Automatic batch aligning and positioning method and system based on visual AI

Abstract

The invention discloses an automatic batch aligning and positioning method and system based on visual AI, and relates to the field of automatic adjustment of printing gaps. The method comprises the following steps: acquiring a panoramic image of a platform to analyze the contour, coordinates and material attributes of a workpiece; dividing the surface of the workpiece into a printing area and a non-printing area; matching a height measurement mode according to material attributes to obtain height data; calculating a vertical avoidance gap according to the area height, the space distance and the nozzle size; and when the gap is smaller than the safety threshold value, the Z-axis height is adjusted to be the non-printing area height superposition safety margin, and otherwise, the preset printing height is maintained. By implementing the technical scheme provided by the invention, the printing definition is considered, and the collision risk is reduced.

Description

Technical Field

[0001] This application relates to the field of automatic printing gap adjustment, and in particular to an automatic batch alignment and positioning method and system based on visual AI. Background Technology

[0002] Currently, UV flatbed inkjet printing technology has been widely used in industrial manufacturing, advertising signage, and personalized customization due to its advantages such as fast curing speed, wide material adaptability, and high color reproduction. As the market's demand for production efficiency and customization continues to increase, printing platforms often need to handle large volumes of workpieces continuously.

[0003] In related technologies, to prevent physical collisions between the printhead and the workpiece during movement, a height measurement and obstacle avoidance system is typically configured on the printing equipment. This involves integrating laser displacement sensors or mechanical contact probes on the side of the printhead carriage. Before or during the printing operation, sensors scan or mark points on the workpiece on the platform to obtain its surface height data. Once the sensor detects that the height at a certain location exceeds the current printhead's set height, or scans the highest point on the entire surface, the Z-axis mechanism is controlled to raise the entire printhead to a safe height that can safely pass over that highest point, thus avoiding the risk of collision.

[0004] However, when printing on irregularly shaped workpieces involving mixed materials or complex surface structures, related technologies typically employ a global or row-wide uniform lifting strategy based on the highest detected point. When there are localized protrusions on the workpiece surface outside the printing area, or when encountering special materials such as transparent or highly reflective materials that cause distortion in the distance sensor data, a high-position printing strategy is often adopted to ensure safety. This results in an excessive distance between the printhead and the actual effective printing plane, leading to problems such as ink droplet scattering, droplet deviation, and image blurring. Therefore, when dealing with such complex workpieces, it is difficult to simultaneously ensure print clarity and mitigate printhead collision risks. Summary of the Invention

[0005] This application provides a method and system for automatic batch alignment and positioning based on visual AI, which balances print clarity and reduces the risk of collision.

[0006] The first aspect of this application provides an automatic batch alignment and positioning method based on visual AI, the method comprising: The system acquires a panoramic image of the printing platform, performs image segmentation and feature extraction to obtain the contour information, position coordinates, and surface material attribute category of the workpiece to be printed. Based on the contour information, the surface area of ​​the workpiece to be printed is divided into a printing area and a non-printing area. Based on the surface material attribute category, the corresponding height measurement mode is selected to obtain height data. Based on the height data, the height information of the non-printing area, the spatial distance between the printing area and the non-printing area, and the physical size parameters of the printhead assembly, the vertical clearance value of the printhead assembly relative to the non-printing area when operating on the printing area at a preset printing height is calculated. It is then determined whether the vertical clearance value is less than a preset safety threshold. If so, the Z-axis height during printing is adjusted to the height information of the non-printing area plus the preset safety margin, and a corresponding printing control command is generated. If not, the Z-axis height during printing is set to the preset printing height.

[0007] In the above embodiments, a spatial geometric relationship between the printhead and the non-printing obstacle area of ​​the workpiece is constructed through visual perception and zone height measurement technology. Based on the calculated vertical clearance, safe and dangerous operating conditions are distinguished: under the premise of ensuring no physical interference, the optimal low printing height is maintained first to ensure the accuracy of ink droplet landing; only when a collision risk is determined, the Z-axis is raised as needed for active avoidance. This avoids the image quality loss caused by blindly raising the printhead, thus achieving a balance between print clarity and reduced collision risk.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, based on height data, height information of the non-printing area, spatial distance between the printing area and the non-printing area, and physical dimension parameters of the printhead assembly, the vertical clearance value of the printhead assembly relative to the non-printing area when operating on the printing area at a preset printing height is calculated, specifically including: Based on surface material properties and contour information, the coefficient of thermal expansion of the workpiece to be printed is retrieved. If there are overhanging edge features in the non-printing area with geometric stiffness lower than the preset value, the preset thermodynamic deformation prediction model is invoked. Based on the coefficient of thermal expansion and the preset radiation power of the UV curing lamp, the dynamic thermal warpage increment of the overhanging edge features during the printing process is calculated. The dynamic thermal warpage increment is superimposed on the height information of the non-printing area to obtain the corrected non-printing area height. Based on the height data, the corrected non-printing area height, the spatial distance, and the physical dimension parameters, a dynamic spatial position relationship model between the nozzle assembly and the workpiece surface is constructed. Based on the dynamic spatial position relationship model, the minimum physical distance of the nozzle assembly relative to the corrected non-printing area height when operating on the printing area at the preset printing height is calculated, and the minimum physical distance is used as the vertical clearance value.

[0009] In the above embodiments, a thermodynamic deformation prediction mechanism is used to quantitatively calculate the dynamic warpage increment caused by thermal expansion during UV curing, based on material properties and structural characteristics. Combining static measurement with dynamic thermal simulation enables the anti-collision logic to predict the actual spatial shape changes of the workpiece after heating, reducing the potential for hysteresis interference caused by neglecting thermal deformation. This ensures that the nozzle maintains the optimal working distance while avoiding thermal deformation obstacles, achieving a dual guarantee of operational safety and image quality.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, after using the minimum physical distance as the vertical clearance value, the method further includes: If the vertical clearance value is less than the preset safety threshold and the corrected non-printing area height includes dynamic thermal warpage increment, a time-segmented UV power modulation strategy is generated based on the spatial offset of the UV curing lamp relative to the nozzle in the printhead assembly and the position coordinates of the overhanging edge features. Based on the time-segmented UV power modulation strategy, the dynamic thermal warpage increment is updated to zero or a preset low thermal deformation value, and the newly corrected non-printing area height and the new vertical clearance value are recalculated. If the new vertical clearance value is not less than the preset safety threshold, the Z-axis height is kept at the preset printing height, and the printing task is executed according to the time-segmented UV power modulation strategy.

[0011] In the above embodiments, by using a time-segmented UV power modulation strategy based on spatial location, the thermal energy input that causes thermal warping of the suspended edge of the workpiece is actively cut off from the source, suppressing the mechanism of eliminating obstacles by thermal deformation. There is no need to raise the nozzle to avoid potential warping risks, thereby avoiding thermal interference while always maintaining an ideal low spray height, achieving a balance between high-precision operation quality and equipment operation safety.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, after setting the Z-axis height during printing to a preset printing height, the method further includes: Based on the difference between the height information and the height data of the non-printing area, and the spatial distance between the printing area and the non-printing area, the cavity depth-to-width ratio of the printing area relative to the non-printing area is calculated; combined with the surface material property category, the density parameters of the workpiece to be printed are obtained; based on the cavity depth-to-width ratio characteristics and the preset scanning movement speed of the nozzle assembly, an aerodynamic flow field simulation model is constructed to calculate the maximum Bernoulli negative pressure suction force of the workpiece to be printed during the printing process; when the maximum Bernoulli negative pressure suction force is greater than the sum of the gravity of the workpiece to be printed and the platform adsorption force, the Z-axis height is kept at the preset printing height, and variable speed printing control commands are generated.

[0013] In the above embodiments, aerodynamic flow field analysis is used to predict the adsorption effect of Bernoulli negative pressure generated by high-speed scanning on lightweight workpieces. When the risk of workpiece floating is detected, a strategy of actively reducing the scanning speed is adopted to weaken the aerodynamic lift at the source. Speed ​​adjustment is used to resolve airflow disturbances, preventing accidental interference caused by workpiece suction and ensuring safe operation while maintaining the optimal close-range spray height.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, generating variable speed printing control instructions specifically includes: Based on the density parameters of the workpiece to be printed, the aspect ratio of the cavity, and the geometric span of the printing area, the first-order natural vibration frequency of the workpiece as a flexible thin film structure is calculated. Based on the target velocity value after the scanning movement speed is reduced, the aerodynamic excitation frequency generated when the airflow passes through the edge of the non-printing area is calculated. When the difference between the aerodynamic excitation frequency and the first-order natural vibration frequency is within the preset resonance risk range, the target velocity value is adjusted to a non-resonance velocity outside the resonance risk range, and a variable speed printing control command containing the non-resonance velocity is generated.

[0015] In the above embodiments, by avoiding the resonance range between aerodynamic excitation and the workpiece's natural frequency, the flutter instability of the flexible structure under high-speed airflow is eliminated. Speed ​​optimization is used to lock the workpiece shape, preventing accidental contact caused by resonance amplitude encroaching on the printing gap, ensuring that the workpiece surface remains stable while maintaining a tight spray distance.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, height data is obtained by selecting a corresponding height measurement mode based on the surface material property category, specifically including: Based on the image segmentation results, the boundary contour lines between different material attribute categories on the surface of the workpiece to be printed are identified, and a preset width is extended to both sides of the boundary contour lines to obtain an optically unstable region mask. The ranging sensor is controlled to scan and sample the printing area to obtain the original height point cloud data. Local point cloud data falling within the optically unstable region mask are extracted, and the height gradient change rate of the local point cloud data is calculated. If the height gradient change rate exceeds a preset physical abrupt change threshold, and the image segmentation results show that the region is visually a continuous plane, the local point cloud data is determined to be optical artifact noise. A bilinear interpolation algorithm is used to smoothly reconstruct the optical artifact noise using the effective height data outside the optically unstable region mask, generating corrected height data.

[0017] In the above embodiments, by constructing a dual verification of visual semantics and physical distance measurement, and utilizing the planar continuity features of the image, false height artifacts caused by optical abrupt changes at material interfaces are identified and eliminated, restoring the true topological structure of the workpiece surface. This avoids invalid lifting forced by sensor misreading data, ensuring that the printhead can always maintain optimal close-range operation based on real data, thereby improving the accuracy of ink droplet landing point while ensuring physical safety.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, after calculating the vertical clearance value relative to the non-printing area when the printhead assembly operates on the printing area at a preset printing height, the method further includes: Extract the surface gloss characteristics of the non-printing area. When determining whether the non-printing area belongs to a high specular reflection area, construct a geometric model of the UV light reflection path based on the installation angle, radiation cone angle, and vertical clearance value of the UV curing lamp on the printhead assembly. Based on the geometric model of the UV light reflection path, simulate and calculate the reflected light energy density reaching the printhead nozzle panel surface after reflection from the non-printing area. When the reflected light energy density exceeds the preset critical curing threshold of the UV ink, calculate the minimum optical attenuation distance required to attenuate the reflected light energy density below the critical curing threshold, and update the vertical clearance value to the larger of the minimum optical attenuation distance and the original vertical clearance value.

[0019] In the above embodiments, a reflection energy analysis model based on optical path tracing was established to quantify the light pollution threat posed by highly reflective areas to the nozzle panel. For the ink anti-solidification crust that may occur on the nozzle surface due to specular reflection, the minimum physical distance at which reflected light can be attenuated to a safe threshold was calculated. This prevents nozzle clogging caused by solidified deposits and the resulting risk of hard scratching, while also avoiding ink droplet scattering due to excessive lifting, achieving a balance between imaging sharpness and equipment operational safety.

[0020] In a second aspect, embodiments of this application provide an automatic batch alignment and positioning system, the automatic batch alignment and positioning system comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the automatic batch alignment and positioning system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on an automatic batch alignment and positioning system, cause the automatic batch alignment and positioning system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on an automatic batch alignment and positioning system, cause the automatic batch alignment and positioning system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the automatic batch alignment and positioning system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the automatic batch alignment and positioning method based on visual AI provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. This application utilizes visual perception and zone height measurement technology to construct the spatial geometric relationship between the printhead and the non-printing obstacle area of ​​the workpiece. Based on the calculated vertical clearance, it distinguishes between safe and hazardous operating conditions: prioritizing the maintenance of an optimal low printing height to ensure ink droplet accuracy while ensuring no physical interference; only when a collision risk is detected is the Z-axis raised as needed for active avoidance. This avoids the image quality loss caused by blindly raising the printhead, thus achieving a balance between print clarity and reduced collision risk.

[0025] 2. This application utilizes a thermodynamic deformation prediction mechanism to quantitatively calculate the dynamic warpage increment caused by thermal expansion during UV curing, based on material properties and structural characteristics. Combining static measurement with dynamic thermal simulation enables the anti-collision logic to predict the actual spatial shape changes of the workpiece after heating, reducing the potential for hysteresis interference caused by neglecting thermal deformation. This ensures that the nozzle maintains the optimal working distance while avoiding thermal deformation obstacles, achieving a dual guarantee of operational safety and image quality.

[0026] 3. This application proactively cuts off the thermal energy input that causes thermal warping of the suspended edge of the workpiece at the source by adopting a time-segmented UV power modulation strategy based on spatial location. It suppresses the mechanism of eliminating obstacles by suppressing thermal deformation, without having to raise the nozzle to avoid potential warping risks. Thus, while avoiding thermal interference, it always maintains an ideal low spray height, achieving a balance between high-precision operation quality and equipment operation safety. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating an automatic batch alignment and positioning method based on visual AI in an embodiment of this application. Figure 2 This is another flowchart illustrating the automatic batch alignment and positioning method based on visual AI in the embodiments of this application; Figure 3 This is an exemplary hardware structure diagram of the automatic batch alignment and positioning system in the embodiments of this application. Detailed Implementation

[0028] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0030] In related technologies, to prevent physical collisions between the printhead and the workpiece during movement, height measurement and obstacle avoidance systems are typically installed on the printing equipment. For example, laser displacement sensors or mechanical contact probes are integrated into the side of the printhead carriage. Before the printing job begins, sensors scan the workpiece on the platform to obtain height data. Once the highest point is detected, the Z-axis mechanism is controlled to raise the entire printhead to a safe height that can safely pass over that point. However, when dealing with multi-material mixed or irregularly shaped workpieces with complex surface structures, if there are local protrusions in the non-printing areas of the workpiece, or if transparent or highly reflective materials cause sensor data distortion, a conservative high-position printing strategy is often forced. While this approach avoids collisions, it results in an excessive distance between the printhead and the actual effective printing plane, leading to problems such as ink droplet scattering, droplet deviation, and image blurring. It is difficult to ensure print clarity while simultaneously guaranteeing printhead collision safety.

[0031] This application proposes an automatic batch alignment and positioning method based on visual AI. Through panoramic image acquisition and intelligent segmentation, the contour, position, and material properties of the workpiece are identified, and the workpiece surface is divided into printing and non-printing areas. Based on this, the accurate height is obtained by combining material characteristics, and a precise spatial geometric relationship is constructed between the printhead and the non-printing obstacle area to calculate the vertical clearance. Through this differentiated partitioning height measurement and judgment logic, safe and dangerous working conditions are distinguished: under the premise of ensuring no physical interference, the optimal low printing height is maintained first to ensure the accuracy of ink droplet landing; only when the vertical clearance is determined to be insufficient is the Z-axis raised as needed for active avoidance. This strategy avoids the image quality loss caused by blindly raising the printhead, thus achieving a technical effect that balances print clarity and reduces collision risk.

[0032] Figure 1 This is a flowchart illustrating the automatic batch alignment and positioning method based on visual AI used in the embodiments of this application, including the following steps: S101. Acquire panoramic images of the printing platform, and perform image segmentation and feature extraction on the panoramic images to obtain the contour information, position coordinates, and surface material attribute category of the workpiece to be printed.

[0033] The printing platform refers to the physical plane on which the industrial UV printer holds the workpiece to be printed, typically a suction platform or a glass platform; the panoramic image refers to a high-resolution digital image covering the entire effective working area of ​​the printing platform, with a field of view encompassing all workpieces to be printed; image segmentation refers to the process of dividing the pixels in the panoramic image into several regions with specific semantics (such as workpiece area and background area); feature extraction refers to the process of extracting recognizable geometric features (such as edges and corners) and texture features from the segmented image regions; contour information is used to represent the outer boundary shape data of the workpiece on a two-dimensional plane; position coordinates refer to the coordinates (X, Y) of the center point or key point of the workpiece in the printer coordinate system; surface material attribute category refers to the classification of the physical and optical properties of the workpiece surface, such as metal, wood, acrylic, glass, etc.

[0034] Specifically, the automated batch alignment and positioning system (hereinafter referred to as the system) controls a high-resolution industrial camera mounted above the printer beam or on a stand to take overhead images of the printing platform. After acquiring the images, the system processes them using a pre-trained deep convolutional neural network (such as the U-Net architecture). First, multi-scale features of the image are extracted through the backbone layer; then, a region proposal network (RPN) generates candidate boxes containing the workpieces; next, a classification branch identifies the material properties of the workpieces (e.g., distinguishing between metal and plastic by analyzing surface gloss and texture roughness), while a regression branch precisely locates the pixel-level mask of the workpieces. Finally, a coordinate mapping matrix is ​​used to convert the contour and position data in the image pixel coordinate system into the physical pulse coordinate data of the printer.

[0035] In some embodiments, this step can be implemented in several ways: Optionally, an RGB-D depth camera can be used to acquire images, and by combining color information with depth point cloud information, the spatial contour and material type of the workpiece can be directly obtained using a 3D point cloud segmentation algorithm (such as PointNet). It is understood that images can also be acquired by scanning the gantry with a line scan camera; this is not limited here.

[0036] S102. Based on the contour information, the surface area of ​​the workpiece to be printed is divided into a printing area and a non-printing area.

[0037] Among them, the surface area refers to the visible physical surface range of the workpiece from a top-down view; the printing area refers to the area on the workpiece surface that actually needs to be sprayed with ink according to the original printout (pattern); and the non-printing area refers to the blank areas on the workpiece surface area that do not need to be sprayed with ink, the areas blocked by edge clamps, or the areas with raised decorative structures that exist on the workpiece itself.

[0038] Specifically, the system first reads the uploaded pattern file to be printed and, based on the position coordinates and contour information obtained in S101, virtually maps the pattern onto the digital model of the workpiece. Next, the system executes Boolean operations: defining the physical contour set of the workpiece as set A, and defining the effective pixel coverage set of the printed pattern as set B. The printing area is the intersection of set A and set B (A∩B), while the non-printing area is the complement of set A minus set B (A−B), as well as any non-planar auxiliary structures identified within the workpiece contour (such as a ring buckle on a phone case, physical buttons on a panel, etc.). Through this division, the system can distinguish which areas are working surfaces and which are potential obstacle surfaces.

[0039] In some embodiments, this step can be implemented in a variety of ways: Optionally, by vector path analysis, the vector contour lines of the workpiece and the vector boundaries of the printed pattern are extracted, and the vector polygon descriptions of the printed area and the non-printed area are calculated using computational geometry algorithms (such as polygon clipping algorithms).

[0040] S103. Select the corresponding height measurement mode to obtain height data according to the surface material property category.

[0041] Here, "height measurement mode" refers to the specific combination of sensor operating parameters or measurement algorithm strategy used for surfaces with different optical properties; "height data" refers to the vertical distance of the workpiece surface relative to the Z-axis reference plane of the printing platform. "Matching" refers to matching based on a pre-established material-measurement strategy mapping table.

[0042] Specifically, the system dynamically calls the underlying driver interface based on the material category identified by S101. For example, for diffuse reflective materials (such as wood or paper), the system selects the triangulation laser ranging mode to quickly obtain the height using laser triangulation; for highly reflective materials (such as mirrored metal), conventional lasers will cause specular reflection, preventing the receiver from imaging, so the system switches to confocal spectral ranging mode or capacitive sensing ranging mode (if it is metal); for transparent materials (such as glass), optical sensors may penetrate directly, so the system switches to ultrasonic ranging mode or mechanical probe contact mode. The system controls the selected sensor to move above the printing area to perform single-point or multi-point sampling, and filters the sampled data (such as removing extreme values ​​and Gaussian smoothing) to finally obtain representative height data for that area.

[0043] In some embodiments, this step can be implemented in several ways: Optionally, a multi-sensor fusion module can be preset, and specific sensors can be activated according to the material type, such as eddy current sensor when metal is identified, and laser displacement sensor when plastic is identified. It is understood that a vision-based structured light height measurement method can also be used, adjusting the frequency and phase of the projection grating according to the reflective characteristics of the material, which is not limited here.

[0044] S104. Based on the height data, the height information of the non-printing area, the spatial distance between the printing area and the non-printing area, and the physical dimension parameters of the printhead assembly, calculate the vertical clearance value of the printhead assembly relative to the non-printing area when it operates on the printing area at a preset printing height.

[0045] Among them, the height information of the non-printing area refers to the Z-axis height of the non-printing area (i.e., potential obstacles) obtained by the height measurement sensor; the spatial distance refers to the Euclidean distance between the edge of the printing area and the edge of the non-printing area in the X / Y plane; the physical dimension parameters of the printhead assembly refer to the width of the printhead base plate, the tilt angle of the printhead side, and the installation position dimensions of auxiliary components such as UV lamps; the vertical clearance value is used to represent the predicted vertical margin between the lowest point of the printhead and the highest point of the non-printing area when the printhead assembly passes over the non-printing area on its movement trajectory; the preset printing height refers to the optimal distance between the printhead nozzle and the printing surface required by the process specification to ensure printing quality, and this value is experimentally determined and stored in the system configuration file.

[0046] Specifically, the system constructs a three-dimensional virtual geometric space, abstracting the printhead assembly as a moving body with specific length, width, height, and base shape, and abstracting the printing and non-printing areas of the workpiece as stepped surfaces of different heights. Assuming the printhead is operating in the printing area, its absolute Z-axis coordinate is Z... head =H print_zone +H preset When the printhead moves to print edge pixels, the physical housing (such as a large base plate) may extend over the non-printing area. The system calculates the difference in the Z-axis direction between the equation of the printhead base plate plane and the equation of the top surface of the non-printing area. Specifically, if the printhead assembly has a bevel (chamfer), the system also considers spatial distance to calculate whether the printhead sidewalls will interfere with the sidewalls of the non-printing area. The final calculated vertical clearance value is G=Z. head -H non_print_zone This value may be negative, indicating that a theoretical collision exists.

[0047] In some embodiments, this step can be implemented in several ways: Optionally, a voxelized collision model can be established, and a Boolean intersection operation can be performed on the volume swept by the nozzle motion trajectory and the volume of the non-printing area to calculate the overlap depth as the negative gap value; Optionally, a simplified 2D section projection method can be used to calculate the minimum perpendicular distance between the projected profile of the nozzle assembly in the scanning direction and the profile of the workpiece surface. It is understood that a fast detection algorithm based on bounding boxes can also be used to calculate the distance between the bottom surface of the nozzle bounding box and the top surface of the non-printing area bounding box, which is not limited here.

[0048] In other embodiments, in special scenarios where the workpiece to be printed is in a low-humidity environment and the non-printing area is made of a highly insulating, high-static-electrostatic material (such as a large area of ​​acrylic or PVC film), the automatic batch alignment and positioning system can also perform an electrostatic breakdown risk assessment step. Specifically, because the surface of a high-static-electrostatic material may accumulate electrostatic charges of up to several thousand volts, even if the printhead is not physically in contact with the workpiece, an excessively small gap may induce air ionization and generate an electric arc, breaking down the delicate piezoelectric crystal circuit inside the printhead. In this scenario, the system introduces an electrostatic safety distance parameter when calculating the vertical clearance value. The system estimates the surface potential based on ambient humidity sensor data and material properties, and uses Paschen's law to calculate the air breakdown distance under the current environment. At this time, the calculated vertical clearance value is no longer just the geometric distance in physical space, but the effective electrical safety clearance after subtracting the air breakdown distance from the physical clearance. If there is no physical collision but electrical breakdown is possible, the system will still determine that the clearance is insufficient, thereby forcibly raising the Z-axis in subsequent steps to ensure both physical and electrical safety.

[0049] In other embodiments, in scenarios where the non-printing area is made of a highly specular reflective material (such as metal clamps or chrome-plated decorative parts), the automatic batch alignment and positioning system can also perform anti-curing backlash calculation steps based on optical path simulation. By constructing a UV light reflection model, it assesses and eliminates the risk of nozzle clogging caused by reflected light, thereby achieving optical protection on the basis of physical collision protection.

[0050] Specifically, the system first uses a visual sensor to extract the surface gloss characteristics of the non-printing area (e.g., by analyzing the pixel brightness distribution histogram of the highlight area). If it is determined to be a high specular reflection area, the optical path simulation module is activated. The system reads the CAD geometric parameters of the printhead assembly, including the installation angle of the UV curing lamp relative to the nozzle panel, the radiation cone angle of the UV lamp, and the currently calculated vertical clearance value. Based on the principles of geometric optics, the system constructs a geometric model of the UV light reflection path: assuming a beam of edge light emitted by the UV lamp strikes the surface of the non-printing area at an incident angle, according to the law of reflection, the reflection angle equals the incident angle, and the light will be reflected back to the printhead direction at the same angle. The system tracks the path of this reflected light and determines whether it will fall within the effective area of ​​the nozzle panel. If the light falls in, the system further applies the principles of radiometry and the inverse square law of distance to simulate and calculate the reflected light energy density reaching the nozzle surface. The preset critical curing threshold for UV ink refers to the minimum ultraviolet energy density required for the ink to undergo a gelation reaction; this value is usually provided by the ink supplier or determined through rheological experiments. If the simulated reflected light energy density exceeds the threshold, it means that even without a physical collision, the intense reflected ultraviolet light will instantly solidify the minute amount of ink mist on the nozzle surface, forming an ink column that causes blockage. In this case, the system solves the optical path equation in reverse to calculate the minimum distance required for the reflected light energy density to attenuate below the safe threshold, i.e., the minimum optical attenuation distance. The system compares this distance with the original physical clearance value, takes the larger value as the updated vertical clearance value, and forcibly increases the Z-axis distance, utilizing the attenuation effect of space to dilute the harmful reflected light energy.

[0051] The aforementioned technical steps prevent nozzle clogging and damage caused by photocuring backlash. By introducing optical-level safety checks, the lifespan of expensive nozzles is extended, and continuous and stable production capabilities are ensured in complex, highly reflective workpiece environments.

[0052] S105. Determine whether the vertical clearance value is less than the preset safety threshold. If so, proceed with step S106 below; If not, proceed to step S107 below.

[0053] The preset safety threshold refers to the minimum safe buffer distance reserved to accommodate mechanical transmission errors, workpiece flatness errors, and sensor measurement errors. This value is usually set by those skilled in the art based on the accuracy level of the equipment.

[0054] Specifically, the system compares the vertical clearance value calculated in S104 with a preset safety threshold. If the calculated clearance value is less than the safety threshold, it means there is an extremely high risk of physical collision or electrical breakdown, and the current state is determined to be unsafe, with the logic flow proceeding to S106 for intervention; conversely, if the clearance value is greater than or equal to the safety threshold, it means that it is safe for the printhead to pass through the non-printing area at the current height, and the logic flow proceeds to S107 to maintain the original state.

[0055] S106. Adjust the Z-axis height during printing to the height information of the non-printing area plus the preset safety margin, and generate the corresponding printing control command.

[0056] The preset safety margin refers to an additional vertical distance above the highest point of the non-printing area to ensure absolute safety during passage; the print control command refers to the underlying code that drives the printer's motion control card to execute the Z-axis motor movement.

[0057] Specifically, when a risk is detected, the system abandons the original optimal printing height and prioritizes ensuring safety. The system calculates the new target Z-axis height Z. new =H non_print_zone +Margin safety This adjustment means the printhead will be raised to a position higher than obstacles in the non-printing area for printing. While this increases the distance between the printhead and the printing area (i.e., increases the jetting distance), potentially causing ink droplet scattering (ink splatter) and a slight decrease in clarity, it avoids printhead collisions. The system will calculate Z... new The control data package is encapsulated in the printing task, or commands are sent in real time to control the rotation of the Z-axis motor, driving the crossbeam or platform to rise and fall to the safe height.

[0058] In some embodiments, this step can be implemented in several ways: optionally, a global lift command is generated to maintain the safe height throughout the printing process; alternatively, a dynamic jump command is generated to lift the Z-axis only when the printhead moves above the non-printing area and lower the Z-axis when it returns to the printing area and space permits. It is understood that a variable speed command can also be generated in conjunction with height adjustment to reduce the carriage scanning speed while raising the printhead, in order to reduce ink droplet droplet deviation caused by the increased jetting distance; this is not limited here.

[0059] S107. Set the Z-axis height during printing to the preset printing height.

[0060] The preset printing height refers to the printhead height that achieves optimal ink droplet accuracy and color saturation under ideal conditions without any obstructions, and is set in advance by the operator.

[0061] Specifically, when S105 determines there is no collision risk, the system does not need to perform additional avoidance lifting on the Z-axis. It directly reads the standard print height parameters in the configuration file and, combined with the actual surface height of the print area measured by S103, calculates the absolute coordinate position of the Z-axis. The system generates standard print control commands to keep the printhead close to the workpiece surface, thereby ensuring that the ink droplets reach the media surface with the shortest flight path, achieving high-definition printing results.

[0062] In some embodiments, this step can be implemented in several ways: Optionally, a closed-loop control method can be used to monitor the height in real time during printing and fine-tune the Z-axis to maintain a constant preset printing height; alternatively, the Z-axis can be positioned to the preset printing height and locked once before printing begins. It is understood that the preset printing height can also be fine-tuned according to the image resolution requirements, for example, automatically reducing the height by 0.1mm (within a safe range) when printing high-precision text, which is not limited here.

[0063] In some embodiments, when the workpiece to be printed is a lightweight thin sheet with a deep cavity structure, the automatic batch alignment and positioning system can also perform anti-suction and anti-resonance calculation steps based on fluid-structure interaction analysis. By adjusting the scanning speed, the system can eliminate the risks of pneumatic negative pressure adsorption and structural resonance, thereby achieving safe printing while ensuring that the Z-axis height remains unchanged.

[0064] Specifically, when the printhead sweeps across the recessed printing area at high speed, according to Bernoulli's principle, the increased flow velocity leads to a decrease in pressure, thus generating an upward suction force on the workpiece surface. First, the cavity depth-to-width ratio (i.e., the ratio of the height difference of the non-printing area to the width of the printing area) is calculated; this characteristic determines the acceleration of the airflow within the cavity. Simultaneously, the system retrieves density parameters from a database based on the material type. Next, the system constructs a simplified aerodynamic flow field simulation model (e.g., a simplified solution based on potential flow theory or the Navier-Stokes equations), inputs the current preset scanning speed, and calculates the maximum Bernoulli negative pressure suction force on the workpiece. If this suction force is greater than the sum of the workpiece's own weight and the platform's adsorption force (e.g., vacuum adsorption force), it indicates a risk of the workpiece being sucked away and impacting the printhead. In this case, the system does not change the Z-axis height (to maintain print clarity) but instead generates a variable-speed printing control command to actively reduce the scanning speed, because dynamic pressure is proportional to the square of the speed, and reducing the speed significantly reduces the suction force.

[0065] Furthermore, to prevent the reduced airflow frequency from unexpectedly inducing workpiece resonance, the system performs vibration modal analysis. The system treats the workpiece as a flexible membrane with fixed or simply supported ends, and calculates the first-order natural vibration frequency using thin-plate vibration theory formulas based on density, aspect ratio, and geometric span (printed area length). Simultaneously, the airflow passing through the edge of the non-printed area (at abrupt step changes) generates periodic vortex shedding, the frequency of which is the aerodynamic excitation frequency, proportional to the flow velocity (Strauhal number relationship). The preset resonance risk range refers to the frequency band near the natural frequency, typically set based on structural dynamics principles. If the calculated excitation frequency falls within this range, the system will fine-tune the target velocity to deviate from the resonance point, generating the final non-resonant velocity command.

[0066] The above technical steps, through dual verification by fluid mechanics and structural dynamics, mitigate the risk of workpieces being sucked up and colliding or ripples caused by resonance by using speed adjustment without sacrificing printing height (i.e., without sacrificing clarity), thus achieving a balance between printing efficiency, quality and safety.

[0067] In the above embodiments, a spatial geometric relationship between the printhead and the non-printing obstacle area of ​​the workpiece is constructed through visual perception and zone height measurement technology. Based on the calculated vertical clearance, safe and dangerous operating conditions are distinguished: under the premise of ensuring no physical interference, the optimal low printing height is maintained first to ensure the accuracy of ink droplet landing; only when a collision risk is determined, the Z-axis is raised as needed for active avoidance. This avoids the image quality loss caused by blindly raising the printhead, thus achieving a balance between print clarity and reduced collision risk.

[0068] In other embodiments of this application, when printing workpieces made of heat-sensitive materials with overhanging edges, the continuous thermal radiation from the UV lamp may cause the workpiece to warp dynamically due to heat, thereby encroaching on the originally safe clearance and causing a collision. The automatic batch alignment and positioning method based on visual AI provided in this application can predict the amount of thermal deformation based on the material's thermal expansion coefficient and structural stiffness, and dynamically correct the height of the non-printing area or modulate the UV power accordingly, thereby avoiding the risk of thermally induced collisions.

[0069] like Figure 2 The diagram shown is another flowchart illustrating the automatic batch alignment and positioning method based on visual AI provided in this application embodiment, including the following steps: S201. Acquire panoramic images of the printing platform, and perform image segmentation and feature extraction on the panoramic images to obtain the contour information, position coordinates, and surface material attribute category of the workpiece to be printed.

[0070] S202. Based on the contour information, the surface area of ​​the workpiece to be printed is divided into a printing area and a non-printing area.

[0071] S203. Select the corresponding height measurement mode to obtain height data according to the surface material property category.

[0072] In some embodiments, in special scenarios where multiple materials are spliced ​​on the surface of the workpiece to be printed and optical properties change abruptly at the interface, the automatic batch alignment and positioning system can also perform artifact removal and reconstruction steps based on multimodal data fusion, and correct abnormal data of the ranging sensor through visual semantic constraints, thereby obtaining true and accurate height information.

[0073] Specifically, when a laser beam sweeps across the boundary between two materials with different reflectivities (such as alternating black and white) or different transparency (such as a glass and metal composite), the echo signal often becomes distorted, causing false and drastic jumps in the measured height value. The system first uses the high-precision visual image segmentation results in S201 to locate the boundary contour line between different material property categories. Considering the sensor spot's certain diameter and mechanical motion errors, the system extends a preset width (e.g., 1.5 times the spot diameter) outwards from this contour line to generate a strip-shaped optically unstable region mask.

[0074] Next, the system controls a ranging sensor (such as a laser displacement sensor) to perform a full-coverage scan of the printing area, acquiring raw height point cloud data. The system spatially aligns the point cloud data with the mask, extracts local point cloud data falling within the mask's range, and calculates the height gradient change rate (i.e., the ratio of the height difference to the distance between adjacent points). If the calculated change rate is extremely large, exceeding a preset physical abrupt change threshold (this threshold is set based on the limits of the workpiece processing technology; for example, a 90-degree micron-level cliff is impossible on the surface of an injection-molded part), it indicates data abnormality. At this point, the system introduces visual semantics for arbitration: checking the corresponding image segmentation results, if the visual appearance shows that the texture of the area is continuous and without breaks or step features (i.e., it belongs to a continuous plane), then the abrupt change data is optical artifact noise.

[0075] To recreate the true surface, the system removes noise points and employs either bilinear interpolation or Gaussian process regression. Using the effective height data from the outside of the mask (i.e., the area with uniform material and accurate measurement) as boundary conditions, the voids inside the mask are mathematically fitted and smoothly reconstructed, ultimately generating the corrected height data.

[0076] The above technical steps correct the physical error of laser ranging by using the semantic information of visual images, avoiding the miscalculation of extremely small clearances or false collision risks due to sensor misreading, ensuring the authenticity and reliability of height data, and laying the foundation for subsequent Z-axis control.

[0077] Steps S201-S203 and Figure 1Steps S101-S103 in the illustrated embodiment are similar and can be found in the descriptions of steps S101-S103, which will not be repeated here.

[0078] S204. Based on the surface material property category and contour information, retrieve the coefficient of thermal expansion of the workpiece to be printed.

[0079] The coefficient of thermal expansion refers to the relative change in length or volume of a substance when the temperature increases by 1 K due to thermal expansion and contraction. It is usually expressed in units of 10. −6 / K. Searching refers to the operation of finding and matching materials in the system's built-in material property database.

[0080] Specifically, during the UV printing process, the UV lamp generates a large amount of accompanying heat (infrared radiation), and the volume change of different materials after heating varies greatly. The system uses the material category identified by S201 (such as polycarbonate PC, aluminum alloy, ABS plastic) as the primary index key, and combines the contour information obtained by S201 (used to assist in judgment, such as certain specific contours corresponding to specific grades of industrial standard parts) as the secondary index key to query the corresponding linear thermal expansion coefficient in the local or cloud database.

[0081] In some embodiments, this step can be implemented in several ways: Optionally, a local SQLite database can be built to store ISO standard property tables of common industrial materials, which can be directly obtained through SQL query statements.

[0082] S205. If there are overhanging edge features in the non-printing area with geometric stiffness lower than the preset value, call the preset thermodynamic deformation prediction model, and calculate the dynamic thermal warping increment of the overhanging edge features during the printing process based on the coefficient of thermal expansion and the preset radiation power of the UV curing lamp.

[0083] Among them, geometric stiffness refers to the structure's ability to resist deformation, which depends on the elastic modulus of the material and the moment of inertia of the structure's cross section; cantilever edge characteristics refer to thin-walled or slender structures in the workpiece that are not adsorbed and fixed by the printing platform and are in a free cantilever state (such as the edge of the button hole of a mobile phone case, or the unadsorbed corner of a film); preset value refers to the stiffness threshold set based on experience, below which the structure is considered to be prone to thermal deformation; thermodynamic deformation prediction model refers to a set of mathematical equations or a finite element simulation model constructed based on thermoelasticity theory; preset radiation power of UV curing lamp refers to the energy output intensity of the UV lamp when it is working; dynamic thermal warpage increment refers to the upward displacement of the workpiece in the Z-axis direction after it is heated.

[0084] Specifically, the system first analyzes the workpiece's geometric topology, identifying areas without bottom support or with weak adhesion (i.e., overhanging edges). If such areas exist, the system initiates thermodynamic calculations. The system inputs the retrieved coefficient of thermal expansion, the UV lamp's radiant power (converted to heat flux input), and the workpiece's thickness parameters. Based on the theory of non-uniform temperature fields, the model calculates the temperature gradient between the workpiece's upper surface (the illuminated surface) and lower surface (the unilluminated surface). Due to the higher temperature and greater expansion on the upper surface, and the lower temperature and smaller expansion on the lower surface, this non-uniform expansion generates a thermal bending moment, causing the overhanging edge to curl upwards. The system calculates the maximum Z-axis displacement that this edge may reach during the printing process through integration, which is the dynamic thermal warpage increment.

[0085] In some embodiments, this step can be implemented in several ways: Optionally, a simplified bimetallic sheet bending formula can be used, treating the workpiece as a heated beam and directly substituting the parameters to calculate the maximum deflection; alternatively, a lightweight finite element analysis (FEA) solver can be run to mesh the suspended region, apply temperature loads, and solve for the nodal displacement vectors. It is understood that a data-driven approach can also be used, utilizing a regression model trained on historical printing data, inputting material and thickness to directly predict the warpage; this is not limited here.

[0086] In some embodiments, in special scenarios where the workpiece to be printed is a hygroscopic polymer material (such as nylon PA6 or wood fiberboard) and the ambient humidity changes drastically, the automatic batch alignment and positioning system can also perform a hygrothermal coupling deformation compensation step. Specifically, such materials not only expand when heated, but also expand in volume after absorbing moisture (hygroscopic expansion). When UV lamp heating causes rapid evaporation of local moisture, a complex superposition effect of wet stress and thermal stress is generated, resulting in deformation far exceeding the simple prediction of thermal expansion. In this scenario, the system reads ambient humidity sensor data and calls the hygrothermal coupling constitutive model. When calculating the dynamic thermal warpage increment, this model adds an additional term for the hygroscopic warpage component caused by the humidity gradient and considers the effect of temperature rise on the material's hygroscopic diffusion coefficient. The system vectorically superimposes the thermal deformation and the hygroscopic deformation to calculate a more extreme total warpage, thereby guiding subsequent avoidance calculations and preventing accidental collisions caused by ignoring humidity factors.

[0087] S206. The dynamic thermal warpage increment is superimposed on the height information of the non-printing area to obtain the corrected height of the non-printing area.

[0088] Among them, the corrected non-printing area height refers to the predicted absolute coordinates of the top surface of the non-printing area on the Z-axis after taking into account thermal deformation factors.

[0089] Specifically, the height data measured in S203 is the static height of the workpiece in a cold state (room temperature). However, printing is a dynamic heating process, and the workpiece height changes over time. The system will use the positive (upward) dynamic thermal warpage increment ΔZ calculated in S205. warp Add to the original measured non-printing area height H static Above, i.e., H corrected =H static +ΔZ warp This correction value represents the highest position the obstacle can reach in the worst-case scenario (i.e., when thermal deformation is at its maximum).

[0090] In some embodiments, this step can be implemented in several ways: Optionally, a scalar addition operation can be performed directly, adding the maximum predicted warpage value to the entire non-printing area height as a conservative estimate; alternatively, a region mapping overlay can be performed, overlaying the warpage increment only in the identified overhanging edge regions, while keeping the original height unchanged in the well-adhesive central regions, generating a fine height distribution map. It is understood that a time factor can also be introduced, linearly increasing the warpage amount according to the printing progress to simulate the heat accumulation process; this is not limited here.

[0091] S207. Based on the height data, the height of the non-printing area, the spatial distance, and the physical dimension parameters are corrected to construct a dynamic spatial position relationship model between the nozzle assembly and the workpiece surface.

[0092] Among them, the dynamic spatial position relationship model refers to a three-dimensional geometric scene description that describes the relative position of the nozzle and the workpiece, which changes with time or printing progress.

[0093] Specifically, the system no longer treats the workpiece as a static rigid body, but rather as a dynamic surface whose height varies with position and temperature (defined by the corrected height in S206). Simultaneously, based on the nozzle's motion trajectory planning, the system determines the spatial coordinates (X(t), Y(t), Z) of the nozzle at any given time t. head The model calculates or simulates in real time the Euclidean distance field between the nozzle enclosure and the dynamic curved surface of the workpiece. This model not only includes static geometric dimensions but also implicitly includes the growth trend of the workpiece due to thermal deformation.

[0094] In some embodiments, this step can be implemented in several ways: Optionally, a time-step-based discretized model can be constructed, dividing the printing process into several time slices, updating the workpiece height and checking the nozzle position in each time slice; alternatively, a parametric surface model can be constructed, fitting the workpiece surface with a NURBS surface, describing the nozzle motion with parametric equations, and analyzing the positional relationships by solving the surface intersection problem. It is understood that GPU-accelerated collision detection libraries can also be used to render invisible virtual scenes in real time for physical simulation, which is not limited here.

[0095] S208. Based on the dynamic spatial position relationship model, calculate the minimum physical distance of the printhead assembly relative to the height of the corrected non-printing area when it operates on the printing area at a preset printing height, and use the minimum physical distance as the vertical clearance value.

[0096] The minimum physical distance refers to the minimum Euclidean distance that may occur between any point on the surface of the printhead assembly and any point on the workpiece correction surface during the entire printing and scanning process.

[0097] Specifically, the system performs traversal search or optimization solutions within a dynamic spatial positional relationship model. Assuming the printhead operates at an ideal preset printing height, the system simulates the printhead's movement along the scanning path. When passing over non-printing areas (especially overhanging edges predicted to experience thermal warping), the system calculates the distance between the printhead's bottom surface and the corrected height H at that location. corrected The system records the minimum value along the entire path. This minimum value is the actual vertical clearance value after considering the risk of thermal deformation. If this value is negative, it means that although there is no problem in the cold state, the workpiece will tilt up and hit the nozzle in the hot state.

[0098] In some embodiments, this step can be implemented in several ways: Optionally, a path sampling method can be used, sampling a point at regular intervals along the nozzle's movement path, calculating the perpendicular distance between the nozzle and the workpiece at that point, and taking the minimum value of all sampled points; Optionally, an analytical geometry method can be used to calculate the algebraic distance between the nozzle's bottom plane equation and the highest point of the workpiece (after correction). It is understood that probabilistic analysis can also be combined to calculate the collision probability within the range of thermal expansion coefficient error, and the minimum distance satisfying a 99.9% confidence level can be taken as the result; this is not limited here.

[0099] In other embodiments, in special cases where a potential risk of thermal warping collision is detected and that the risk originates from the thermal radiation of the UV lamp, the automatic batch alignment and positioning system can also perform a UV power dynamic modulation step based on spatial location to eliminate thermal deformation by suppressing heat input at the source, thereby achieving safe printing without raising the Z-axis.

[0100] Specifically, when the system determines that the vertical clearance is insufficient and the cause is the dynamic thermal warping increment predicted by S205, it no longer passively raises the Z-axis but instead controls the heat source. The system first reads the geometric parameters of the nozzle assembly to determine the spatial offset of the UV curing lamp relative to the nozzle (i.e., the distance the UV lamp lags behind or leads the nozzle in the X-axis direction). Combining this with the position coordinates of the overhanging edge features, the system calculates the precise time window for the UV lamp to sweep across the sensitive area. Based on this, the system generates a time-segmented UV power modulation strategy. This strategy is a set of timing control commands: when the nozzle is printing a pattern near the overhanging edge, the UV lamp maintains normal power; however, when the nozzle moves away and the UV lamp is about to move above the overhanging edge, the system immediately reduces the UV lamp's radiant power to a preset low level or directly turns it off. This cuts off the heat source input that causes material warping.

[0101] Subsequently, the system reassessed the risk based on this strategy. Since the heat source was cut off, the previously predicted severe thermal warpage would no longer occur. The system updated the dynamic thermal warpage increment in the model to zero (or a very small residual thermal deformation value) and calculated a new corrected non-printing area height (which is now almost equal to the static height in the cold state). Next, the system again substituted the model from S207 to calculate a new vertical clearance value. If this new clearance value meets the safety threshold requirements, it means that as long as the lights are properly controlled, there is no longer a physical risk of collision. At this point, the system locked the Z-axis height to the optimal preset printing height and issued the final printing command, which included power modulation logic.

[0102] The aforementioned technical steps, through regional power control, suppress the thermal deformation tendency of the workpiece and eliminate potential collision obstacles. This eliminates the need for the printhead to be forced to rise as a whole to avoid potentially warped edges, thus maintaining optimal spraying distance and ensuring print clarity and droplet accuracy in edge areas.

[0103] S209. Determine whether the vertical clearance value is less than the preset safety threshold. If so, proceed with step S210 below; If not, proceed to step S211 below.

[0104] S210. Adjust the Z-axis height during printing to the height information of the non-printing area plus the preset safety margin, and generate the corresponding printing control command.

[0105] S211. Set the Z-axis height during printing to the preset printing height.

[0106] Steps S209-S211 and Figure 1 Steps S105-S107 in the illustrated embodiment are similar and can be found in the descriptions of steps S105-S107, which will not be repeated here.

[0107] In the above embodiments, by introducing a thermodynamic deformation prediction mechanism, the dynamic warpage increment caused by thermal expansion during UV curing is quantitatively calculated based on the workpiece material properties and suspended structure characteristics. Combining static measurement with dynamic thermal simulation enables the anti-collision logic to predict the actual spatial shape of the workpiece after heating. This eliminates the risk of delayed collisions caused by neglecting thermally induced deformation, achieving dual protection of physical safety and printing accuracy when handling thin-walled or heat-sensitive irregularly shaped workpieces.

[0108] The following describes an exemplary automatic batch alignment and positioning system 300 provided in an embodiment of this application. Figure 3 This is an exemplary hardware structure diagram of the automatic batch alignment and positioning system 300 provided in the embodiments of this application.

[0109] In some embodiments, the automatic batch alignment and positioning system 300 is a computer device or includes a computer device. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device stores data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements the methods in the embodiments of this application.

[0110] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0111] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0112] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0113] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0114] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An automatic batch alignment and positioning method based on visual AI, characterized in that, include: A panoramic image of the printing platform is acquired, and the panoramic image is segmented and feature extracted to obtain the contour information, position coordinates and surface material attribute category of the workpiece to be printed. Based on the contour information, the surface area of ​​the workpiece to be printed is divided into a printing area and a non-printing area; Based on the surface material property category, select the corresponding height measurement mode to obtain height data; Based on the height data, the height information of the non-printing area, the spatial distance between the printing area and the non-printing area, and the physical size parameters of the printhead assembly, calculate the vertical clearance value of the printhead assembly relative to the non-printing area when it operates on the printing area at a preset printing height. Determine whether the vertical clearance value is less than a preset safety threshold; If so, the Z-axis height during printing will be adjusted to the height information of the non-printing area plus a preset safety margin, and a corresponding printing control command will be generated. If not, then set the Z-axis height during printing to the preset printing height.

2. The method according to claim 1, characterized in that, The step of calculating the vertical clearance value of the printhead assembly relative to the non-printing area when the printhead assembly operates on the printing area at a preset printing height, based on the height data, the height information of the non-printing area, the spatial distance between the printing area and the non-printing area, and the physical size parameters of the printhead assembly, specifically includes: Based on the surface material property category and the contour information, the coefficient of thermal expansion of the workpiece to be printed is retrieved; If there are overhanging edge features in the non-printing area with geometric stiffness lower than the preset value, the preset thermodynamic deformation prediction model is invoked to calculate the dynamic thermal warping increment of the overhanging edge features during the printing process based on the thermal expansion coefficient and the preset radiation power of the UV curing lamp. The dynamic thermal warpage increment is superimposed on the height information of the non-printing area to obtain the corrected non-printing area height. Based on the height data, the corrected non-printing area height, the spatial distance, and the physical dimension parameters, a dynamic spatial position relationship model between the nozzle assembly and the workpiece surface is constructed. Based on the dynamic spatial position relationship model, the minimum physical distance of the printhead assembly relative to the height of the corrected non-printing area when it operates on the printing area at a preset printing height is calculated, and the minimum physical distance is used as the vertical clearance value.

3. The method according to claim 2, characterized in that, After setting the minimum physical distance as the vertical clearance value, the method further includes: If the vertical clearance value is less than a preset safety threshold and the corrected non-printing area height includes the dynamic thermal warpage increment, then a time-segmented UV power modulation strategy is generated based on the spatial offset of the UV curing lamp relative to the nozzle in the printhead assembly and the position coordinates of the overhanging edge feature; the strategy is configured to reduce or turn off the radiation power when the UV curing lamp passes over the overhanging edge feature. Based on the time-segmented UV power modulation strategy, the dynamic thermal warpage increment is updated to zero or a preset low thermal deformation value, and the newly corrected non-printing area height and the new vertical clearance value are recalculated. If the new vertical clearance value is not less than the preset safety threshold, the Z-axis height is kept at the preset printing height, and the printing task is executed according to the time-segmented UV power modulation strategy.

4. The method according to claim 1, characterized in that, After setting the Z-axis height during printing to the preset printing height, the method further includes: Based on the difference between the height information of the non-printing area and the height data, and the spatial distance between the printing area and the non-printing area, the cavity depth-to-width ratio feature of the printing area relative to the non-printing area is calculated; Based on the surface material property category, the density parameter of the workpiece to be printed is obtained; Based on the cavity aspect ratio characteristics and the preset scanning movement speed of the nozzle assembly, an aerodynamic flow field simulation model is constructed to calculate the maximum Bernoulli negative pressure suction force of the workpiece to be printed during the printing process. When the maximum Bernoulli negative pressure suction force is greater than the sum of the weight of the workpiece to be printed and the platform's adsorption force, the Z-axis height is maintained at the preset printing height, and a variable speed printing control command is generated; the variable speed printing control command is to reduce the scanning movement speed so that the maximum Bernoulli negative pressure suction force is reduced to a preset safe range.

5. The method according to claim 4, characterized in that, The generation of variable speed printing control commands specifically includes: Based on the density parameters of the workpiece to be printed, the aspect ratio characteristics of the cavity, and the geometric span of the printing area, the first-order natural vibration frequency of the workpiece to be printed as a flexible thin film structure is calculated. Based on the target velocity value after the scanning movement speed is reduced, calculate the aerodynamic excitation frequency generated when the airflow passes through the edge of the non-printing area. When the difference between the aerodynamic excitation frequency and the first-order natural vibration frequency is within a preset resonance risk range, the target speed value is adjusted to a non-resonance speed outside the resonance risk range, and a variable speed printing control command containing the non-resonance speed is generated.

6. The method according to claim 1, characterized in that, The step of selecting the corresponding height measurement mode to obtain height data based on the surface material property category specifically includes: Based on the image segmentation results, the boundary contour lines between different material property categories on the surface of the workpiece to be printed are identified, and a preset width is extended to both sides with the boundary contour lines as the center to obtain an optically unstable region mask. The distance measuring sensor is controlled to scan and sample the printed area to obtain the original height point cloud data; Extract local point cloud data that falls within the optically unstable region mask, and calculate the height gradient rate of change of the local point cloud data; If the height gradient change rate exceeds a preset physical abrupt change threshold, and the image segmentation result shows that the region visually belongs to a continuous plane, then the local point cloud data is determined to be optical artifact noise. A bilinear interpolation algorithm is used to smoothly reconstruct the optical artifact noise using the effective height data outside the mask in the optically unstable region, generating corrected height data.

7. The method according to claim 1, characterized in that, After calculating the vertical clearance value relative to the non-printing area when the printhead assembly operates on the printing area at a preset printing height, the method further includes: Extract the surface gloss characteristics of the non-printing area. When the non-printing area is a high specular reflection area, construct a geometric model of the UV light reflection path based on the installation angle of the UV curing lamp on the nozzle assembly, the radiation cone angle, and the vertical clearance value. Based on the aforementioned geometric model of the UV light reflection path, the energy density of the reflected light reaching the surface of the printhead nozzle panel after reflection from the non-printing area is simulated and calculated. When the reflected light energy density exceeds the preset critical curing threshold of the UV ink, the minimum optical attenuation distance required to attenuate the reflected light energy density to below the critical curing threshold is calculated, and the vertical clearance value is updated to the larger of the minimum optical attenuation distance and the original vertical clearance value.

8. An automatic batch alignment and positioning system, characterized in that, The automatic batch alignment and positioning system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the automatic batch alignment and positioning system to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on an automatic batch alignment and positioning system, the automatic batch alignment and positioning system performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the automatic batch alignment and positioning system, the automatic batch alignment and positioning system performs the method as described in any one of claims 1-7.

Images