Automatic laser galvanometer calibration method for forging printing equipment

Abstract

The invention is suitable for the field of additive manufacturing, and provides a laser galvanometer automatic calibration method for forging printing equipment, which comprises the following steps: camera installation, camera calibration, substrate adjustment, printing laser galvanometer calibration, ultrafast laser galvanometer calibration, offset calculation and calibration completion. According to the invention, automatic calibration is realized, the efficiency is greatly improved, the calibration precision is improved, the forming quality is ensured, the resource consumption is reduced, the cost is saved, the applicability is high, the method is not limited by equipment models, the operation is simple and convenient, and the skill requirements are reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of additive manufacturing, and particularly relates to a laser galvanometer automatic calibration method for a forging printing device. BACKGROUND

[0002] Additive manufacturing technology has been rapidly and widely developed in the industrial field due to its significant features such as one-time forming of complex structures and high material utilization rate. As an important branch of additive manufacturing technology, the core principle of forging printing technology is to use the high-energy shock wave of ultrafast laser to process the material surface in the layer stacking process, thereby effectively eliminating or weakening the problems such as incomplete fusion of powder and pores generated in the layer stacking process, and further improving the quality of the final formed part.

[0003] To achieve the purpose of the above-mentioned forging printing technology, a key prerequisite is to ensure that the actual physical sintering positions of the printing laser and the ultrafast laser in the substrate of the forging printing device are consistent. Only in this way, can the material printed in each layer be effectively processed, and the energy of the ultrafast laser can be fully utilized. If the part area position strengthened by the ultrafast laser is inconsistent with the part area position sintered by the printing laser, two undesirable consequences will occur: first, the ultrafast laser does not strengthen the entire part surface sintered by the printing laser, affecting the quality of the formed part; second, the ultrafast laser strengthens the non-part area, causing waste of laser energy and reduction of processing efficiency. Therefore, before the forging printing device works, the printing laser and the ultrafast laser must be calibrated to make the part sintering areas of the two lasers consistent.

[0004] In the past, the dual-laser calibration of the forging printing device was carried out in an artificial manner. The specific process is as follows: first, use the printing laser to sinter a known printing model on black paper; then, use the same model to sinter on the same black paper using the ultrafast laser; after sintering, use a two-dimensional image measuring instrument to measure the difference distance between the two laser-sintered models, and adjust the parameters of the ultrafast laser according to the difference distance; then, replace the new black paper, and perform laser sintering and distance difference measurement again, and adjust the laser parameters; repeat the above steps two or three times to realize the coincidence of the sintering areas of the two lasers.

[0005] This artificial laser calibration method has obvious disadvantages: time-consuming, and each calibration requires repeated sintering, measurement and adjustment steps; The measurement step is complicated and requires high skills of the operator; It is easy to cause waste of human and financial resources, and is not conducive to improving production efficiency; Therefore, a laser galvanometer automatic calibration method for a forging printing device is needed to solve the above problems. SUMMARY

[0006] The embodiment of the present application aims to provide a laser galvanometer automatic calibration method for a forging printing device to solve the problems in the background art.

[0007] To achieve the above-mentioned purpose, the present application provides the following technical scheme: A laser galvanometer automatic calibration method for a forging printing device, comprising the following steps: S1, camera installation: an industrial camera is fixedly installed at the top plate position of the forging printing device, the field of view range of the industrial camera needs to completely cover the entire area of the device substrate, so as to ensure that all sintering areas of the laser on the substrate can be collected by the image; the aperture and focal length parameters of the industrial camera are adjusted, so that the collected image clarity meets the subsequent processing requirements; S2, camera calibration: a high-precision ceramic calibration plate is used to calibrate the industrial camera installed in step S1, and the camera internal parameters (including distortion coefficients and other camera distortion parameters) and camera external parameters (including the pose relationship parameters between the camera and the substrate) are calculated and obtained by collecting images of the calibration plate at different positions and different inclination angles; S3, substrate adjustment: the substrate of the forging printing device is leveled by using a scraper ruler and a depth ruler, so that the substrate surface remains horizontal; at the same time, the substrate height is adjusted to the laser sintering reference surface, so as to ensure that the imaging surface of the industrial camera and the laser sintering surface are in the same plane; S4, printing laser galvanometer calibration: a black thin steel plate is placed on the adjusted substrate, the black thin steel plate has high parallelism and the size is consistent with the substrate; according to the thickness of the black thin steel plate, the substrate height is adjusted again, so that the surface of the thin steel plate is consistent with the laser sintering reference surface and the camera imaging surface; the printing laser is controlled to perform laser sintering on the black thin steel plate according to the preset theoretical model; the position of the black thin steel plate is kept unchanged, and the sintered image is collected by the industrial camera; the camera internal parameters obtained in step S2 are used to correct the collected image, and then the corrected image is converted to the world coordinate system and adjusted to the main viewing angle (the main viewing angle is the angle of the operator observing the substrate from the device forming bin door) through the camera external parameters; the image is processed by noise reduction and contrast enhancement through the traditional algorithm of computer vision, and then the printing laser sintering area in the image is extracted through the convolutional neural network; according to the pixel coordinates of the extracted sintering area in the image, combined with the pixel equivalent (conversion ratio of physical size and pixel) obtained by camera calibration, the pixel coordinates are converted to actual physical coordinates; the error between the actual physical coordinates and the theoretical model coordinates is calculated. S5, ultrafast laser galvanometer calibration: keep the position of the black thin steel plate in step S4 unchanged, control the ultrafast laser to perform laser sintering on the black thin steel plate according to the same theoretical model in step S4; collect the sintered image through the industrial camera, which contains the sintering area of the printing laser and the ultrafast laser; process the image according to the same image processing method (distortion correction, coordinate system conversion, image enhancement, convolution neural network extraction) in step S4, and then separate the sintering area of the ultrafast laser from the image by using the image difference algorithm; convert the pixel coordinates of the sintering area of the ultrafast laser into actual physical coordinates combined with the pixel equivalent, and calculate the error between the actual physical coordinates and the theoretical model coordinates; S6, offset calculation: design a theoretical model with an angle, replace the black thin steel plate, and control the printing laser and the ultrafast laser to sinter at the same position according to the angle model; collect the sintered image through the industrial camera, and extract the sintering area features of the two lasers after processing; take the actual physical coordinate system of the printing laser as the reference, combine the model theoretical size and the pixel equivalent, calculate the offset of the sintering area of the two lasers in the X direction ΔX, the offset in the Y direction ΔY, and the angle offset Δθ, and the calculation formula is: ΔX= (c1-c2) × pd, ΔY= (r2-r1) × pd, Δθ= θ1- θ2, wherein (r1, c1) is the image row and column value of the actual sintering area of the printing laser, θ1 is the rotation angle thereof, (r2, c2) is the image row and column value of the actual sintering area of the ultrafast laser, θ2 is the rotation angle thereof, and pd is the pixel equivalent value; S7, calibration completion: feed the error of the printing laser in step S4, the error of the ultrafast laser in step S5, and the offset of the two lasers in step S6 to the control system of the 3D printing device, and the control system automatically adjusts the parameters of the printing laser galvanometer and the ultrafast laser galvanometer to complete the calibration.

[0008] Further technical solutions, in step S1, the industrial camera is an industrial CCD camera with high resolution, and the resolution is not less than 20 million pixels, so as to ensure clear collection of the details of the laser sintering area.

[0009] Further technical solutions, in step S2, the flatness error of the high-precision ceramic calibration plate is not more than 0.001 mm, and the grid spacing error on the calibration plate is not more than 0.005 mm, so as to ensure the accuracy of the camera calibration parameters.

[0010] Further technical solutions, in step S4, the thickness of the black thin steel plate is 0.5-2 mm, the surface roughness Ra is ≤0.8 μm, and it has high temperature resistance and can withstand high temperature without deformation during laser sintering.

[0011] Further technical solutions, in steps S4 and S5, the preset theoretical model is an array pattern containing multiple feature points, the pattern includes at least 9 uniformly distributed circular or square features, and the spacing between the features is 10-50mm, to ensure the comprehensiveness of error calculation.

[0012] Further technical solutions, in steps S4 and S5, the computer vision traditional algorithm includes a Gaussian filter algorithm (for noise reduction) and an adaptive histogram equalization algorithm (for contrast enhancement), and the convolutional neural network is a trained U-Net network, which has an extraction accuracy of sintering area of not less than 99%.

[0013] Further technical solutions, in step S7, when the control system adjusts the laser galvanometer parameters, the adjustment accuracy is not less than 0.001mm in X and Y directions, and not less than 0.001° in angular direction, to ensure the coincidence accuracy of the two laser sintering areas.

[0014] Compared with the prior art, the beneficial effects of the present application are: The present application realizes automatic calibration and greatly improves efficiency: through industrial camera image acquisition, combined with computer vision algorithm and upper computer control, it realizes the full automatic process of printing laser and ultrafast laser galvanometer calibration without manual intervention. Compared with the traditional manual calibration method, the steps of replacing black paper, manual measurement and adjusting parameters are saved, and the single calibration time is shortened from several hours to tens of minutes, which significantly improves the calibration efficiency and reduces the equipment downtime. The present application improves the calibration accuracy and ensures the molding quality: through camera calibration to obtain accurate internal and external parameters, high-precision conversion of image coordinates and actual physical coordinates is realized; convolutional neural network is used to extract sintering area, which has high accuracy and eliminates image noise and external interference through algorithm processing; the offset calculation is based on the actual physical coordinate system, which ensures the accuracy of the measurement of the offset of the two lasers; finally, the coincidence accuracy of the two laser sintering areas is effectively improved, which avoids the quality problems caused by inaccurate calibration and ensures the molding quality of the forging printing equipment. The present application reduces resource consumption and saves cost: black thin steel plate is used as the carrier, which has the characteristics of high temperature resistance and reusability, and greatly reduces the consumption of consumables compared with the one-time use of black paper in the traditional manual calibration. At the same time, the automatic process reduces the dependence on operating personnel and reduces the cost of human resources, thereby realizing the saving of resources and the reduction of cost. The present application has strong applicability and is not limited by the model of the equipment: the calibration method is based on the physical coordinates of the substrate coordinate system and the coordinate difference value for calculation, only a reasonable theoretical model is designed according to the actual substrate size of different forging printing equipment, and the optical device (such as camera lens) is reasonably selected, which can be applied to various models of forging printing equipment, has wide applicability and promotional value. The present application is simple to operate and reduces skill requirements: traditional manual calibration requires higher skills and experience of the operator, while the calibration process of the present application is automatically controlled by the upper computer software, and the operator only needs to start the calibration program to complete the entire calibration process, greatly reducing the skill requirements of the operator, and facilitating the popularization and application in industrial production.

[0015] In order to more clearly illustrate the structural features and effects of the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 Positional schematic view of the laser, ultrafast laser and industrial camera of the forging printing equipment of the present application; Figure 2 Schematic view of the laser calibration process of the forging printing equipment of the present application; Figure 3 Schematic view of the calibration plate for camera calibration to obtain internal and external parameters of the present application; Figure 4 Schematic view of the process for calculating the galvanometer error of the printing laser of the present application; Figure 5 Schematic view of the process for calculating the galvanometer error of the ultrafast laser of the present application; Figure 6 Schematic view of the laser automatic calibration process of the forging printing equipment of the present application; Figure 7 Angle model for calculating the deviation amount of the two lasers of the present application; Figure 8 Sintering original drawing of the printing laser galvanometer calibration (left) and image after processing and correcting to the front perspective coordinate system using camera internal and external parameters (right) of the present application; Figure 9 Original drawing (left) and corrected image (right) obtained after sintering of the ultrafast laser of the present application, and then processing by the difference algorithm; Figure 10 Both lasers of the present application are sintered using the theoretical model with angles, and the image after sintering (left) and the image corrected using camera internal and external parameters (right); Figure 11 The left side is the area sintered by the printing laser, and the right side is the image after sintering by the ultrafast laser. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application more clear, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0018] The specific implementation of the present application is described in detail below in connection with specific embodiments.

[0019] Embodiment 1 As shown in the figure, the embodiment of the present application provides a laser galvanometer automatic calibration method for a forging printing device, comprising the following steps: Figures 1-11 S1, camera installation: A 2000 million pixel industrial CCD camera is fixedly installed in the middle position of the top plate of the forging printing device through a support; according to the size of the device substrate (300mmx300mm), a lens is selected to make the camera field of view range completely cover the substrate area; the camera aperture is adjusted to f / 8, the focus ring is rotated to change the focal length, and through real-time preview of the imaging of the substrate surface, until the fine texture on the substrate is clearly distinguishable, to ensure that the details of the laser sintering area (such as sintering spot boundary, size) can be clearly collected; the camera as the core component of image acquisition, captures the sintering image of the laser on the black thin steel plate in real time, to provide original data support for subsequent image processing. S2, camera calibration:

[0020] A high-precision ceramic calibration plate (flatness error ≤0.001mm, grid spacing error ≤0.005mm) is used, the calibration plate is placed stably on the substrate, and 15-20 calibration plate images are collected within the camera field of view range from different directions (at least covering angles of 0°, 45°, 90°, etc.) and different heights (offset from the substrate reference plane ±5mm); Zhang Zhengyou calibration algorithm is used to process the collected images, to solve the camera intrinsic parameters (including focal length, principal point coordinates, radial and tangential distortion coefficients) and extrinsic parameters (rotation matrix and translation vector of the camera relative to the substrate), to establish the conversion relationship between the image pixel coordinates and the substrate physical coordinates, and to eliminate the influence of camera distortion and installation pose on the measurement accuracy. S3, substrate adjustment:

[0021] A scraper ruler is used to adhere to the surface of the substrate, and is slid along the X and Y axis directions of the substrate to detect the flatness of the substrate, and the substrate flatness error is adjusted to ≤0.01mm by adjusting the support bolts at the bottom of the substrate; then the distance from the substrate surface to the reference plane is measured by using a depth ruler with the laser sintering reference plane as a reference, the substrate lifting mechanism is fine-tuned, the height of the substrate is adjusted to coincide with the laser sintering reference plane, and the imaging plane of the camera is made to be coplanar with the laser sintering plane, to lay a foundation for the consistency of the laser sintering position and the image acquisition coordinates. S4, printing laser galvanometer calibration:

[0022] ​The carrier preparation and installation: a high parallelism black quality thin steel plate with a thickness of 1 mm, a surface roughness Ra = 0.8 μm, and a size of 300 mm x 300 mm is selected, the surface oil and impurities are cleaned, and then it is laid on the substrate. Because the thin steel plate has a thickness, the substrate lifting mechanism is controlled to lower the substrate height by 1 mm, so that the surface of the thin steel plate is strictly aligned with the laser sintering reference surface and the camera imaging surface, and the height difference deviation between the laser sintering position and the image acquisition coordinates is ensured. The black quality thin steel plate has the characteristics of high temperature resistance and good surface light absorption, can stably bear the laser sintering process, and makes the sintering traces (such as sintering spots) clearly present, which is convenient for image feature extraction.

[0023] Laser sintering and image acquisition: the device control system is called, the printing laser parameters (power 30 W, scanning speed 2000 mm / s) are set, the preset theoretical model (19 x 19 solid circle array, circle diameter 10 mm, circle center distance 15 mm) is loaded, and the printing laser galvanometer is controlled to sinter point by point on the surface of the thin steel plate according to the model trajectory. After sintering is completed, the camera is triggered to acquire images, and the original sintering image data is stored.

[0024] Image processing and error calculation: Distortion correction: the internal parameters (distortion coefficients) obtained by camera calibration are used to correct the distortion of the collected images through the undistort function in the OpenCV library, and the geometric deformation of the images caused by the distortion of the camera lens is corrected to obtain regular corrected images.

[0025] Coordinate conversion and view angle adjustment: according to the camera external parameters (rotation matrix, translation vector), the perspectiveTransform function is used to convert the pixel coordinates of the corrected image into the world coordinate system coordinates of the substrate, and through image rotation and translation operations, the main view angle (simulating the view angle of the operator observing the substrate from the device forming bin) is adjusted, so that the image coordinates and the actual observation coordinates are logically consistent.

[0026] Image enhancement and feature extraction: first, the Gaussian filter algorithm (kernel size 5 x 5) is used to denoise the image and suppress random noise interference; then the adaptive histogram equalization is used to improve the image contrast, so that the sintering spots and the background difference are more obvious; the trained U-Net convolutional neural network model is called, the processed image is input, the printing laser sintering area contour is extracted, and the sintering spot center pixel coordinates are output.

[0027] Error calculation: Combine the pixel equivalent of the camera calibration (converted by the grid spacing of the calibration board and the number of grid pixels in the image, such as 1 mm grid spacing corresponds to 10 pixels in the image, then the pixel equivalent is 0.1 mm / pixel), convert the center pixel coordinates of the sintering spot into physical coordinates, compare them with the physical coordinates of the center of the circle in the theoretical model (based on the preset world coordinate system of the substrate), calculate the X and Y direction errors (such as the actual coordinates (x1, y1) and the theoretical coordinates (x0, y0), the error Δx=x1-x0, Δy=y1-y0), and record the calibration error data of the printing laser galvanometer.

[0028] S5, Ultrafast Laser Galvanometer Calibration: Laser sintering and image acquisition: Keeping the position and orientation of the black thin steel plate unchanged, the equipment control system is invoked, and the ultrafast laser parameters (power, scanning speed, etc. are adapted to the printing laser, such as power 30W and scanning speed 2000mm / s) are set. The same theoretical model (19×19 solid circular array) as the printing laser is loaded, and the ultrafast laser galvanometer is controlled to sinter along the trajectory. After sintering is completed, the camera is triggered to acquire images containing the printing and ultrafast laser sintering traces, and the data is stored.

[0029] Image processing and error calculation: Repeat the steps of "distortion correction, coordinate transformation and viewpoint adjustment, and image enhancement" in the calibration of the ultrafast laser galvanometer to obtain a clear and coordinate-aligned image; use an image difference algorithm (such as the absdiff function in OpenCV) to subtract the current image from the printed laser sintering image to separate the ultrafast laser sintering region and extract the center pixel coordinates of the sintering spot; after converting to physical coordinates, compare with the coordinates of the theoretical model to calculate the X and Y direction errors and record the ultrafast laser galvanometer calibration error data.

[0030] Offset calculation and final calibration: Sintering with Angled Models and Image Acquisition: Replace with a new black thin steel plate, load an angled theoretical model (such as an equilateral triangle with a side length of 50mm and a 60° apex angle), and control the printing laser and ultrafast laser to sinter in the same area of ​​the substrate (laser parameters are adapted to the previous description); after sintering, the camera acquires images containing the sintering marks of the two lasers.

[0031] S6, Offset Calculation: Following the aforementioned image processing procedure, extract the features of the two laser sintering regions (such as the pixel coordinates and angles of the apex of an equilateral triangle) and convert them into physical coordinates. Using the physical coordinate system of the printing laser as a reference, calculate the offset of the ultrafast laser in the X direction ΔX = (c1-c2) × pixel equivalent, the offset in the Y direction ΔY = (r2-r1) × pixel equivalent, and the angle offset Δθ = θ1-θ2 (θ1 is the characteristic angle of the printing laser sintering, and θ2 is the characteristic angle of the ultrafast laser sintering).

[0032] S7. Calibration Complete: Upload the offset data to the equipment control system. The system automatically adjusts the ultrafast laser galvanometer parameters (such as galvanometer motor angle compensation and position offset loading), and re-controls the ultrafast laser to sinter according to the theoretical model. Repeat the image acquisition and error calculation process until the overlap of the two laser sintering areas meets the requirements (such as X and Y direction errors ≤ 0.01mm and angle errors ≤ 0.1°). The calibration is then completed, and the calibration parameters are saved for subsequent production.

[0033] Example 2 To adapt to large-size forging printing equipment (substrate size 500mm×500mm), adjust the camera selection, calibration, and adjustment parameters. The following steps are included: S1. Camera Installation: Select a 30-megapixel industrial CCD camera with a wide field of view lens (16mm focal length) and install it on the top plate of the equipment to ensure that the field of view covers a 500mm×500mm substrate; adjust the aperture to f / 11 and finely adjust the focal length to make the image clear within a 500mm span on the substrate surface, meeting the image acquisition requirements of large-size substrates. S2. Camera Calibration: Using a ceramic calibration plate of the same precision (compatible with a 500mm×500mm substrate, grid spacing of 2mm), within the camera's field of view, acquire 20-25 images of the calibration plate from more directions (at least 12 different angles) and different heights (offset from the substrate reference plane ±10mm). Use Zhang Zhengyou's algorithm to calculate the camera's intrinsic and extrinsic parameters and establish the coordinate transformation relationship of the large-size substrate. S3, Substrate Adjustment: Use a longer scraper ruler (compatible with 500mm substrate) to check the flatness, and adjust the support bolts to ensure that the flatness error of the substrate is ≤0.02mm; with the laser sintering reference surface as a reference, adjust the substrate height to coincide with it, and ensure that the camera imaging surface and the laser sintering surface are coplanar. S4. Printing laser galvanometer calibration: Carrier preparation and installation: Select a high parallelism black thin steel plate with a thickness of 1.5mm, a surface roughness Ra=1.0μm, and a size of 500mm×500mm, lay it flat on the substrate, and control the substrate lifting mechanism to lower it by 1.5mm so that the surface of the thin steel plate is aligned with the laser sintering reference surface and the camera imaging surface. Laser sintering and image acquisition: Set the printing laser parameters (power 40W, scanning speed 1800mm / s), load the preset theoretical model (25×25 solid circle array, circle diameter 15mm, circle center distance 20mm), control the printing laser galvanometer sintering, and acquire the image after completion.

[0034] Image processing and error calculation: The distortion correction, coordinate transformation and other steps are the same as in Example 1, and the corresponding OpenCV functions are called for processing; due to the large size of the substrate, the size of the input image of the convolutional neural network is adjusted (e.g., adjusted to 1024×1024 pixels) to ensure the accuracy of feature extraction.

[0035] The pixel equivalent is calculated based on the grid spacing (2mm) of the calibration board and the number of grid pixels in the image. For example, if the grid corresponds to 20 pixels in the image, the pixel equivalent is 0.1mm / pixel. The error is calculated after converting the physical coordinates.

[0036] S5, Ultrafast Laser Galvanometer Calibration: Following the same procedure as in Example 1, the position of the thin steel plate remains unchanged, and the ultrafast laser parameters are set (to adapt to large-size substrates, such as power 40W and scanning speed 1800mm / s). Images are acquired, processed, and errors are calculated.

[0037] S6, Offset Calculation and S7, Final Calibration: Load the angled theoretical model adapted to the large-size substrate (such as a regular quadrilateral with a side length of 80mm and a sharp corner angle of 90°), calculate the offset and adjust the laser galvanometer parameters according to the process in Example 1 until the overlap meets the requirements (X and Y direction errors ≤ 0.02mm, angle error ≤ 0.2°), and complete the calibration.

[0038] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A laser galvanometer automatic calibration method for a forging printing device, characterized in that, Comprising the following steps: S1, camera installation: an industrial camera is fixedly installed at the top plate position of the forging printing equipment, the field of view range of the industrial camera needs to completely cover the entire area of the equipment substrate, to ensure that all sintering areas of the laser on the substrate can be captured by the image; the aperture and focal length parameters of the industrial camera are adjusted, so that the clarity of the captured image meets the subsequent processing requirements; S2, camera calibration: a high-precision ceramic calibration plate is used to calibrate the industrial camera installed in step S1, by capturing images of the calibration plate at different positions and different inclination angles, the camera intrinsic parameters (including distortion coefficients and other camera distortion parameters) and camera extrinsic parameters (including the pose relationship parameters between the camera and the substrate) are calculated and obtained; S3, substrate adjustment: the substrate of the forging printing equipment is leveled by using a scraper ruler and a depth ruler, so that the substrate surface remains horizontal; at the same time, the height of the substrate is adjusted to the laser sintering reference surface, to ensure that the imaging plane of the industrial camera and the laser sintering surface are in the same plane; S4, printing laser galvanometer calibration: a black thin steel plate is placed on the adjusted substrate, the black thin steel plate has high parallelism and the size is consistent with the substrate; According to the thickness of the black thin steel plate, the height of the substrate is adjusted again, so that the surface of the thin steel plate is consistent with the laser sintering reference surface and the camera imaging plane; the printing laser is controlled to perform laser sintering on the black thin steel plate according to the preset theoretical model; the position of the black thin steel plate remains unchanged, and the image after sintering is captured by the industrial camera; the camera intrinsic parameters obtained in step S2 are used to correct the distortion of the captured image, and then the corrected image is converted to the world coordinate system and adjusted to the main viewing angle (the main viewing angle is the angle at which the operator observes the substrate from the equipment forming bin door); the image is processed by traditional computer vision algorithms for noise reduction and contrast enhancement, and then the printing laser sintering area in the image is extracted by a convolutional neural network; according to the pixel coordinates of the extracted sintering area in the image, and in combination with the pixel equivalent (conversion ratio of physical size to pixel) obtained by camera calibration, the pixel coordinates are converted to actual physical coordinates; the error between the actual physical coordinates and the theoretical model coordinates is calculated; S5, ultrafast laser galvanometer calibration: the position of the black thin steel plate in step S4 remains unchanged, and the ultrafast laser is controlled to perform laser sintering on the black thin steel plate according to the same theoretical model in step S4; the image after sintering is captured by the industrial camera, which contains the sintering areas of the printing laser and the ultrafast laser; the image is processed in the same way as in step S4 (distortion correction, coordinate system conversion, image enhancement, convolutional neural network extraction), and then the image difference algorithm is used to separate the ultrafast laser sintering area from the image; in combination with the pixel equivalent, the pixel coordinates of the ultrafast laser sintering area are converted to actual physical coordinates, and the error between the actual physical coordinates and the theoretical model coordinates is calculated; S6, offset calculation: a theoretical model with an angle is designed, after replacing the black thin steel plate, the printing laser and the ultrafast laser are controlled to sinter at the same position according to the angle model; the image after sintering is captured by the industrial camera, and the sintering area features of the two lasers are extracted after processing; With the actual physical coordinate system of the printing laser as the reference, the offset amount ΔX of the two laser sintering areas in the X direction, the offset amount ΔY of the two laser sintering areas in the Y direction, and the angle offset amount Δθ are calculated according to the model theoretical size and the pixel equivalent, and the calculation formula is: ΔX = (c1-c2)×pd, ΔY = (r2-r1)×pd, Δθ = θ1-θ2, wherein (r1, c1) is the image row and column value of the actual sintering area of the printing laser, θ1 is the rotation angle thereof, (r2, c2) is the image row and column value of the actual sintering area of the ultrafast laser, θ2 is the rotation angle thereof, and pd is the pixel equivalent value; S7, calibration is completed: the error of the printing laser in step S4, the error of the ultrafast laser in step S5, and the offset amount of the two lasers in step S6 are fed back to the control system of the forging printing device, and the control system automatically adjusts the parameters of the printing laser galvanometer and the ultrafast laser galvanometer, and the calibration is completed.

2. The laser galvanometer auto-calibration method for a forging printing device according to claim 1, characterized in that, In step S1, the industrial camera is an industrial CCD camera with high resolution, and the resolution is not less than 20 million pixels, so as to ensure clear collection of the details of the laser sintering area.

3. The laser galvanometer auto-calibration method for a forging printing device according to claim 1, wherein, In step S2, the flatness error of the high-precision ceramic calibration plate is not more than 0.001 mm, and the grid spacing error on the calibration plate is not more than 0.005 mm, so as to ensure the accuracy of the camera calibration parameters.

4. The laser galvanometer auto-calibration method for a forging printing device according to claim 1, wherein, In step S4, the black steel sheet has a thickness of 0.5-2 mm, a surface roughness Ra≤0.8 μm, and a high-temperature resistance, and can withstand high temperature without deformation during laser sintering.

5. The laser galvanometer auto-calibration method for a forging printing device according to claim 1, wherein, In steps S4 and S5, the preset theoretical model is an array pattern containing a plurality of feature points, which includes at least 9 uniformly distributed circular or square features, and the spacing between the features is 10-50 mm, so as to ensure the comprehensiveness of the error calculation.

6. The laser galvanometer auto-calibration method for a forging printing device according to claim 1, wherein, In steps S4 and S5, the computer vision traditional algorithm includes a Gaussian filter algorithm (for noise reduction) and an adaptive histogram equalization algorithm (for contrast enhancement), and the convolutional neural network is a trained U-Net network, and the extraction accuracy of the sintering area is not less than 99%.

7. The laser galvanometer auto-calibration method for a stamping printing apparatus according to claim 1, wherein, In step S7, when the control system adjusts the laser galvanometer parameters, the adjustment accuracy in the X and Y directions is not less than 0.001 mm, and the adjustment accuracy in the angle direction is not less than 0.001°, so as to ensure the coincidence accuracy of the two laser sintering areas.

Images