A machine vision-based uvw platform self-calibration compensation method

Abstract

The application discloses a UVW platform self-calibration compensation method based on machine vision and belongs to the technical field of precise motion control and machine vision alignment. In view of the fact that long-term operation of an existing UVW alignment platform can easily cause alignment precision drift due to mechanical wear, shaft gap, perpendicularity deviation and rotation center offset, the application builds a machine vision and UVW platform linkage calibration architecture, relies on a high-precision calibration plate as a reference, controls the UVW platform to traverse a plurality of groups of target positions in a full stroke, synchronously collects images and completes sub-pixel level feature extraction, and accurately solves the actual position of the platform. By comparing the command position with the actual position, the application completes multi-dimensional mechanical error identification of translation zero point drift, shaft perpendicularity, rotation center offset and return gap, constructs an inverse self-calibration compensation model of the command coordinates and the actual position, and realizes online pre-compensation correction of the platform motion coordinates.

Description

Technical Field

[0001] This invention relates to the fields of precision motion control, machine vision intelligent alignment, and automated equipment calibration and compensation technology, specifically to a machine vision-based UVW platform self-calibration and compensation method. Background Technology

[0002] Existing UVW alignment platforms suffer from unavoidable accuracy degradation and calibration lag during long-term continuous industrial operation. The continuous reciprocating motion of the equipment leads to wear on the transmission structure and increased clearance in the shaft system. Simultaneously, factory installation and long-term load operation introduce inherent deviations and deformation deviations, manifesting as multi-dimensional mechanical errors such as platform zero-point drift, X / Y axis perpendicularity deviation, rotation center offset, and excessive shaft return clearance. These errors are systematic and inherent, causing a continuous deviation between the platform's input coordinates and actual motion posture, resulting in a gradual drift in alignment accuracy and leading to quality problems such as poor alignment, misalignment, and assembly deviations in batch products.

[0003] Currently, the main method for correcting accuracy deviations on the UVW platform relies on manual offline calibration. This requires professional operators to stop the machine for inspection and calibration using precision measuring tools such as dial indicators, levels, and reference rulers. This has several drawbacks: First, manual calibration is complex, requires a high level of expertise, and is time-consuming. Frequent machine stops for calibration severely disrupt continuous production line operations, significantly reducing production efficiency. Second, manual calibration is greatly affected by human error, resulting in poor consistency in calibration accuracy and making it impossible to achieve micron-level high-precision calibration. Third, manual calibration can only correct significant deviations in a single dimension and cannot systematically identify and compensate for multi-dimensional composite mechanical errors, thus failing to address the root cause of long-term accuracy drift in the equipment. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a machine vision-based self-calibration compensation method for UVW platforms. This method solves the technical problems of low efficiency, high downtime losses, poor calibration accuracy, inability to compensate for multi-dimensional systematic mechanical errors, and serious long-term accuracy drift of traditional UVW platforms. It enables unmanned, online, and fully automated self-calibration of the equipment, ensuring the long-term high-precision and high-stability operation of the UVW platform.

[0005] The objective of this invention can be achieved through the following technical solutions: A machine vision-based self-calibration compensation method for a UVW platform includes the following steps: S1: Build a self-calibration hardware architecture, fix a high-precision calibration board on the UVW platform, fix an image acquisition module on the top of the UVW platform, establish signal interconnection between the image acquisition module and the motion controller of the UVW platform, and complete system initialization and camera parameter calibration. S2: Generate a full-stroke calibration motion sequence, control the UVW platform to drive the calibration plate to move sequentially to multiple preset target poses, the target poses cover the full working stroke of the platform, including multiple translation points, rotation points and forward and reverse motion points; S3: Visual synchronous acquisition and pose calculation. After the UVW platform reaches a set of target poses stably, the image acquisition module synchronously acquires the calibration board image, preprocesses the image and extracts sub-pixel level feature points, and combines the camera calibration parameters to convert the pixel coordinates into world coordinates to calculate the actual pose parameters of the platform. S4: Multi-dimensional mechanical error identification, matching and comparing each set of preset command poses with the corresponding actual pose data, and fitting and calculating to obtain four core mechanical error parameters of the UVW platform: translation zero point drift, X / Y axis perpendicularity error, rotation center offset, and shaft return clearance. S5: Construct a self-calibration compensation model. Based on the identified multi-dimensional error parameters, establish an inverse compensation mapping relationship between platform command coordinates and actual motion poses. Solve for the optimal compensation parameters by fitting using the least squares method, and generate and store the compensation parameter table. S6: Online real-time error compensation. During normal alignment operations on the UVW platform, the compensation parameter table is called to pre-compensate and correct the target alignment coordinates, and the corrected motion commands are output to offset the platform's mechanical system errors and achieve high-precision alignment.

[0006] Preferably, in step S2, the calibration motion sequence includes uniformly distributed lateral and longitudinal translation points and multi-angle rotation points. The forward and reverse motion points are used to collect shaft return clearance deviation data. Each set of target poses is repeatedly collected multiple times, and the average value is taken as the effective pose data to eliminate random collection errors.

[0007] Preferably, in step S3, the calibration plate is a ceramic substrate checkerboard calibration plate or a dot array calibration plate, the feature point extraction adopts a sub-pixel corner detection algorithm or a sub-pixel circle center detection algorithm, the pose calculation accuracy reaches the sub-pixel level, and the camera calibration includes intrinsic parameter calibration and hand-eye calibration to achieve accurate mapping between pixel coordinates and world coordinates.

[0008] Preferably, the specific method for identifying multi-dimensional mechanical errors in step S4 is as follows: by linear fitting of multiple points on the X and Y axes, the zero-point drift and motion proportionality coefficient of the X and Y axes are solved; the verticality deviation angle of the X / Y axes is calculated by the coordinate deviation of the orthogonal points; the offset of the platform rotation center is solved by fitting the coordinate changes of the feature points of multi-angle rotation; and the return clearance value of each axis is solved by the pose difference value of the forward and reverse motion at the same point.

[0009] Preferably, in step S5, the self-calibration compensation model includes a linear compensation term and a nonlinear compensation term. The linear compensation term is used to correct zero drift, shaft proportional deviation, perpendicularity error, and return clearance error. The nonlinear compensation term uses a piecewise interpolation method to correct local nonlinear accuracy deviations within the platform's travel range.

[0010] Preferably, the coordinate correction calculation formula for the linear compensation term is:

[0011] in, To compensate for the corrected target coordinates, The original command coordinates, These represent the zero-point drift amounts on the X and Y axes, respectively. These are the motion proportionality coefficients for the X and Y axes, respectively. This is the perpendicularity deviation angle of the X / Y axes.

[0012] Preferably, it also includes a dynamic iterative update step: during normal production, the equipment supports automatic or manual one-click triggering of the self-calibration process, real-time collection of the current platform pose data, and iterative update of the compensation parameter table; when the platform alignment error is detected to continuously exceed the preset threshold, the full-stroke calibration is automatically started to dynamically correct the accuracy drift.

[0013] A machine vision-based UVW platform self-calibration compensation system, used to implement a machine vision-based UVW platform self-calibration compensation method, characterized by comprising: The UVW alignment platform is used to support the workpiece to be processed and realize precise fine-tuning alignment with three degrees of freedom: X translation, Y translation, and θ rotation. A high-precision calibration plate is fixedly installed at the center of the UVW platform table surface as a reference for machine vision calibration. The image acquisition module is fixedly mounted above the UVW platform and is used to acquire images of the calibration plate and the workpiece alignment. The motion control unit is used to output calibration motion commands and alignment motion commands to control the UVW platform to accurately complete pose adjustment; The vision processing unit is used to complete image preprocessing, sub-pixel feature extraction, pose calculation, multi-dimensional error identification and compensation model construction. The compensation execution unit is used to call the compensation parameter table, perform pre-compensation correction on the target alignment coordinates, and output the corrected motion command.

[0014] Preferably, the visual processing unit has a built-in image preprocessing module, a sub-pixel feature detection module, a camera calibration module, a multi-dimensional error identification module, and a compensation model fitting module, which can realize micron-level pose calculation and full-dimensional mechanical error parameter extraction.

[0015] Preferably, the system is configured with both manual one-click calibration and timed automatic calibration modes. The entire calibration process is completed automatically without human intervention, and the updated calibration parameters take effect immediately without the need to stop and restart the equipment, making it suitable for continuous industrial production scenarios.

[0016] The beneficial effects of this invention are: This invention introduces an integrated self-calibration algorithm for UVW alignment platforms, boasting five major advantages: First, it achieves fully automatic online calibration without downtime or manual operation, supporting timed / one-click calibration, eliminating downtime losses and improving equipment uptime; second, it identifies four types of core mechanical errors across all dimensions, achieving micron-level high-precision compensation through sub-pixel vision and least squares fitting, improving long-term operational accuracy drift; third, it employs a linear + nonlinear piecewise interpolation composite compensation model, balancing global and local errors to ensure uniform alignment accuracy throughout the entire stroke; fourth, it dynamically iterates compensation parameters during production, adapting to equipment wear and deformation to avoid error accumulation; and fifth, it is compatible with mainstream UVW platforms and vision systems, requiring no hardware modifications, requiring only algorithm integration for implementation, offering good versatility and low modification costs, suitable for various precision automated alignment production lines. Attached Figure Description

[0017] The invention will now be further described with reference to the accompanying drawings.

[0018] Figure 1 This is a flowchart of an algorithm according to an embodiment of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1 As shown, a machine vision-based UVW platform self-calibration compensation method includes the following steps: S1: Build a self-calibration hardware architecture, fix a high-precision calibration board on the UVW platform, fix an image acquisition module on the top of the UVW platform, establish signal interconnection between the image acquisition module and the motion controller of the UVW platform, and complete system initialization and camera parameter calibration. S2: Generate a full-stroke calibration motion sequence, control the UVW platform to drive the calibration plate to move sequentially to multiple preset target poses, the target poses cover the full working stroke of the platform, including multiple translation points, rotation points and forward and reverse motion points; S3: Visual synchronous acquisition and pose calculation. After the UVW platform reaches a set of target poses stably, the image acquisition module synchronously acquires the calibration board image, preprocesses the image and extracts sub-pixel level feature points, and combines the camera calibration parameters to convert the pixel coordinates into world coordinates to calculate the actual pose parameters of the platform. S4: Multi-dimensional mechanical error identification, matching and comparing each set of preset command poses with the corresponding actual pose data, and fitting and calculating to obtain four core mechanical error parameters of the UVW platform: translation zero point drift, X / Y axis perpendicularity error, rotation center offset, and shaft return clearance. S5: Construct a self-calibration compensation model. Based on the identified multi-dimensional error parameters, establish an inverse compensation mapping relationship between platform command coordinates and actual motion poses. Solve for the optimal compensation parameters by fitting using the least squares method, and generate and store the compensation parameter table. S6: Online real-time error compensation. During normal alignment operations on the UVW platform, the compensation parameter table is called to pre-compensate and correct the target alignment coordinates, and the corrected motion commands are output to offset the platform's mechanical system errors and achieve high-precision alignment.

[0021] In step S2, the calibration motion sequence includes uniformly distributed lateral and longitudinal translation points and multi-angle rotation points. The forward and reverse motion points are used to collect shaft return clearance deviation data. Each target pose is repeatedly collected multiple times, and the average value is taken as the effective pose data to eliminate random collection errors.

[0022] In step S3, the calibration board is a ceramic substrate checkerboard calibration board or a dot array calibration board. The feature point extraction adopts a sub-pixel corner detection algorithm or a sub-pixel circle center detection algorithm. The pose calculation accuracy reaches the sub-pixel level. The camera calibration includes intrinsic parameter calibration and hand-eye calibration to achieve accurate mapping between pixel coordinates and world coordinates.

[0023] The specific method for identifying multi-dimensional mechanical errors in step S4 is as follows: by linear fitting of multiple points on the X and Y axes, the zero-point drift and motion proportionality coefficient of the X and Y axes are solved; the verticality deviation angle of the X / Y axes is calculated by the coordinate deviation of the orthogonal points; the offset of the platform rotation center is solved by fitting the coordinate changes of the feature points of multi-angle rotation; and the return clearance value of each axis is solved by the pose difference value of the forward and reverse motion at the same point.

[0024] In step S5, the self-calibration compensation model includes linear compensation terms and nonlinear compensation terms. The linear compensation terms are used to correct zero drift, shaft proportional deviation, perpendicularity error and return clearance error. The nonlinear compensation terms use piecewise interpolation to correct local nonlinear accuracy deviations within the platform's travel range.

[0025] The formula for calculating the coordinate correction of the linear compensation term is:

[0026] in, To compensate for the corrected target coordinates, The original command coordinates, These represent the zero-point drift amounts on the X and Y axes, respectively. These are the motion proportionality coefficients for the X and Y axes, respectively. This is the perpendicularity deviation angle of the X / Y axes.

[0027] It also includes dynamic iterative update steps: During normal equipment production, it supports automatic or manual one-click triggering of the self-calibration process, real-time collection of current platform pose data, and iterative update of the compensation parameter table; when the platform alignment error is detected to continuously exceed the preset threshold, it automatically starts full-stroke calibration to dynamically correct accuracy drift.

[0028] A machine vision-based UVW platform self-calibration compensation system, used to implement a machine vision-based UVW platform self-calibration compensation method, characterized by comprising: The UVW alignment platform is used to support the workpiece to be processed and realize precise fine-tuning alignment with three degrees of freedom: X translation, Y translation, and θ rotation. A high-precision calibration plate is fixedly installed at the center of the UVW platform table surface as a reference for machine vision calibration. The image acquisition module is fixedly mounted above the UVW platform and is used to acquire images of the calibration plate and the workpiece alignment. The motion control unit is used to output calibration motion commands and alignment motion commands to control the UVW platform to accurately complete pose adjustment; The vision processing unit is used to complete image preprocessing, sub-pixel feature extraction, pose calculation, multi-dimensional error identification and compensation model construction. The compensation execution unit is used to call the compensation parameter table, perform pre-compensation correction on the target alignment coordinates, and output the corrected motion command.

[0029] The vision processing unit has built-in image preprocessing module, sub-pixel feature detection module, camera calibration module, multi-dimensional error identification module and compensation model fitting module, which can realize micron-level pose calculation and full-dimensional mechanical error parameter extraction.

[0030] The system is configured with both manual one-click calibration and timed automatic calibration modes. The entire calibration process is completed automatically without human intervention. The updated calibration parameters take effect immediately without the need to stop and restart the equipment, making it suitable for continuous industrial production scenarios.

[0031] Example 1 This embodiment provides a machine vision-based UVW platform self-calibration compensation system, which is applied to the UVW alignment equipment in a mobile phone screen precision bonding production line to achieve fully automatic platform precision calibration and error compensation, ensuring screen bonding alignment accuracy.

[0032] The system hardware includes a high-precision UVW alignment platform, a ceramic checkerboard calibration plate, an image acquisition module, a motion control unit, a vision processing unit, and a compensation execution unit. The UVW alignment platform has an effective X / Y axis travel of ±50mm, a θ-axis rotation angle of ±5°, and a factory repeatability of ±1μm. The ceramic checkerboard calibration plate is fixed to the center of the platform surface; the checkerboard unit has a side length of 1mm and a machining accuracy of ±0.001mm, featuring high flatness, high wear resistance, and no deformation, and can serve as a long-term calibration benchmark. The image acquisition module consists of a 5-megapixel industrial camera, dual telecentric lenses, and a coaxial shadowless light source, fixedly suspended above the platform, providing a complete field of view covering the platform's working area with clear and distortion-free imaging. The motion control unit uses a high-precision servo motion controller, enabling precise output of minute displacements. The vision processing unit is equipped with an industrial control computer, integrating a complete set of image algorithms and error identification algorithms. The compensation execution unit is embedded in the motion control link, achieving real-time coordinate pre-compensation and correction.

[0033] The system software enables fully automated control of the entire process, including camera calibration, point motion control, image acquisition, feature calculation, error fitting, model generation, online compensation, and dynamic updates. It can autonomously complete calibration and compensation tasks.

[0034] Example 2 This embodiment, based on the above system, fully executes the self-calibration compensation process of the present invention, with the specific steps as follows: First, the system is initialized. After powering on, the device automatically completes camera intrinsic parameter calibration and hand-eye calibration, establishing a precise mapping relationship between pixel coordinates and world coordinates, and confirming that the calibration board image is clear and the device is operating normally. The system automatically generates a full-stroke calibration motion sequence, including 36 sets of X / Y axis translation points and 12 sets of multi-angle rotation points, while also configuring forward and reverse motion points. The points are evenly distributed throughout the platform's full-stroke working range, comprehensively covering various error conditions of the device.

[0035] After initiating the self-calibration process, the motion control unit sequentially outputs motion commands, controlling the UVW platform to move the calibration board to the preset target pose one by one. Once the platform is in place and stable, the camera automatically triggers image acquisition, repeatedly acquiring image data three times at each point to effectively avoid random imaging errors. The vision processing unit performs noise reduction, distortion correction, and image enhancement preprocessing on the acquired images. It uses a sub-pixel corner detection algorithm to extract checkerboard feature points and calculates the actual X and Y translation coordinates and θ rotation angle of the platform corresponding to each pose.

[0036] After data acquisition, the system automatically matches and fits all command poses with the actual pose data to complete multi-dimensional error identification: solving the zero-point drift and motion proportionality coefficient of the X and Y axes through multi-point linear fitting; calculating the verticality deviation angle of the X / Y axes through orthogonal point position deviation; solving the rotation center offset through multi-angle rotation data fitting; and calculating the return clearance parameters of each axis through forward and reverse point position difference.

[0037] Based on all identified error parameters, the system constructs a composite self-calibration compensation model. The optimal compensation parameters are solved iteratively using the least squares method, generating and storing a standardized compensation parameter table. The model corrects global systematic deviations through linear compensation terms and corrects local accuracy deviations in the stroke through piecewise nonlinear interpolation, comprehensively optimizing the platform's motion accuracy.

[0038] After the equipment enters the normal production alignment operation state, the compensation execution unit intervenes in the control link in real time, receives the original alignment coordinates sent by the host computer, calls the compensation parameter table to complete the coordinate pre-correction through the compensation formula, and sends the corrected accurate coordinates to the motion controller to drive the UVW platform to complete the alignment action, effectively offsetting various mechanical errors and achieving high-precision alignment.

[0039] Example 3 This embodiment details the system's dynamic update and anomaly protection mechanisms to ensure long-term stable operation of the equipment. The system supports a timed automatic calibration function, which can be set to automatically calibrate upon daily startup or after every 1000 products produced. No manual operation is required; compensation parameters are automatically updated to adapt to accuracy changes caused by long-term equipment wear. Simultaneously, the system has an accuracy monitoring function, comparing alignment deviations in real time. When multiple consecutive alignment deviations exceed a preset warning threshold, it automatically forces a full-stroke self-calibration to quickly correct accuracy drift issues.

[0040] Multiple anomaly protection mechanisms are implemented during the calibration process. If issues such as image blurring, feature point extraction failure, platform motion deviations, or data fitting anomalies occur, the system automatically terminates the calibration, retaining the original valid compensation parameters to prevent erroneous parameters from causing equipment accuracy failure. After calibration, the system automatically performs accuracy verification. Parameters are only updated after successful verification, maximizing the safety and reliability of the calibration operation.

[0041] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A machine vision-based UVW platform self-calibration compensation method, characterized in that, Includes the following steps: S1: Build a self-calibration hardware architecture, fix a high-precision calibration board on the UVW platform, fix an image acquisition module on the top of the UVW platform, establish signal interconnection between the image acquisition module and the motion controller of the UVW platform, and complete system initialization and camera parameter calibration. S2: Generate a full-stroke calibration motion sequence, control the UVW platform to drive the calibration plate to move sequentially to multiple preset target poses, the target poses cover the full working stroke of the platform, including multiple translation points, rotation points and forward and reverse motion points; S3: Visual synchronous acquisition and pose calculation. After the UVW platform reaches a set of target poses stably, the image acquisition module synchronously acquires the calibration board image, preprocesses the image and extracts sub-pixel level feature points, and combines the camera calibration parameters to convert the pixel coordinates into world coordinates to calculate the actual pose parameters of the platform. S4: Multi-dimensional mechanical error identification, matching and comparing each set of preset command poses with the corresponding actual pose data, and fitting and calculating to obtain four core mechanical error parameters of the UVW platform: translation zero point drift, X / Y axis perpendicularity error, rotation center offset, and shaft return clearance. S5: Construct a self-calibration compensation model. Based on the identified multi-dimensional error parameters, establish an inverse compensation mapping relationship between platform command coordinates and actual motion poses. Solve for the optimal compensation parameters by fitting using the least squares method, and generate and store the compensation parameter table. S6: Online real-time error compensation. During normal alignment operations on the UVW platform, the compensation parameter table is called to pre-compensate and correct the target alignment coordinates, and the corrected motion commands are output to offset the platform's mechanical system errors and achieve high-precision alignment.

2. The machine vision-based UVW platform self-calibration compensation method according to claim 1, characterized in that, In step S2, the calibration motion sequence includes uniformly distributed lateral and longitudinal translation points and multi-angle rotation points. The forward and reverse motion points are used to collect shaft return clearance deviation data. Each target pose is repeatedly collected multiple times, and the average value is taken as the effective pose data to eliminate random collection errors.

3. The machine vision-based UVW platform self-calibration compensation method according to claim 1, characterized in that, In step S3, the calibration plate is a ceramic substrate checkerboard calibration plate or a dot array calibration plate. The feature point extraction adopts a sub-pixel corner detection algorithm or a sub-pixel circle center detection algorithm. The pose calculation accuracy reaches the sub-pixel level. The camera calibration includes intrinsic parameter calibration and hand-eye calibration to achieve accurate mapping between pixel coordinates and world coordinates.

4. The machine vision-based UVW platform self-calibration compensation method according to claim 1, characterized in that, The specific method for identifying multi-dimensional mechanical errors in step S4 is as follows: by linear fitting of multiple points on the X and Y axes, the zero-point drift and motion proportionality coefficient of the X and Y axes are solved; the verticality deviation angle of the X / Y axes is calculated by the coordinate deviation of the orthogonal points; the offset of the platform rotation center is solved by fitting the coordinate changes of the feature points of multi-angle rotation; and the return clearance value of each axis is solved by the pose difference value of the forward and reverse motion at the same point.

5. The machine vision-based UVW platform self-calibration compensation method according to claim 1, characterized in that, In step S5, the self-calibration compensation model includes a linear compensation term and a nonlinear compensation term. The linear compensation term is used to correct zero drift, shaft proportional deviation, perpendicularity error and return clearance error. The nonlinear compensation term uses a piecewise interpolation method to correct local nonlinear accuracy deviations within the platform's travel range.

6. The machine vision-based UVW platform self-calibration compensation method according to claim 5, characterized in that, The formula for calculating the coordinate correction of the linear compensation term is as follows: ； in, To compensate for the corrected target coordinates, The original command coordinates, These represent the zero-point drift amounts on the X and Y axes, respectively. These are the motion proportionality coefficients for the X and Y axes, respectively. This is the perpendicularity deviation angle of the X / Y axes.

7. The machine vision-based UVW platform self-calibration compensation method according to claim 1, characterized in that, It also includes dynamic iterative update steps: During normal equipment production, it supports automatic or manual one-click triggering of the self-calibration process, real-time collection of current platform pose data, and iterative update of the compensation parameter table; when the platform alignment error is detected to continuously exceed the preset threshold, it automatically starts full-stroke calibration to dynamically correct accuracy drift.

8. A machine vision-based UVW platform self-calibration compensation system, used to implement the machine vision-based UVW platform self-calibration compensation method according to any one of claims 1-7, characterized in that, include: The UVW alignment platform is used to support the workpiece to be processed and realize precise fine-tuning alignment with three degrees of freedom: X translation, Y translation, and θ rotation. A high-precision calibration plate is fixedly installed at the center of the UVW platform table surface as a reference for machine vision calibration. The image acquisition module is fixedly mounted above the UVW platform and is used to acquire images of the calibration plate and the workpiece alignment. The motion control unit is used to output calibration motion commands and alignment motion commands to control the UVW platform to accurately complete pose adjustment; The vision processing unit is used to complete image preprocessing, sub-pixel feature extraction, pose calculation, multi-dimensional error identification and compensation model construction. The compensation execution unit is used to call the compensation parameter table, perform pre-compensation correction on the target alignment coordinates, and output the corrected motion command.

9. The machine vision-based UVW platform self-calibration compensation system according to claim 8, characterized in that, The vision processing unit has built-in image preprocessing module, sub-pixel feature detection module, camera calibration module, multi-dimensional error identification module and compensation model fitting module, which can realize micron-level pose calculation and full-dimensional mechanical error parameter extraction.

10. The machine vision-based UVW platform self-calibration compensation system according to claim 8, characterized in that, The system is configured with both manual one-click calibration and timed automatic calibration modes. The entire calibration process is completed automatically without human intervention. The updated calibration parameters take effect immediately without the need to stop and restart the equipment, making it suitable for continuous industrial production scenarios.

Images