectory planning submodule is used to divide the high-precision area into the ordinary area, and dynamically adjust the spray gun speed and distance, while training the obstacle avoidance logic of the robotic arm according to the reinforcement learning algorithm. The real-time positioning and attitude calibration submodule solves the problems of positioning drift and attitude deviation under complex working conditions through multi-sensor fusion and dynamic compensation algorithms.

3. The intelligent spraying system for enamel products according to claim 2, characterized in that, The 3D perception and modeling submodule includes a data acquisition unit, a high-precision calibration unit, a data preprocessing unit, and a 3D reconstruction optimization unit; The data acquisition unit includes a depth camera array, an active infrared supplementary lighting subunit, and a lidar scanner. The depth camera array consists of multiple structured light cameras and a line laser profilometer deployed at the processing position. The depth camera array is used to synchronously acquire point cloud data from multiple angles around the workpiece, covering the entire surface. The active infrared supplementary lighting subunit is used to project infrared structured light in low-light or high-reflectivity environments and calculate surface depth information by analyzing the deformation of the light spot. The lidar scanner is used to acquire continuous contour data by sliding a line lidar along a guide rail. The high-precision calibration unit is used to establish the mapping relationship between the coordinate system of each camera and the world coordinate system based on the checkerboard calibration plate and the pinhole imaging principle, thereby eliminating lens distortion. It also uses PTP protocol hardware synchronization signal to ensure the synchronization of data acquisition time of multiple cameras. The data preprocessing unit is used to remove noise points using a radius filtering statistical algorithm, retain effective surface data, and then synthesize a multi-frequency phase map by projecting phase shift patterns of different frequencies. The synthesis adopts point cloud stitching and registration by extracting SIFT key points in the point cloud and using the ICP iterative nearest point algorithm to achieve accurate stitching of multi-view point clouds. After stitching, the divergence-free wavelet technology is applied to process the original scan data to reconstruct a high-fidelity depth map. The 3D reconstruction optimization unit includes a triangular patch meshing subunit and an adaptive mesh subdivision subunit. The triangular patch meshing subunit is used to convert point cloud data into a triangular mesh model, retaining geometric features, and optimizing mesh noise through an edge-preserving smoothing algorithm while sharpening edge features. The adaptive mesh subdivision subunit is used to automatically refine the mesh in high curvature areas and simplify the mesh in low curvature areas, balancing data volume and accuracy.

4. The intelligent spraying system for enamel products according to claim 3, characterized in that, The defect detection and compensation submodule includes a thermal imaging and thickness sensing unit, a defect identification unit, a classification and evaluation unit, and a compensation decision unit. The thermal imaging and thickness sensing unit is used to monitor the temperature field distribution and thickness change of the coating in real time and identify areas of uneven spraying. The thermal imaging and thickness sensing unit consists of an infrared thermal imager and a laser thickness gauge. The infrared thermal imager detects the coating distribution based on the Seebeck effect, and the laser thickness gauge dynamically scans the film thickness through the triangulation principle. The defect identification unit is used to extract defect features from multi-source data, where the defect features are scratches, bubbles and color differences. The defect identification unit locates the defect area by combining threshold segmentation, morphological operations and grayscale histogram analysis. At the same time, it uses a U-Net network to perform semantic segmentation of the defect and outputs a pixel-level mask image. The classification and evaluation unit distinguishes defect categories based on LBP texture features and gradient features, classifies defect types, and evaluates mild or severe severity levels. The compensation decision unit is used to adjust the spraying flow rate, distance, and angle according to the defect type, and to automatically switch faulty nozzles. The compensation decision unit performs compensation in collaboration with AI through a rule base, increasing the paint flow rate when the coating is too thin, and adjusting the nozzle angle and reducing the spraying speed when there is a sagging defect. The flow rate formula is as follows: ,in This is for thickness deviation.

5. The intelligent spraying system for enamel products according to claim 4, characterized in that, The dynamic trajectory planning submodule includes a 3D model processing unit, an initial path planning unit, a multi-objective optimization unit, a real-time trajectory correction unit, an environmental coupling compensation unit, a process parameter matching unit, a zoned spraying strategy unit, and an execution control unit. The 3D model processing unit includes a geometric topology analysis subunit and a feature semantic segmentation unit. The geometric topology analysis subunit calculates vertex normal vectors based on the area weighted interpolation of adjacent triangular facets. Simultaneously, it combines curvature analysis to analyze the STL format 3D point cloud model generated by the vision module and extracts the topological relationships of the triangular mesh. The topological relationships include vertex, edge, and facet set. The feature semantic segmentation unit uses the U-Net deep learning network to segment the surface. At the same time, it uses multi-scale Gaussian filtering combined with principal curvature analysis to distinguish edges, corners, and planar regions, and divides the workpiece surface into functional regions, including inner walls, outer edges, and grooves, and outputs a spraying priority mask map. The initial path planning unit generates intersection trajectories based on the triangular mesh domain slicing algorithm, generates tool paths using the surface method, and supports hovering and detouring around marked obstacle areas, thereby generating a basic spraying path. The multi-objective optimization unit uses quadratic curves to model the thickness of adjacent trajectory superposition, dynamically adjusts the spray width, and optimizes the transition path between groups, thereby balancing spray uniformity, efficiency, and paint consumption. The real-time trajectory correction unit trains the robotic arm to avoid protruding parts, compensates for accumulated errors by combining IMU data, and dynamically adjusts the spray gun speed and distance according to the position deviation, thereby reducing the pose drift caused by mechanical vibration and temperature and humidity fluctuations. The environmental coupling compensation unit is used to monitor humidity changes, dynamically increase the spray gun voltage, and adjust the atomization pressure based on aerosol particle counter data, thereby offsetting the effect of high humidity environment on electrostatic adsorption force. The process parameter matching unit is used to associate the paint characteristic database with the spraying parameters. The process parameter matching unit optimizes the parameters through a weighted negative binomial distribution to adapt to different glaze characteristics, and simulates 100,000 working conditions on the digital twin platform to preview the parameter effects. The partitioned spraying strategy unit is used to dynamically adjust the spraying mode according to the surface characteristics, adjusting the spraying mode to a wide fan-shaped spray for flat workpieces and a narrow cone-shaped spray for curved workpieces; The execution control unit is used to convert the planned trajectory into six-axis robotic arm joint commands. The execution control unit calculates joint variables based on the target pose of the end effector and fine-tunes the contour spray gun to ensure continuous changes in speed and flow rate at corners and avoid coating accumulation. The execution control unit also controls the switching of heterogeneous nozzles through relays and automatically cleans the clogged unit after triggering.

6. The intelligent spraying system for enamel products according to claim 5, characterized in that, The real-time positioning and attitude calibration submodule includes a multimodal sensor array and a multi-coordinate system synchronization engine; The multimodal sensor array is used to collect workpiece spatial pose data in real time, covering six degrees of freedom information of position and angle. The multi-coordinate system synchronization engine is used to unify the visual coordinate system, the robot arm base coordinate system, and the workpiece coordinate system.

7. The intelligent spraying system for enamel products according to claim 6, characterized in that, The multimodal sensor array includes an inertial measurement unit (IMU), a visual target unit, and a laser rangefinder. The IMU includes a three-axis gyroscope and a three-axis accelerometer, and is used to output pitch, roll, and yaw angles in real time. The visual target unit consists of a 4×4 dot matrix reflective target and an infrared camera, and is used to identify the physical coordinates of the target. The laser rangefinder is used to assist in calibrating the Z-axis depth and compensate for blind spots. The multi-coordinate system synchronization engine solves the camera extrinsic parameters through the PnP algorithm, maps pixel coordinates to physical coordinates, and refreshes the pose data every 120Hz to ensure that the robot arm trajectory is synchronized with the actual pose of the workpiece.

8. The intelligent spraying system for enamel products according to claim 7, characterized in that, The intelligent environmental control module includes an environmental monitoring submodule, a temperature and humidity control submodule, a dehumidification unit, an air purification submodule, a real-time monitoring submodule, and a linkage control submodule. The environmental monitoring submodule is used to synchronously collect temperature and humidity data using distributed temperature and humidity sensors, detect dust concentration using the β-ray method, and monitor VOCs data using an electrochemical sensor. The temperature and humidity control submodule is used to dynamically adjust spraying parameters based on environmental data to counteract the effects of temperature and humidity fluctuations. The dehumidification unit is used to maintain a constant temperature and humidity in the spraying isolation room; The air purification submodule is used to capture spraying dust and degrade VOCs to ensure environmental safety. The real-time monitoring submodule is used to display visual environmental parameter curves and trigger audible and visual alarms when abnormalities occur. The linkage control submodule is used to link the vision guidance module and the spraying execution unit to achieve dynamic matching of environment and process.

9. The intelligent spraying system for enamel products according to claim 8, characterized in that, The auxiliary function module includes a human-computer interaction sub-module, a mobile collaboration platform, a full lifecycle database, and an intelligent analysis engine. The human-computer interaction submodule is a touch screen used to display real-time process parameter curves, defect heat maps and environmental monitoring data. The mobile collaboration platform is used to push alarm information and supports remote start / stop and policy adjustment. The full lifecycle database is used to store workpiece IDs, spraying parameters, defect records, and environmental data, and supports SQL queries and cloud synchronization. The intelligent analysis engine is used to calculate the mean and standard deviation of parameters using regression analysis to evaluate process stability, and to aggregate data from multiple production lines through a federated learning framework to optimize the generalization ability of the AI ​​model.

10. An enamel product, characterized in that, The enamel articles are processed according to any one of claims 1-9.

Images