Remaining plate material identification method and system for laser cutting machine

Abstract

The invention relates to the technical field of laser cutting, in particular to an excess plate recognition method and system for a laser cutting machine, and the method comprises the following steps: S1, configuring line laser vision equipment, installing the line laser vision equipment on a cross beam of the laser cutting machine, enabling the line laser vision equipment to move along with the cross beam, and achieving the full-breadth scanning of a working area of a machine tool; s2, performing visual calibration on the line laser visual equipment by using at least two calibration plates, and obtaining a conversion relation delta between a coordinate system of the line laser visual equipment and a coordinate system of a machine tool; and S3, the line laser vision equipment is controlled to scan the excess materials on the machine tool, point cloud data of the excess materials are obtained, the point cloud data are converted into a machine tool coordinate system through the conversion relation delta, and a DXF file used for representing the outlines of the excess materials is generated. The DXF file used for representing the contour of the excess material is generated through the point cloud data, automation of excess material recognition is achieved, manual intervention is reduced, and production efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of laser cutting technology, specifically to a method and system for identifying residual material in a laser cutting machine. Background Technology

[0002] In the field of laser cutting, it has become an industry consensus to reuse the unprocessed areas (i.e., "leftovers") on the finished sheet metal to improve material utilization and reduce production costs. However, since leftovers are usually irregular in shape, traditional methods make it difficult to directly obtain their contour information, resulting in low efficiency in leftover identification and reuse.

[0003] Currently, common methods for identifying scrap materials mainly fall into two categories: manual measurement and machine vision. Manual measurement relies on operators using calipers, measuring tapes, and other tools to measure the scrap material segment by segment and then manually draw a contour drawing. This method is not only time-consuming and labor-intensive but also prone to large measurement errors, making it difficult to meet the demands of high-precision processing. Furthermore, manual operation is heavily influenced by experience, resulting in significant differences in contour drawings drawn by different personnel, which is detrimental to subsequent automated layout and cutting. Regarding machine vision, some companies have attempted to use area scan cameras or 2D vision systems to photograph and identify scrap materials. However, due to interference factors such as reflections, scratches, and stains on the board surface, 2D images struggle to accurately distinguish the scrap material from the machine tool table or background, leading to insufficient contour extraction accuracy. In addition, 2D vision systems cannot obtain the height information of the scrap material and lack the ability to perceive thickness changes caused by warping, stacking, or local deformation, further limiting their practical application effectiveness.

[0004] In summary, existing technologies struggle to balance efficiency, accuracy, and automation in scrap identification, necessitating an intelligent solution to achieve rapid identification of laser-cut scrap. Summary of the Invention

[0005] To address the aforementioned problems, this invention provides a method and system for identifying leftover material in a laser cutting machine.

[0006] In a first aspect, the present invention provides a method for identifying residual material in a laser cutting machine, comprising the following steps: S1. Configure a line laser vision device and install it on the crossbeam of the laser cutting machine so that it moves with the crossbeam to achieve full-area scanning of the machine tool's working area; S2. Use at least two calibration plates to perform visual calibration on the line laser vision device and obtain the transformation relationship δ between the coordinate system of the line laser vision device and the machine tool coordinate system. S3. Control the line laser vision device to scan the scrap material on the machine tool, obtain the point cloud data of the scrap material, and convert the point cloud data to the machine tool coordinate system through the transformation relationship δ to generate a DXF file representing the outline of the scrap material.

[0007] By mounting a line laser vision device on the crossbeam of a laser cutting machine, full-area scanning of the machine tool's working area is achieved without the need for additional mechanical devices, thus improving scanning efficiency and coverage. Visual calibration is performed using at least two calibration plates to obtain the transformation relationship δ between the line laser vision device's coordinate system and the machine tool's coordinate system, ensuring high-precision conversion of point cloud data and improving the accuracy of subsequent processing.

[0008] As a further limitation of the technical solution of the present invention, S1 specifically includes: S11. Install the line laser vision device on one side of the crossbeam via a connector and communicate with the CNC system of the laser cutting machine. S12. Set global scan range parameters, including the length and width of the scan area; S13. Calculate the scanning path for the device to perform full-area scanning based on the global scanning range parameters, and control the crossbeam to move along the scanning path so that the line laser vision device can complete the full-area scanning of the machine tool's working area.

[0009] The line laser vision device is mounted on the crossbeam via connectors and communicates with the CNC system of the laser cutting machine, achieving equipment integration, reducing reliance on external equipment, and improving system stability and reliability. By setting global scanning range parameters, including the length and width of the scanning area, the scanning range can be flexibly adjusted according to different sized plates, improving the system's versatility and adaptability. The scanning path is calculated based on the global scanning range parameters, and the crossbeam is controlled to move along the planned path, achieving efficient full-area scanning and reducing scanning time and mechanical wear.

[0010] As a further limitation of the technical solution of the present invention, the specific steps in S13 for calculating the scanning path based on the global scanning range parameters include: S13a. Calculate the required number of scan strips ViewY in the Y direction based on the width ScanWidth of the scan area and the single scan field width W_fov of the line laser vision device. S13b. Determine the X-coordinate X2 of the scan path endpoint based on the X-coordinate X0 of the scan starting point, the length of the scan area (ScanLength), and the parity of the scan strip number (ViewY): If ViewY is odd, then X2 = X0 + ScanLength; If ViewY is even, then X2 = X0; S13c. Determine the Y coordinate Y2 of the scanning endpoint based on the width ScanWidth of the scanning area. Y2 = Y0 + ScanWidth; S13d, based on the starting point (X0, Y0), ending point (X2, Y2), number of scan strips ViewY, and length of the scan area ScanLength, generates an S-shaped scanning path in which the beam moves in opposite directions between adjacent scan strips.

[0011] By calculating the number of scan strips (ViewY) in the Y direction and determining the X coordinate of the scan path endpoint, an S-shaped scan path is generated, ensuring the accuracy and efficiency of the scan path and reducing scanning time and the complexity of mechanical movements. Based on the parity of the scan strip number, the X coordinate of the scan endpoint is flexibly adjusted to adapt to scan areas of different widths, improving the system's flexibility and adaptability. Precise calculation of the Y coordinate of the scan endpoint ensures the integrity and accuracy of the scan path, improving scanning efficiency and data integrity.

[0012] As a further limitation of the technical solution of the present invention, in S13, the number of scans in the Y direction is calculated based on the width of the scan area, that is, the number of scan strips ViewY in the Y direction; ViewY = ceil(ScanWidth / W_fov) Where W_fov is the field of view of the line laser vision device in a single scan at the installation height, i.e., the camera field of view, and ceil() is the round-up function.

[0013] The number of scans (ViewY) in the Y direction is calculated using the floor function ceil(), ensuring the scientific validity and rationality of the scan count, avoiding missed or repeated scans, and improving scanning efficiency. The single-scan field of view width (W_fov) of the line laser vision device is introduced as a calculation parameter, ensuring the accuracy of the scan count calculation and improving the system's adaptability and reliability.

[0014] As a further limitation of the technical solution of the present invention, the specific steps of visual calibration in S2 using at least two calibration plates include: S21. Fix two calibration plates to two different positions with a height difference on the machine tool worktable, the two positions including the lower left area and the upper right area of ​​the machine tool worktable; predefine corner points on each calibration plate; the corner points include the lower left corner point or the upper left corner point of the calibration plate; S22. Set calibration parameters in the CNC system of the laser cutting machine. The calibration parameters include at least the calibration plate size, calibration motion speed, and calibration range. S23. Control the indicator light spot of the laser cutting head to move sequentially to the predefined corner points on the two calibration plates, and record the coordinates of the corner points in the machine tool coordinate system; S24. Control the crossbeam to move the line laser vision device along a preset calibration path, so that the line laser vision device scans the two calibration plates and generates corresponding calibration plate point cloud data. S25. Extract the image coordinates of the calibration board corner points from the point cloud data, and combine them with the coordinates of the corresponding corner points in the machine tool coordinate system obtained in S23. Calculate the transformation relationship δ between the coordinate system of the line laser vision device and the machine tool coordinate system using a coordinate system transformation algorithm.

[0015] By placing two calibration plates on the machine tool's worktable and setting calibration parameters in the CNC system, precise visual calibration was achieved, ensuring a high-precision transformation relationship δ between the line laser vision equipment's coordinate system and the machine tool's coordinate system. Placing the calibration plates in the lower left and upper right areas of the machine tool's worktable ensured the stability and reliability of the calibration process, improving calibration accuracy. By controlling the indicator light spot of the laser cutting head to move to the corner of the calibration plate and recording the corner coordinates, an automated calibration process was achieved, reducing manual intervention and improving calibration efficiency.

[0016] As a further limitation of the technical solution of the present invention, S3 specifically includes: S31. Set the task parameters for this scan in the CNC system of the laser cutting machine. The task parameters include the scanning range, scanning speed and scanning start point. The scanning range is not greater than the global scanning range defined in step S12. S32. Control the crossbeam to move along the planned scanning path and trigger the line laser vision device to synchronously acquire images and obtain raw scan image data; S33. Process the acquired raw image data to generate initial three-dimensional point cloud data, and perform filtering and noise reduction processing on the initial three-dimensional point cloud data. S34. Based on the height threshold, the processed point cloud data is segmented to distinguish the point cloud of the scrap material from the point cloud of the machine tool table, and the effective point cloud dataset representing the shape of the scrap is extracted. S35. Using the transformation relationship δ, the effective point cloud dataset is transformed from the line laser vision device coordinate system to the machine tool coordinate system; S36. Perform contour fitting processing on the effective point cloud data converted to the machine tool coordinate system to generate a DXF vector graphic file representing the outline of the scrap material.

[0017] Setting scanning task parameters in the CNC system of a laser cutting machine, including scanning range, scanning speed, and scanning start point, allows for flexible adjustment of these parameters according to actual needs, improving the system's flexibility and adaptability. Processing the acquired raw image data generates initial 3D point cloud data, which is then filtered and denoised, improving the quality and accuracy of the point cloud data and providing a reliable data foundation for subsequent point cloud segmentation and contour fitting. Point cloud data is segmented based on a height threshold, distinguishing the point cloud of the scrap material from the point cloud of the machine tool table, extracting an effective point cloud dataset representing the shape of the scrap, thus improving the accuracy and reliability of scrap identification. Contour fitting of the effective point cloud data generates a DXF vector graphic file representing the scrap's outline, achieving automated scrap contour generation, reducing manual intervention, and improving production efficiency.

[0018] As a further limitation of the technical solution of the present invention, the specific method for segmenting the processed point cloud data based on the height threshold in S34 is as follows: Calculate the statistical value of the Z-axis coordinate in the point cloud data, and set the height threshold according to the formula Δ=μ+ kσ, where μ is the mean of the Z-axis coordinate, σ is the standard deviation of the Z-axis coordinate, and k is a constant; Point cloud data with Z-axis coordinate values ​​less than the height threshold Δ are identified as machine tool table point clouds and excluded. Point cloud data with height values ​​greater than or equal to the height threshold Δ are retained as valid point cloud data for scrap materials.

[0019] By calculating the statistical value of the Z-axis coordinate in the point cloud data and setting a height threshold according to the formula Δ=μ+kσ, the scientific validity and rationality of the height threshold are ensured, thus improving the accuracy and reliability of point cloud segmentation. Point cloud data with Z-axis coordinate values ​​less than the height threshold Δ are identified as machine tool table point clouds and excluded, while point cloud data with values ​​greater than or equal to the height threshold Δ are retained as valid point cloud data for leftover materials. This achieves precise point cloud segmentation and improves the accuracy of leftover material identification.

[0020] As a further limitation of the technical solution of the present invention, S36 specifically includes: S36a. Project the three-dimensional effective point cloud data in the machine tool coordinate system onto the XY plane of the machine tool to form a two-dimensional scattered point set; S36b. Extract edge points from the two-dimensional scatter set to obtain an edge point sequence that represents the outline of the remaining material. S36c, Fit the edge point sequence into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments; S36d. The fitted closed two-dimensional contour line is encapsulated according to the DXF file format standard to generate the final DXF vector graphics file.

[0021] In S36c, a piecewise fitting algorithm is used to fit the edge point sequence into a closed two-dimensional contour line composed of smoothly connected straight line segments and circular arc segments.

[0022] By projecting effective 3D point cloud data onto the XY plane of the machine tool to form a 2D scattered point set, and then performing edge point extraction and contour fitting to generate a closed 2D contour line, an efficient contour fitting process is achieved, improving the efficiency and accuracy of contour generation. The fitted closed 2D contour line is encapsulated according to the DXF file format standard to generate the final DXF vector graphics file, ensuring the standardization and compatibility of the generated file and facilitating subsequent nesting programming.

[0023] As a further limitation of the technical solution of the present invention, the specific steps of S36c include: S36c.1. Traverse the edge point sequence and, based on the changes in turning angle or curvature between adjacent points, preliminarily identify feature points that may be the boundary points between straight line segments and arc segments. S36c.2. Divide the point set segment between adjacent feature points into a unit to be fitted, and use the least squares method to perform straight line fitting or circular arc fitting respectively, and calculate the fitting error of each fitting method. S36c.3. Compare the straight line fitting error and the circular arc fitting error of the same unit to be fitted. Based on the error magnitude and the preset accuracy threshold, select the model with the smaller fitting error as the final contour line of the segment. And ensure that adjacent straight line segments and circular arc segments satisfy the tangent connection relationship through geometric constraints. S36c.4 Check whether all the line segments and arcs generated by the fitting are connected end to end to form a complete closed two-dimensional contour line.

[0024] A piecewise fitting algorithm is employed to fit the edge point sequence into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments. The least squares method is used for straight line fitting or arc fitting, and the model with the smaller fitting error is selected as the final contour line, ensuring the accuracy and smoothness of the contour fitting. Geometric constraints ensure that adjacent straight line segments and arc segments are tangent, improving the continuity and smoothness of the contour and avoiding contour breaks or sharp corners. Checking whether all fitted line segments and arcs are connected end-to-end to form a complete closed two-dimensional contour line ensures the integrity and accuracy of the contour, improving the quality and reliability of the generated file.

[0025] Secondly, the present invention provides a sheet metal scrap identification system for a laser cutting machine, comprising: Line laser vision equipment is installed on the crossbeam of a laser cutting machine to move with the crossbeam and perform full-area scanning of the machine tool's working area; The control module is used to control the scanning path and data acquisition of the line laser vision device; The calibration module is used to perform visual calibration on the line laser vision device using at least two calibration plates, and to obtain the transformation relationship δ between the line laser vision device coordinate system and the machine tool coordinate system. The image processing module is used to process the image data obtained by scanning and generate point cloud data; The point cloud processing module is used to segment point cloud data based on a height threshold and extract a valid point cloud dataset that represents the shape of the scrap material. A coordinate transformation module is used to transform effective point cloud data from the line laser vision device coordinate system to the machine tool coordinate system using the transformation relationship δ. The contour generation module is used to perform contour fitting processing on the converted point cloud data to generate a DXF vector graphic file representing the contour of the scrap material.

[0026] As a further limitation of the technical solution of the present invention, the control module includes: The scanning parameter setting unit is used to set the length and width of the scanning area, the scanning start point, and the scanning speed; The scanning path calculation unit is used to calculate the scanning path of the line laser vision device based on the scanning area parameters, and control the crossbeam to move along an S-shaped path to complete the full-area scanning.

[0027] By integrating line laser vision equipment, control modules, calibration modules, image processing modules, point cloud processing modules, coordinate transformation modules, and contour generation modules, a complete laser cutting machine sheet metal scrap identification system is formed, achieving automation and integration of scrap identification and improving system stability and reliability. From scanning by the line laser vision equipment to the final generation of DXF vector graphics files, the entire process is automated, reducing manual intervention and improving production efficiency and scrap utilization. Through precise coordinate transformation, point cloud segmentation, and contour fitting, a high-precision scrap contour DXF file is generated, providing accurate drawing information for subsequent layout programming and improving production quality and efficiency.

[0028] As a further limitation of the technical solution of the present invention, the scanning path calculation unit includes: The Y-direction scan strip count calculation subunit is used to calculate the Y-direction scan strip count ViewY based on the scan area width and the single field of view width of the line laser vision device. The X-axis endpoint coordinate calculation subunit is used to determine the X-coordinate of the endpoint of the scan path based on the parity of the number of scan strips. The path generation sub-unit is used to generate an S-shaped scan path based on the starting point, ending point, number of scan strips, and region length.

[0029] As a further limitation of the technical solution of the present invention, the number of scan strips in the Y direction, ViewY, is calculated in the following way: ViewY = rounded up (ScanWidth / W_fov), Where ScanWidth is the width of the scanning area, and W_fov is the field of view width of a single scan of the line laser vision device.

[0030] As a further limitation of the technical solution of the present invention, the calibration module includes: The calibration plate arrangement unit is used to place two calibration plates in the lower left and upper right areas of the machine tool worktable, respectively. The corner point recording unit is used to control the indicator light spot of the laser cutting head to move sequentially to the preset corner point on the calibration plate and record its coordinates in the machine tool coordinate system; The calibration scanning control unit is used to control the line laser vision equipment to scan two calibration boards and generate calibration point cloud data; The coordinate transformation calculation unit is used to extract the image coordinates of corner points in the calibration point cloud data and calculate the transformation relationship δ by combining them with the coordinates in the machine tool coordinate system.

[0031] As a further limitation of the technical solution of the present invention, the point cloud processing module includes: The height threshold calculation unit is used to calculate the height threshold Δ = μ + kσ based on the statistical values ​​of the Z-axis coordinates of the point cloud data, where μ is the mean of the Z-axis coordinates, σ is the standard deviation, and k is a constant. Point cloud segmentation unit is used to retain point cloud data with Z-axis coordinates greater than or equal to the height threshold Δ as valid point cloud datasets.

[0032] As a further limitation of the technical solution of the present invention, the contour generation module includes: The projection unit is used to project three-dimensional effective point cloud data onto the XY plane of the machine tool to form a two-dimensional scattered point set. The edge extraction unit is used to extract edge points from a two-dimensional scatter set to obtain an edge point sequence. The contour fitting unit is used to fit a sequence of edge points into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments. The DXF generation unit is used to encapsulate closed two-dimensional contour lines in DXF format and generate DXF vector graphics files.

[0033] As a further limitation of the technical solution of the present invention, the contour fitting unit includes: The feature point recognition subunit is used to identify the boundary points between straight line segments and circular arc segments based on the turning angle or curvature changes between adjacent points; The fitting error calculation subunit is used to perform straight line fitting and circular arc fitting on each point set and calculate the fitting error. The model selection sub-unit is used to select the fitting model based on the error magnitude and accuracy threshold, and to ensure tangential connection between adjacent segments through geometric constraints. The closure check sub-unit is used to check whether the fitted line segment and the arc are connected end to end to form a complete closed contour.

[0034] As can be seen from the above technical solutions, this application has the following advantages: The line laser vision device is mounted on the laser head beam, and the scanning motion can be completed using the existing CNC axis system, eliminating the need for additional robots or gantry mechanisms; after initial configuration, subsequent processing does not require manual parameter setting, truly achieving one-click scanning and significantly reducing operational intensity. The number of stripes in the Y direction (ViewY) is automatically calculated based on the field of view width W_fov and the plate width, and an odd-even reverse S-shaped path is adopted to minimize idle travel and acceleration / deceleration times; under the same working area, the scanning time is shorter than that of a line-by-line unidirectional path, improving the effective utilization rate of the machine tool. A dual calibration plate scheme of lower left + upper right is adopted, and a rigid correspondence between the camera coordinate system and the machine tool coordinate system is established in one step through corner point extraction and homogeneous transformation matrix δ; the calibration process is automatically controlled by the CNC system to align the laser head red light, laying the foundation for subsequent high-precision contour reconstruction. Based on the μ+kσ adaptive height threshold, the point cloud of machine tool structures such as blades and support teeth is automatically removed, and only the surface data of the remaining material is retained; it can still stably segment complex situations such as warping, sinking, and multi-layer stacking, avoiding secondary screening.

[0035] By generating DXF files representing the outline of scrap material using point cloud data, the identification of scrap material is automated, reducing manual intervention and improving production efficiency. Attached Figure Description

[0036] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart of a method provided in an embodiment of the present invention.

[0038] Figure 2 This is a flowchart of the visual calibration process in an embodiment of the present invention.

[0039] Figure 3 This is a flowchart of the scanning process in an embodiment of the present invention.

[0040] Figure 4 A block diagram of the apparatus provided in an embodiment of the present invention. Detailed Implementation

[0041] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0042] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this application and in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0043] like Figure 1 As shown in the figure, this embodiment of the invention provides a method for identifying sheet metal scrap in a laser cutting machine, including the following steps: S1. Configure a line laser vision device and install it on the crossbeam of the laser cutting machine so that it moves with the crossbeam to achieve full-area scanning of the machine tool's working area; S2. Use at least two calibration plates to perform visual calibration on the line laser vision device and obtain the transformation relationship δ between the coordinate system of the line laser vision device and the machine tool coordinate system. S3. Control the line laser vision device to scan the scrap material on the machine tool, obtain the point cloud data of the scrap material, and convert the point cloud data to the machine tool coordinate system through the transformation relationship δ to generate a DXF file representing the outline of the scrap material.

[0044] In some embodiments, S1 specifically includes: S11. Install the line laser vision device on one side of the crossbeam via a connector and communicate with the CNC system of the laser cutting machine. S12. Set global scan range parameters, including the length and width of the scan area; input the length and width of the scan range to be scanned in the system. The scan range length is the range of movement in the X direction during contour scanning, and the scan range Y is the range of movement in the Y direction during contour scanning. S13. Calculate the scanning path for the device to perform full-area scanning based on the global scanning range parameters, and control the crossbeam to move along the scanning path so that the line laser vision device can complete the full-area scanning of the machine tool's working area.

[0045] The linear laser vision device is mounted on one side of the machine tool's crossbeam using a connector, and the line laser moves along with the crossbeam. The scanning path is calculated through parameter settings, and the movement of the crossbeam is controlled to allow the linear laser vision device to scan the entire surface of the machine tool. The device collects the deformed linear laser stripes, and the computer processes these stripes to obtain point cloud data of the shape of the remaining material on the machine tool.

[0046] The laser cutting machine described in this application has a gantry structure, including a bed, a crossbeam that can move along the Y-axis guide rail, and a laser head that can move along the crossbeam in the X-axis direction. The line laser vision device is mounted on the crossbeam via a connector, in the same way as the laser head, enabling the device to reuse its movement capabilities in the X and Y axes under the control of the machine tool's CNC system, thereby achieving two-dimensional scanning motion relative to the machine tool's worktable.

[0047] Scanning implementation principle: Y-axis scanning: When the CNC system controls the entire crossbeam to move along the bed guide rail, the camera mounted on the crossbeam also moves in the Y-axis direction. This enables the camera to be positioned at different Y-coordinate locations.

[0048] X-axis scanning: The CNC system controls the laser head to move on the crossbeam, thereby moving the line laser vision device in the X-axis direction. During this process, the line laser vision device continuously acquires images at a high frame rate.

[0049] The system plans an S-shaped path covering the entire tabletop by controlling the alternating movement of the X and Y axes. The line laser vision device follows this path, and the continuously acquired two-dimensional image data is stitched and processed to finally synthesize point cloud data describing the three-dimensional shape of the tabletop material.

[0050] The basic process is as follows: During the movement of the beam, the camera takes pictures according to the frame rate. Taking the current frame as an example, the laser in the line laser camera emits a laser beam that is projected onto the surface of the object. The camera takes a picture of this laser beam from another angle. Due to the change in the height of the object surface, the position of the line laser beam in the imaging plane of the camera will shift. Through image processing (using existing algorithms for ROI extraction, filtering and noise reduction, contrast enhancement, Gaussian fitting, and surface reconstruction), the position calculation of the current frame image is completed. At this time, the position information is the point cloud data of a line on the board. The next cycle of movement continues, and finally the scanning of the entire area is completed, forming the final point cloud data.

[0051] In this embodiment of the invention, the specific steps in S13 for calculating the scan path based on the global scan range parameters include: S13a. Calculate the required number of scan strips ViewY in the Y direction based on the width ScanWidth of the scan area and the single scan field width W_fov of the line laser vision device. S13b. Determine the X-coordinate X2 of the scan path endpoint based on the X-coordinate X0 of the scan starting point, the length of the scan area (ScanLength), and the parity of the scan strip number (ViewY): If ViewY is odd, then X2 = X0 + ScanLength; If ViewY is even, then X2 = X0; S13c. Determine the Y coordinate Y2 of the scanning endpoint based on the width ScanWidth of the scanning area. Y2 = Y0 + ScanWidth; S13d, based on the starting point (X0, Y0), ending point (X2, Y2), number of scan strips ViewY, and length of the scan area ScanLength, generates an S-shaped scanning path in which the beam moves in opposite directions between adjacent scan strips.

[0052] In fact, in S13, an S-shaped scanning path is planned based on the scanning start point coordinates (X0, Y0), the length and width of the scanning area, and the camera field of view, so that the crossbeam moves in opposite directions between adjacent scanning strips. The endpoint coordinates (X2, Y2) of the scanning path are determined by the parity of the number of scanning strips in the Y direction. If the total number of stripes ViewY is odd, then the laser head is located at the rightmost side of the scanning area when the scan ends, i.e., X2 = X0 + ScanLength; If the total number of stripes ViewY is even, then the laser head will be located at the leftmost edge of the scanning area when the scan ends, i.e., X2 = X0; At the end of the scan, the laser head is positioned on the other side of the scanned area in the Y direction, i.e., Y2 = Y0 + ScanWidth. During the scan, after each strip scan is completed, the laser head moves a step distance in the Y direction of ΔY = ScanWidth / (ViewY-1).

[0053] Here, in S13, the number of scans in the Y direction, i.e., the number of scan strips in the Y direction ViewY, is calculated based on the width of the scan area. ViewY = ceil(ScanWidth / W_fov) Wherein, W_fov is the field of view of the line laser vision device in a single scan at the installation height, i.e., the camera's field of view, and ceil() is the rounding function. In this embodiment of the invention, it is recommended that the line laser camera be installed above the laser head at a distance of approximately 850mm ± 100mm from the cutting plate, at which point W_fov is 1000.

[0054] In some embodiments, such as Figure 2 As shown, the specific steps for visual calibration in S2 using at least two calibration plates include: S21. Fix two calibration plates to two different positions with a height difference on the machine tool worktable, the two positions including the lower left area and the upper right area of ​​the machine tool worktable; predefine corner points on each calibration plate; the corner points include the lower left corner point or the upper left corner point of the calibration plate; S22. Set calibration parameters in the CNC system of the laser cutting machine. The calibration parameters include at least the calibration plate size, calibration motion speed, and calibration range. S23. Control the indicator light spot of the laser cutting head to move sequentially to the predefined corner points on the two calibration plates, and record the coordinates of the corner points in the machine tool coordinate system; S24. Control the crossbeam to move the line laser vision device along a preset calibration path, so that the line laser vision device scans the two calibration plates and generates corresponding calibration plate point cloud data. S25. Extract the image coordinates of the calibration board corner points from the point cloud data, and combine them with the coordinates of the corresponding corner points in the machine tool coordinate system obtained in S23. Calculate the transformation relationship δ between the coordinate system of the line laser vision device and the machine tool coordinate system using a coordinate system transformation algorithm.

[0055] It should be noted here that when obtaining the coordinates of the corner points mentioned on the calibration board in the machine tool coordinate system, the physical corner point positions are the same.

[0056] In this embodiment of the invention, before performing line laser scanning, the line laser camera is visually calibrated using two square calibration plates. The purpose of visual calibration is to obtain the correspondence δ between the line laser coordinate system and the machine tool coordinate system. This invention includes a calibration method for a line laser vision device. The two calibration plates are placed in the lower left and upper right regions respectively. Calibration parameters are set in the system, such as the size of the two calibration plates, calibration speed, and calibration range. The corner coordinates of the two calibration plates are obtained. The calibration process involves movement in a predetermined direction. After calibration, the correspondence between the camera coordinate system and the machine tool coordinate system is automatically obtained through coordinate transformation. The calibration range determines the movement distance of the machine tool, so the calibration parameters need to be modified accordingly for different situations.

[0057] The obtained corner coordinates are transformed to correspond one-to-one with the coordinates of the actual machine tool, thus obtaining the correspondence δ between the camera coordinate system and the machine tool coordinate system. The transformation between the camera coordinate system and the actual machine tool coordinate system is a rigid body transformation in three-dimensional space. The transformation relationship δ uses a homogeneous transformation matrix:

[0058] in Represents the rotation matrix. 0 represents the translation variable, and 0 represents the zero vector.

[0059] The rotation matrix can be automatically calculated using the calibration method described above. R Translation vector T The specific numerical value of the rotation matrix. R It is a unit orthogonal matrix that satisfies And det( R )=1; the translation vector The amount These represent the offsets of the optical center of the line laser vision device along the three coordinate axes in the machine tool coordinate system. Rotation matrix R It is a 3×3 matrix whose column vectors and row vectors are unit vectors and orthogonal to each other. This means... R Each column and each row is a vector of length 1, and the dot product of any two distinct column vectors or row vectors is 0.

[0060] In some embodiments, such as Figure 3 As shown, S3 specifically includes: S31. Set the task parameters for this scan in the CNC system of the laser cutting machine. The task parameters include the scanning range, scanning speed and scanning start point. The scanning range is not greater than the global scanning range defined in step S12. S32. Control the crossbeam to move along the planned scanning path and trigger the line laser vision device to synchronously acquire images and obtain raw scan image data; S33. Process the acquired raw image data to generate initial three-dimensional point cloud data, and perform filtering and noise reduction processing on the initial three-dimensional point cloud data. S34. Based on the height threshold, the processed point cloud data is segmented to distinguish the point cloud of the scrap material from the point cloud of the machine tool table, and the effective point cloud dataset representing the shape of the scrap is extracted. S35. Using the transformation relationship δ, the effective point cloud dataset is transformed from the line laser vision device coordinate system to the machine tool coordinate system; S36. Perform contour fitting processing on the effective point cloud data converted to the machine tool coordinate system to generate a DXF vector graphic file representing the outline of the scrap material.

[0061] In this embodiment of the invention, the specific method for segmenting the processed point cloud data based on a height threshold in step S34 is as follows: Calculate the statistical value of the Z-axis coordinate in the point cloud data, and set the height threshold according to the formula Δ=μ+ kσ, where μ is the mean of the Z-axis coordinate, σ is the standard deviation of the Z-axis coordinate, and k is a constant; Point cloud data with Z-axis coordinate values ​​less than the height threshold Δ are identified as machine tool table point clouds and excluded. Point cloud data with height values ​​greater than or equal to the height threshold Δ are retained as valid point cloud data for scrap materials.

[0062] It should be further explained that S36 specifically includes: S36a. Project the three-dimensional effective point cloud data in the machine tool coordinate system onto the XY plane of the machine tool to form a two-dimensional scattered point set; In this embodiment of the invention, the specific steps are as follows: For each 3D point P_i (x_i, y_i, z_i) in the effective point cloud dataset, the X coordinate x_i and Y coordinate y_i of its coordinate values ​​are directly extracted, while its Z coordinate z_i is ignored, thereby generating a 2D scattered point set consisting only of 2D coordinate pairs (x_i, y_i).

[0063] S36b. Extract edge points from the two-dimensional scatter set to obtain an edge point sequence that represents the outline of the remaining material. In this embodiment of the invention, the Alpha Shapes algorithm is used to extract edge points from a two-dimensional scatter set to obtain an edge point sequence. Specific steps include: S36b.1 Based on the aforementioned two-dimensional scatter set, construct its Delaunay triangulation network; S36b.2. Set a radius threshold α and traverse each edge in the triangular network. If there exists a circle with a radius greater than α that can pass through the two vertices of an edge and does not contain any other points in the scatter set, then the edge is determined to be a boundary edge. S36b.3 Connect all the selected boundary edges to form one or more closed polygonal rings, and the vertex sequence of the polygonal rings is the edge point sequence that represents the outline of the scrap material.

[0064] Here, the value of the radius threshold α is adaptively determined based on the average point spacing of the two-dimensional scatter set, and its value ranges from 1.5 to 3.0 times the average point spacing.

[0065] S36c, Fit the edge point sequence into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments; In this embodiment of the invention, the step of using a piecewise fitting algorithm to fit the edge point sequence into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments includes: S36c.1. Traverse the edge point sequence and, based on the changes in turning angle or curvature between adjacent points, preliminarily identify feature points that may be the boundary points between straight line segments and arc segments. S36c.2. Divide the point set segment between adjacent feature points into a unit to be fitted, and use the least squares method to perform straight line fitting or circular arc fitting respectively, and calculate the fitting error of each fitting method. S36c.3. Compare the fitting error of the straight line with the fitting error of the circular arc of the same unit to be fitted. Based on the error magnitude and the preset accuracy threshold, select the model with the smaller fitting error as the final contour line of the segment. And ensure that the adjacent straight line segments and circular arc segments satisfy the tangent connection relationship through geometric constraints. The method of applying the tangent connection geometric constraints described in S36c.3 is as follows: For adjacent line segments, adjust the center position and radius of the fitted circular arc so that its starting point or ending point coincides with the endpoint of the adjacent line segment, and at the endpoint, the tangent of the circular arc is in the same direction as the tangent of the adjacent line segment.

[0066] S36c.4 Check whether all the line segments and arcs generated by the fitting are connected end to end to form a complete closed two-dimensional contour line.

[0067] S36d. The fitted closed two-dimensional contour line is encapsulated according to the DXF file format standard to generate the final DXF vector graphics file.

[0068] In this embodiment of the invention, the specific steps for encapsulating the file according to the DXF file format standard to generate the final DXF vector graphics file include: S36d.1. Convert the fitted line segments and arc segments into LINE and ARC entities in DXF format, and calculate all the parameters required to write each entity; for line segments, the parameters include the starting coordinates and ending coordinates; for arc segments, the parameters include the center coordinates, radius, starting angle and ending angle. S36d.2. Create a DXF file and write the header segment, table segment, block segment, entity segment, and file end marker in sequence according to the standard DXF file structure; wherein, the entity segment is used to store all LINE and ARC entities generated in step S36d.1; S36d.3. Ensure that all LINE and ARC entities are connected end-to-end in the entity segment to form a closed outline, and write it into the entity segment of the DXF file; S36d.4. Save the packaged DXF file to the specified storage path of the CNC system of the laser cutting machine to complete the output.

[0069] In S36d.2, at least one layer is also created in the table segment of the DXF file, and all LINE and ARC entities generated in S36d.1 are assigned to this layer.

[0070] In this invention, after scanning, corresponding point cloud data is generated based on the shape of the scrap material. The point cloud data contains information in the XYZ directions. The position information in the Z-axis direction can distinguish the scrap material from the machine tool cutter. Since the correspondence between the camera coordinate system and the machine tool coordinate system has been obtained during camera calibration, the point cloud data can be converted to the machine tool coordinate system through the correspondence. Finally, a DXF file containing the shape of the scrap material is generated. Importing the DXF file into the system for subsequent layout operations can greatly improve the utilization rate of the scrap material.

[0071] In this embodiment of the invention, scanning parameters are set, such as scanning range (only residual material exists within the current range, which can further save scanning time), scanning speed, and scanning starting point. After scanning begins, the crossbeam moves according to the calculated scanning path, and the line laser camera moves along with the crossbeam. Simultaneously, the line laser camera begins collecting data and executes the contour scanning logic. Each completed X-direction movement generates a set of point cloud data, which is stored. When all scanning paths are completed, the line point cloud data obtained from each frame is added to an array, forming a three-dimensional point cloud array. Simultaneously, the Z-direction data in the point cloud data is analyzed, and a threshold judgment is performed on the Z-direction data. The purpose is to separate the residual material from the machine tool cutter. For materials that do not meet the threshold range (Δ = μ + kσ), the remaining material is separated. The point cloud data (k=2.5-3.0) is excluded for noise reduction. A weighted fusion algorithm is then used to generate the final point cloud data. This processed point cloud data represents the information of the remaining sheet metal on the current machine tool. During camera calibration, the transformation relationship δ between the machine tool coordinate system and the camera coordinate system is obtained. The point cloud data of the remaining material is converted to the machine tool coordinate system through coordinate transformation, and a DXF drawing file is output. This file contains the drawing information of the sheet metal on the current machine tool bed. The DXF sheet metal file is imported into the system for subsequent layout. Here, μ is the mean of the z-coordinate, and σ is the standard deviation of the z-coordinate.

[0072] like Figure 4 As shown, this embodiment of the invention provides a sheet metal scrap identification system for a laser cutting machine, comprising: Line laser vision equipment is installed on the crossbeam of a laser cutting machine to move with the crossbeam and perform full-area scanning of the machine tool's working area; The control module is used to control the scanning path and data acquisition of the line laser vision device; The calibration module is used to perform visual calibration on the line laser vision device using at least two calibration plates, and to obtain the transformation relationship δ between the line laser vision device coordinate system and the machine tool coordinate system. The image processing module is used to process the image data obtained by scanning and generate point cloud data; The point cloud processing module is used to segment point cloud data based on a height threshold and extract a valid point cloud dataset that represents the shape of the scrap material. A coordinate transformation module is used to transform effective point cloud data from the line laser vision device coordinate system to the machine tool coordinate system using the transformation relationship δ. The contour generation module is used to perform contour fitting processing on the converted point cloud data to generate a DXF vector graphic file representing the contour of the scrap material.

[0073] In some embodiments, the control module includes: The scanning parameter setting unit is used to set the length and width of the scanning area, the scanning start point, and the scanning speed; The scanning path calculation unit is used to calculate the scanning path of the line laser vision device based on the scanning area parameters, and control the crossbeam to move along an S-shaped path to complete the full-area scanning.

[0074] In some embodiments, the scan path calculation unit includes: The Y-direction scan strip count calculation subunit is used to calculate the Y-direction scan strip count ViewY based on the scan area width and the single field of view width of the line laser vision device. The X-axis endpoint coordinate calculation subunit is used to determine the X-coordinate of the endpoint of the scan path based on the parity of the number of scan strips. The path generation sub-unit is used to generate an S-shaped scan path based on the starting point, ending point, number of scan strips, and region length.

[0075] The number of scan stripes in the Y direction, ViewY, is calculated in the following way: ViewY = rounded up (ScanWidth / W_fov), Where ScanWidth is the width of the scanning area, and W_fov is the field of view width of a single scan of the line laser vision device.

[0076] In some embodiments, the calibration module includes: The calibration plate arrangement unit is used to place two calibration plates in the lower left and upper right areas of the machine tool worktable, respectively. The corner point recording unit is used to control the indicator light spot of the laser cutting head to move sequentially to the preset corner point on the calibration plate and record its coordinates in the machine tool coordinate system; The calibration scanning control unit is used to control the line laser vision equipment to scan two calibration boards and generate calibration point cloud data; The coordinate transformation calculation unit is used to extract the image coordinates of corner points in the calibration point cloud data and calculate the transformation relationship δ by combining them with the coordinates in the machine tool coordinate system.

[0077] In some embodiments, the point cloud processing module includes: The height threshold calculation unit is used to calculate the height threshold Δ = μ + kσ based on the statistical values ​​of the Z-axis coordinates of the point cloud data, where μ is the mean of the Z-axis coordinates, σ is the standard deviation, and k is a constant. Point cloud segmentation unit is used to retain point cloud data with Z-axis coordinates greater than or equal to the height threshold Δ as valid point cloud datasets.

[0078] In some embodiments, the contour generation module includes: The projection unit is used to project three-dimensional effective point cloud data onto the XY plane of the machine tool to form a two-dimensional scattered point set. The edge extraction unit is used to extract edge points from a two-dimensional scatter set to obtain an edge point sequence. The contour fitting unit is used to fit a sequence of edge points into a closed two-dimensional contour line composed of smoothly connected straight line segments and arc segments. The DXF generation unit is used to encapsulate closed two-dimensional contour lines in DXF format and generate DXF vector graphics files.

[0079] In some embodiments, the contour fitting unit includes: The feature point recognition subunit is used to identify the boundary points between straight line segments and circular arc segments based on the turning angle or curvature changes between adjacent points; The fitting error calculation subunit is used to perform straight line fitting and circular arc fitting on each point set and calculate the fitting error. The model selection sub-unit is used to select the fitting model based on the error magnitude and accuracy threshold, and to ensure tangential connection between adjacent segments through geometric constraints. The closure check sub-unit is used to check whether the fitted line segment and the arc are connected end to end to form a complete closed contour.

[0080] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method of identifying a plate excess material of a laser cutting machine, characterized by, The method comprises the following steps: S1, configuring a line laser vision device, and installing the line laser vision device on a crossbeam of a laser head to realize synchronous movement with the laser head, so as to realize full-width scanning of a working area of a machine tool; the line laser vision device realizes scanning in a Y direction through Y axis movement of the crossbeam; S2, using at least two calibration plates to perform vision calibration on the line laser vision device, so as to obtain a conversion relationship δ between a coordinate system of the line laser vision device and a coordinate system of the machine tool; S3, controlling the line laser vision device to scan a remaining material on the machine tool, so as to obtain point cloud data of the remaining material, and converting the point cloud data to the coordinate system of the machine tool through the conversion relationship δ, so as to generate a DXF file used for representing a contour of the remaining material.

2. The plate scrap identifying method of a laser cutting machine according to claim 1, characterized by, S1 specifically comprises: S11, installing the line laser vision device on one side of the crossbeam through a connecting piece, and realizing communication connection with a numerical control system of the laser cutting machine; S12, setting a global scanning range parameter, including a length and a width of a scanning area; S13, calculating a scanning path of full-width scanning of the device according to the global scanning range parameter, controlling the crossbeam to move according to the scanning path, and enabling the line laser vision device to complete full-width scanning of the working area of the machine tool.

3. The plate scrap identifying method of a laser cutting machine according to claim 2, characterized by, The specific steps of calculating the scanning path according to the global scanning range parameter in S13 comprise: S13a, calculating a required number of scanning strips ViewY in the Y direction according to the width ScanWidth of the scanning area and a single scanning field width W_fov of the line laser vision device; S13b, determining an X coordinate X2 of an end point of the scanning path according to an X coordinate X0 of a starting point of the scanning, a length ScanLength of the scanning area and parity of the number of scanning strips (ViewY): if the number of scanning strips ViewY is odd, X2 = X0+ScanLength; if the number of scanning strips ViewY is even, X2 = X0; S13c, determining a Y coordinate Y2 of the end point of the scanning according to the width ScanWidth of the scanning area; Y2 = Y0+ScanWidth; S13d, generating an S-shaped scanning path of the crossbeam moving in a reverse direction between adjacent scanning strips based on the starting point (X0, Y0), the end point (X2, Y2), the number of scanning strips ViewY and the length ScanLength of the scanning area.

4. The plate scrap identifying method of a laser cutting machine according to claim 3, characterized by, In S13, the number of scanning times in the Y direction, i.e. the number of scanning strips ViewY in the Y direction, is calculated according to the width of the scanning area; ViewY = ceil(ScanWidth / W_fov) wherein W_fov is a single scanning field width of the line laser vision device at an installation height, i.e. a camera field of view, and ceil() is a rounding-up function.

5. The plate scrap identifying method of a laser cutting machine according to claim 4, characterized by, The specific steps of performing vision calibration through the at least two calibration plates in S2 comprise: S21, fixing two calibration plates at two different positions with a height difference on a working table surface of the machine tool, the two positions including a lower left region and an upper right region of the working table surface; and predefining corner points on each calibration plate; the corner points include a lower left corner point or an upper left corner point of the calibration plate; S22, setting calibration parameters in the numerical control system of the laser cutting machine, the calibration parameters at least including calibration plate size, calibration motion speed and calibration range; S23, controlling the indication light point of the laser cutting head to move to the pre-defined corner points on the two calibration plates in turn, and recording the coordinates of the corner points in the machine tool coordinate system; S24, controlling the cross beam to drive the line laser vision equipment to move according to the pre-set calibration path, so that the line laser vision equipment scans the two calibration plates and generates corresponding calibration plate point cloud data; S25, extracting the image coordinates of the calibration plate corner points in the point cloud data, and combining the coordinates of the corresponding corner points in the machine tool coordinate system obtained in S23, the conversion relationship δ between the coordinate system of the line laser vision equipment and the machine tool coordinate system is calculated through the coordinate system conversion algorithm.

6. The plate scrap identifying method of a laser cutting machine according to claim 5, wherein S3 specifically comprises: S31, setting the task parameters of this scan in the numerical control system of the laser cutting machine, the task parameters including scan range, scan speed and scan starting point, wherein the scan range is not greater than the global scan range defined in step S12; S32, controlling the cross beam to move according to the planned scan path, and triggering the line laser vision equipment to synchronously perform image acquisition, to obtain original scan image data; S33, processing the collected original image data to generate initial three-dimensional point cloud data, and performing filtering and noise reduction processing on the initial three-dimensional point cloud data; S34, segmenting the processed point cloud data based on a height threshold, to distinguish the point cloud of the excess material from the point cloud of the machine tool table, and extract an effective point cloud data set representing the shape of the excess material; S35, using the conversion relationship δ to convert the effective point cloud data set from the line laser vision equipment coordinate system to the machine tool coordinate system; S36, performing contour fitting processing on the effective point cloud data converted to the machine tool coordinate system, to generate a DXF vector graphics file representing the contour of the excess material.

7. The plate scrap identifying method of a laser cutting machine according to claim 6, wherein The specific method for segmenting the processed point cloud data based on a height threshold in S34 is: Calculate the statistical value of the Z-direction coordinates in the point cloud data, and set the height threshold according to the formula Δ=μ+kσ, wherein μ is the mean value of the Z-direction coordinates, σ is the standard deviation of the Z-direction coordinates, and k is a constant; The point cloud data with a Z-direction coordinate value less than the height threshold Δ is determined as the machine tool table point cloud and is excluded, and the point cloud data greater than or equal to the height threshold Δ is retained as the effective point cloud data of the excess material.

8. The plate scrap identifying method of a laser cutting machine according to claim 7, wherein S36 specifically comprises: S36a, projecting the three-dimensional effective point cloud data in the machine tool coordinate system to the X-Y plane of the machine tool to form a two-dimensional scatter point set; S36b, performing edge point extraction on the two-dimensional scatter point set to obtain an edge point sequence representing the contour of the excess material; S36c, fitting the edge point sequence into a closed two-dimensional contour line composed of straight line segments and circular arc segments smoothly connected; S36d, encapsulating the closed two-dimensional contour line obtained by fitting according to the DXF file format standard to generate the final DXF vector graphics file.

9. The plate scrap identifying method of a laser cutting machine according to claim 8, wherein, In S36c, the edge point sequence is fitted into a closed two-dimensional contour line composed of straight line segments and circular arc segments smoothly connected by using a segmented fitting algorithm, and the specific steps include: S36c.1, traverse the edge point sequence, and preliminarily identify feature points that are likely to be straight line segment and circular arc segment demarcation points according to the turning angle or curvature change between adjacent points; S36c.2, divide the point set segment between adjacent feature points into a to-be-fitted unit, and respectively perform straight line fitting or circular arc fitting on the to-be-fitted unit by using a least square method, and calculate fitting errors of each fitting mode; S36c.3, compare the straight line fitting error and the circular arc fitting error of the same to-be-fitted unit, and select the model with smaller fitting error as the final contour line of the segment according to the error size and a preset precision threshold; and ensure that adjacent straight line segments and circular arc segments meet a tangent connection relationship through geometric constraint conditions; S36c.4, check whether all the line segments and circular arcs generated by fitting are connected end to end to form a complete closed two-dimensional contour line.

10. A plate scrap identification system of a laser cutting machine, characterized by, Comprise: A line laser vision device installed on a cross beam of a laser cutting machine, for moving with the cross beam and performing full-width scanning on a working area of the machine tool; A control module for controlling the scanning path and data acquisition of the line laser vision device; A calibration module for calibrating the line laser vision device through at least two calibration boards to obtain a conversion relationship δ between the line laser vision device coordinate system and the machine tool coordinate system; An image processing module for processing image data obtained by scanning to generate point cloud data; A point cloud processing module for segmenting the point cloud data based on a height threshold to extract an effective point cloud data set representing the excess material shape; A coordinate conversion module for converting the effective point cloud data from the line laser vision device coordinate system to the machine tool coordinate system by using the conversion relationship δ; A contour generation module for performing contour fitting processing on the converted point cloud data to generate a DXF vector graphics file representing the excess material contour.

Images