Cutting path generation system based on visual positioning, cutting equipment and cutting method

Abstract

The invention relates to a cutting path generation system based on visual positioning, cutting equipment and a cutting method. The system comprises a scanning camera control module, an image acquisition and recognition module, an image processing module, a cutting path planning module and a cutting control module. According to the system, by combining a Laplace algorithm and an LBP and GLCM texture extraction technology, the boundary recognition problem under the condition that a pattern and a background are in the same color or low contrast is solved; the path smoothness and the cutting quality are improved by adopting contour optimization and curvature compensation algorithms, the cutting sequence is optimized by combining weighting functions of the area, complexity and overlapping degree, and material deformation is reduced. And motor driving is dynamically adjusted through PID control, and stable control over the speed of the tool bit is achieved. The method has the advantages of being high in recognition precision, high in adaptability, intelligent in path planning and stable in cutting process, the automatic cutting efficiency and the finished product quality of complex patterns or same-color materials are remarkably improved, and the overall cutting precision fluctuation is controlled within + / -2%.

Description

Technical Field

[0001] The present invention relates to the technical field of shoe material processing, and specifically to a cutting path generation system, cutting equipment and cutting method based on visual positioning, which are particularly suitable for application scenarios where a cutting knife is used to accurately cut shoe materials. Background Art

[0002] In the shoe material processing industry, traditional pattern and color material cutting technology has significant defects.

[0003] On the one hand, the lack of precise visual positioning means makes it difficult to accurately identify complex patterns on shoe materials and the boundaries of materials of the same color with extremely small color differences (color difference ≤ 5ΔE), resulting in deviations in the cutting position, seriously affecting the quality and aesthetics of the finished shoe materials.

[0004] On the other hand, traditional methods only rely on a single RGB color space for distinction, and have insufficient texture feature extraction capabilities, making it impossible to effectively identify the boundaries of areas with similar colors but different textures, resulting in low cutting accuracy and efficiency.

[0005] In addition, traditional cutting systems cannot automatically adjust cutting parameters according to the characteristics of different materials such as leather and cloth (such as hardness and ductility), which can easily cause problems such as material tearing and increased tool wear, making it difficult to meet the diversified and high-precision production needs of modern shoe material processing. Summary of the Invention

[0006] To address the challenges of existing technologies, this invention aims to achieve precise cutting of patterned and homochromatic shoe materials through innovative technical means. Specifically, by integrating HSV color space segmentation, LBP (local binary pattern), and GLCM (gray-level co-occurrence matrix) texture feature extraction techniques, this invention overcomes the limitations of traditional single-color feature recognition and achieves sub-pixel contour recognition with a matching accuracy exceeding 99%. Even with identical colors, boundaries can be accurately distinguished through texture differences. Furthermore, a material property weighting model is constructed, combined with Bezier curve smoothing and curvature calculation algorithms, to automatically optimize cutting paths and parameters based on different materials. This reduces tool wear by 30% and cut edge burrs by 50%, comprehensively improving the quality and efficiency of shoe material cutting.

[0007] In a first aspect, the technical solution of the present invention provides a cutting path generation system based on visual positioning, which is applied to a cutting device and includes: The scanning camera control module is used to receive speed information fed back by the encoder set in the cutting equipment, adjust the scanning rhythm and exposure parameters of the object to be cut, ensure that the image acquisition ratio matches the movement speed, obtain the acquired image, and ensure the clarity and stability of the acquired image; An image acquisition and recognition module is used to receive the acquired image and automatically adjust acquisition parameters by analyzing the grayscale distribution and brightness characteristics of the acquired image to ensure that the details of the acquired image are clearly visible and generate a clear image; An image processing module is used to receive the clear image and obtain the pattern boundary contour data of the clear image through a threshold segmentation algorithm or a texture algorithm; wherein When the edge color of the pattern in the clear image is similar to the background, the system first uses the Laplace algorithm to enhance and highlight the texture edge details. It then uses the LBP algorithm and the GLCM algorithm to jointly extract the local and global texture features of the clear image. The weighted cross entropy algorithm and the Dice loss function are combined to optimize model training. This allows the system to accurately identify the pattern boundary contour of the clear image based on texture and shape features even in conditions without obvious color contrast. A cutting path planning module is used to receive the pattern boundary contour data and generate an optimal cutting path by calculating a contour optimization algorithm formula, an arc compensation algorithm formula and a cutting sequence algorithm formula; a cutting control module, configured to control a cutting device to perform a cutting operation according to the cutting path generated by the cutting path planning module; and The data storage and management module is used to store the captured image, the clear image, the pattern boundary contour data and the cutting path information.

[0008] Among them, in the texture algorithm of the image processing module, The Laplace algorithm formula is:

[0009] in, is the Laplace transform result of the image at coordinate (x, y), I is the input image (grayscale value matrix), Refers to the second-order partial derivative of the image in the x direction, which is used to detect horizontal edge changes. Refers to the second-order partial derivative of the image in the y direction, which is used to detect vertical edge changes; The formula of the LBP algorithm is:

[0010] in, Refers to the center pixel coordinates, P is the number of neighborhood pixels (such as 8 neighborhoods), R is the neighborhood radius (pixel distance), Refers to the gray value of the center pixel, Refers to the gray value of the pth pixel in the neighborhood, is a sign function used to compare the grayscale difference between the neighborhood and the central pixel; The formula of the GLCM algorithm is:

[0011] in, is the number of gray levels, is the co-occurrence probability of grayscale values ​​i and j in the specified direction and distance; The formula for the weighted cross entropy algorithm is:

[0012] Where N is the total number of image pixels, C is the number of categories (such as foreground / background), is the weight of category c (used to balance sample imbalance), is the true label (0 or 1) of pixel i in category c, It refers to the probability that the model predicts that pixel i belongs to category c; The formula for the Dice loss function is:

[0013] Among them, X refers to the foreground pixel set predicted by the model, and Y refers to the real foreground pixel set. It refers to the number of intersection pixels between the prediction and the true foreground.

[0014] Among them, the image processing module also includes finding the best match between the template pattern and the pattern at any angle, size, and lighting conditions in the current view, and achieving high-robustness recognition and positioning to ensure the accurate generation of the cutting path when the pattern in the clear image is irregularly placed, the direction and scaling ratio are random, and there is occlusion or lighting changes.

[0015] Among them, through the FAST algorithm formula: Detect corner points and combine SIFT and BRIEF descriptors to construct local features of the pattern, and use the Euclidean distance formula: And the Hamming distance formula: , judge the similarity between patterns and achieve accurate matching between template and target pattern. Even if the pattern has changes in scale, angle, lighting, etc., it can maintain robust recognition accuracy. Where P is the center pixel coordinate, is the i-th pixel in the neighborhood, is the gray value of the center pixel, is the gray value of the neighborhood pixel, thresh is the gray difference threshold, which is used to determine whether it is a corner point. refers to the i-th dimension component of two 128-dimensional SIFT descriptors, Refers to the i-th bit of two binary BRIEF descriptors.

[0016] Among them, through the similarity transformation formula: , affine transformation formula: And the normalized cross-correlation coefficient formula: The system can accurately complete the alignment and matching of the template and the pattern in the view in complex scenarios such as pattern rotation, scaling, offset or lighting changes, ensuring the accuracy and stability of subsequent path generation; are the template point coordinates, is the coordinate of the target point after transformation, s is the scaling factor, θ is the rotation angle, is the translation amount, which controls the position of the template in the image, (x, y) is the template pixel coordinate, (u, v) is the corresponding coordinate of the target image, T is the pixel value of the transformed template, and I is the pixel value of the corresponding area of ​​the target image. and is the pixel mean of the template and target area.

[0017] Among them, in the cutting path planning module, the contour optimization algorithm formula is used: And the arc compensation algorithm formula: Together, they ensure the smoothness and safety of the cutting path, avoiding quality problems caused by sawtooth and sharp corners. They work together to achieve mathematical smoothing of the contour and optimization of physical cutting adaptability, ensuring that even complex patterns can generate high-quality and applicable cutting paths. Furthermore, through the cutting sequence algorithm formula: Combining material characteristics and regional information, scientifically formulate cutting sequence to reduce cutting risks and processing time; are the control points on the contour, which determine the shape of the curve. is a k-order basis function that satisfies local support, t is a parameter variable, and a continuous and smooth curve is generated after traversal, t∈[0,1]; is the area of ​​the region, The larger the ratio, the better the priority, to avoid material deformation when cutting large areas; is the contour complexity, The larger the value, the more complex the shape, and the better it is to cut it first to reduce the number of tool adjustments. is the regional overlap, A higher value indicates a higher proportion of intersection with other areas, which should be cut first to avoid secondary damage; is the weight coefficient, which is adjusted according to the material.

[0018] Among them, in the cutting control module, the PID control algorithm formula is also used: Calculate the motor voltage adjustment amount to make the cutter head speed accurately follow the target speed to avoid the reduction of cutting quality caused by speed fluctuation; is the proportional, integral, and differential coefficient, the proportional term Directly respond to the current error. The larger the error, the larger the output. Accumulate historical errors, eliminate system static errors, calculus terms Predict error changing trends and suppress system oscillations.

[0019] In a second aspect, the technical solution of the present invention provides a cutting device, which includes: An object platform, a crossbeam, a line scan camera, a cutting blade device, a material discharge device, a material receiving device, and a computer, wherein the computer is electrically connected to and controls the operation of the crossbeam, the line scan camera, and the cutting blade device; the computer is integrated with the above-mentioned cutting path generation system based on intelligent visual positioning; wherein, The object platform is used to carry the object to be cut, and a crossbeam is slidably connected to the object platform along the Y axis; A discharge device is provided behind the object platform for conveying the object to be cut, and the discharge device has a smoothing mechanism for smoothing wrinkles of the material; A material collecting device is provided in front of the object platform for collecting the materials after cutting; A line scan camera is fixedly connected to the lower end surface of the crossbeam, and a cutting knife device is slidably connected to the crossbeam along the X-axis; The line scan camera includes a plurality of collection elements arranged along the X-axis, the collection elements including an image sensor, a cylindrical lens, a glass window, an LED light source and a processing chip, the image sensor is electrically connected to the processing chip, and the processing chip is electrically connected to a computer; The cutting blade device includes a connecting plate, a blade holder, and a cutting blade. The connecting plate is slidably connected to the crossbeam along the X-axis, and the blade holder is slidably connected to the connecting plate along the Z-axis. The cutting blade is mounted on the blade holder. A blade position sensor is provided on the blade holder. The blade position sensor is used to monitor the status of the blade and feed back the position information to the computer. An inkjet printing device is also installed on the tool holder. The inkjet printing device includes a U-shaped bracket, a printing cylinder, a connecting block and a printing component. The printing component is a spray valve or an inkjet cartridge, which is used to mark or assist in positioning on the material surface.

[0020] In a third aspect, the technical solution of the present invention provides a cutting method, wherein the cutting method comprises the following steps: Step 1: Fix the paper on the object platform, and use the cutting control module to control the printing cylinder to drive the printing component to print the calibration grid lines on the paper surface; Step 2: The camera control module controls the line scan camera to scan the paper with the grid drawn on it. The image data is received and stitched by the image acquisition and recognition module to complete the horizontal alignment of the camera image. Step 3: Perform camera calibration. The image processing module generates a camera grid and compares it with the printed grid in the captured image. If there is any deviation, the X and Y offsets of the camera grid are automatically adjusted until the origin consistency calibration is completed. The calibration image is saved to the data storage and management module. Step 4: Turn on the power of each device, start the discharge device, and transfer the object to be cut to the object platform through the discharge device. During this process, the smoothing mechanism of the discharge device smoothes the surface wrinkles of the object to be cut. The object platform drives the object to be cut forward, and the line scan camera captures the image of the object to be cut. Step 5: The image acquisition and recognition module receives the image data and feeds it back to the camera control module to dynamically adjust the light source brightness to improve image contrast and edge clarity; Step 6: The captured image is transferred to the image processing module, the design artwork is imported, and after the pattern is selected, it is aligned with the real image through scaling and rotation to generate a feature template. The FAST algorithm is used to extract image features and generate a stable template to achieve multi-angle matching and object positioning; Step 7: The image processing module sends the template matching results to the cutting path planning module, which generates the corresponding optimal cutting path based on the template superimposed image; Step 8: The cutting path is converted into a control instruction and sent to the cutting control module. The cutting control module controls the cutting knife to execute the corresponding path cutting. The cutter head position sensor monitors and feeds back the position information in real time to ensure accuracy. Step 9: When the object to be cut on the object platform is cut, the object platform drives the material to move to the material receiving device, and the cutting knife device continues to cut the subsequent objects to be cut. When the material is cut, the device is turned off and cleaned. When the device is used again, the operation can directly enter the scanning and cutting process without the need to recalibrate the camera.

[0021] The beneficial effects of the technical solution of the present invention are: The cutting path generation system of this invention uses image processing algorithms to complete single-image contour recognition within 500ms. Template matching technology achieves sub-pixel alignment of the product outline with the design template, with a matching accuracy exceeding 99%. Furthermore, a threshold segmentation algorithm based on the RGB / HSV color space can identify homochromatic regions on the material surface with a color difference of ≤5ΔE, enabling color-based cutting and improving material utilization. Furthermore, by extracting microtexture differences (such as fabric weave and leather pore distribution), pattern boundaries can be distinguished even within the same color. Texture features are less sensitive to lighting variations than color features, reducing positioning errors caused by light fluctuations.

[0022] The cutting path generation system of this invention effectively combines a threshold segmentation algorithm in the RGB / HSV color space with a texture algorithm, overcoming the limitations of traditional single-feature segmentation and constructing a multimodal feature fusion framework that can adaptively adjust weights based on the different cutting objects. Furthermore, by combining material properties (such as hardness and ductility) with cutting requirements (such as jagged edges and smooth edges), a genetic algorithm or an AI algorithm is used to generate a cutting path with the shortest path or the fewest turns. This improves cutting efficiency by over 30% while reducing tool wear (extending tool life by 20%). After fitting the original (serrated) profile with cubic B-spline fitting, tool wear is reduced by 30% and burrs on the cut edge are reduced by 50%. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A structural block diagram of a computer of a cutting path generation system according to an embodiment of the present invention; Figure 2 A schematic structural diagram of a cutting device according to an embodiment of the present invention; Figure 3 Schematic diagram of the structure of the cutting blade device in an embodiment of the present invention; Figure 4 Schematic diagram of the transmission mechanism of the cutting knife in an embodiment of the present invention; Figure 5 Schematic diagram of the structure of the acquisition element of the line scan camera in an embodiment of the present invention; Figure 6 Schematic diagram of the structure of an inkjet device in an embodiment of the present invention; Figure 7 This is a flow chart of a cutting method of a cutting device according to an embodiment of the present invention; Figure 8 Schematic diagram of the structure of the discharge device in an embodiment of the present invention. DETAILED DESCRIPTION

[0024] In order to better understand the above technical solutions, the technical solutions of the embodiments of this specification are described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of this specification and the specific features in the embodiments are detailed descriptions of the technical solutions of the embodiments of this specification, rather than limitations on the technical solutions of this specification. In the absence of conflict, the embodiments of this specification and the technical features in the embodiments can be combined with each other.

[0025] The technical solution of this embodiment includes a cutting path generation system based on visual positioning, a cutting device, and a cutting method. The components of the cutting device are as follows: 1. Equipment composition See also Figure 2 、 Figure 3 and Figure 8As shown, the embodiment technical scheme provides a cutting device, which comprises an object platform 1, a crossbeam 2, a line-scan camera 3, a cutting knife device 4, a computer 5, a material collecting device 6 and a material feeding device 7, the computer 5 is electrically connected to and controls the operation of the crossbeam 2, the line-scan camera 3 and the cutting knife device 4 respectively; the computer is integrated with a cutting path generation system based on intelligent visual positioning; wherein, The object platform 1 is used for carrying an object to be cut, and the crossbeam 2 is slidably connected to the object platform 1 along the Y axis; The material feeding device 7 is arranged behind the object platform 1 and is used for conveying the material to be cut, and the material feeding device has the function of smoothing the wrinkles of the material; The material collecting device 6 is arranged in front of the object platform 1 and is used for collecting the cut material; The material feeding device comprises a rack and a roller shaft rotatably installed on the rack, the roller shaft rotates to convey the material, and a smoothing mechanism is installed on the rack, the smoothing mechanism comprises a fixed shaft rotatably connected to the rack, a striking wheel is installed on the fixed shaft, the striking wheel is in the form of a gear, the material extends from top to bottom, the striking wheel contacts the front end surface of the material, the fixed shaft is driven to rotate by a motor, and the fixed shaft drives the striking wheel to strike the material, so that the wrinkles of the material tend to be flat, and the material is conveyed to the material platform by multiple extrusion shafts.

[0026] The line-scan camera 3 is fixedly connected to the lower end surface of the crossbeam 2, and the cutting knife device 4 is slidably connected to the crossbeam 2 along the X axis; As shown in the figure, Figure 5 The line-scan camera 3 comprises a plurality of collection elements arranged along the X axis, the collection elements comprise an image sensor 313, a columnar lens 312, a glass window 311, an LED light source 315 and a processing chip 314 integrated in a PCB, the image sensor 313 is electrically connected to the processing chip 314, and the processing chip 314 is electrically connected to the computer 5.

[0027] The line-scan camera comprises a camera shell and a plurality of collection elements installed in the camera shell, the collection elements comprise a processing chip, an image sensor, a columnar lens, glass and an LED light source, the processing chip is electrically connected to the computer, the image sensor is electrically connected to the processing chip, the columnar lens is arranged between the glass and the image sensor, and one LED light source is arranged on the left and right sides of the columnar lens respectively and faces the detection object, the light source refracted by the detection object passes through the glass, the columnar lens and the image sensor in sequence and is converted into an electric signal by the image sensor.

[0028] As shown in the figure, Figure 3 , Figure 4As shown, in this embodiment, a vibrating knife device is used as the cutting knife device. The cutting knife device 4 includes a connecting plate 41, a knife seat 42 and a cutting knife 43. The connecting plate 41 is slidably connected to the crossbeam 2 along the X-axis, and the knife seat 42 is slidably connected to the connecting plate 41 along the Z-axis. The cutting knife 43 is rotatably connected to the knife seat 42 through a sleeve 45 and a bearing 47. A large gear 46 is provided on the sleeve 45 and is connected to the gear motor for transmission. The gear motor is electrically connected to the computer 5. A knife head position sensor 49 is provided on the knife seat 42. The knife head position sensor 49 is used to monitor the status of the knife head and feed back the position information to the computer 5.

[0029] The cutting knife includes a connecting plate slidingly connected to the crossbeam along the X-axis, a knife seat slidingly connected to the connecting plate along the Z-axis, and a knife head rotatably connected to the knife seat. A through hole is provided on the knife seat for the knife head to pass through, and a shaft sleeve is installed in the through hole. The knife head is installed in the shaft sleeve. A large gear and a bearing are provided on the shaft sleeve. The shaft sleeve is rotatably connected to the knife seat through the bearing. The large gear is located above the knife seat. The large gear is connected to a gear motor for transmission. The gear motor is electrically connected to a computer. A sensor identification notch 48 is provided on the shaft sleeve. A knife head position sensor facing the shaft sleeve is installed on the knife seat. The knife head position sensor is electrically connected to the computer.

[0030] like Figure 6 As shown, an inkjet printing device 44 is also installed on the knife holder 42, and the inkjet printing device 44 includes a U-shaped bracket 442 fixedly connected to the knife holder, and a printing cylinder 441 extending in the up and down directions is installed on the U-shaped bracket 442. The movable end of the printing cylinder 441 is fixedly connected to a connecting block 444 in a limiting rod 443 slidingly connected in the U-shaped bracket 442 in the up and down directions. The connecting block 444 is connected to a printing component 445 by bolts. The printing component 445 can be a spray valve or an inkjet cartridge.

[0031] The technical solution of this embodiment includes a cutting path generation system based on visual positioning, a cutting device, and a cutting method. The components of the computer system in the cutting path generation system are as follows: 2. Computer System Composition like Figure 1 As shown, the technical solution of this embodiment provides a cutting path generation system based on visual positioning, which is applied to cutting equipment, including: The scanning camera control module is used to receive speed information fed back by the encoder installed in the cutting equipment, adjust the scanning rhythm and exposure parameters of the object to be cut, ensure that the image acquisition ratio matches the movement speed, and obtain the captured image to ensure the clarity and stability of the captured image; The image acquisition and recognition module is used to receive the acquired image and automatically adjust the acquisition parameters by analyzing the grayscale distribution and brightness characteristics of the acquired image to ensure that the details of the acquired image are clearly visible and generate a clear image; The image processing module is used to receive a clear image and obtain the pattern boundary contour data of the clear image through a threshold segmentation algorithm or a texture algorithm; When the edge color of a clear image pattern is similar to the background color, the system first uses the Laplace algorithm to enhance and highlight the texture edge details. It then uses the LBP algorithm and the GLCM algorithm to jointly extract the local and global texture features of the clear image. The weighted cross entropy algorithm and the Dice loss function are combined to optimize model training. This allows the system to accurately identify the pattern boundary contour of the clear image based on texture and shape features even in conditions without obvious color contrast. The cutting path planning module is used to receive the pattern boundary contour data and generate the optimal cutting path through the calculation of the contour optimization algorithm formula, the arc compensation algorithm formula and the cutting sequence algorithm formula; a cutting control module, configured to control the cutting equipment to perform cutting operations according to the cutting path generated by the cutting path planning module; and The data storage and management module is used to store the collected images, clear images, pattern boundary contour data and cutting path information.

[0032] The technical solution of this embodiment includes a cutting path generation system based on visual positioning, a cutting device, and a cutting method. The cutting method is as follows: 3. System cutting method like Figure 7 As shown, the technical solution of this embodiment provides a cutting method, wherein the cutting method includes the following steps: Step 1: Secure the paper to the object platform. The cutting control module controls the print cylinder to drive the printing unit, printing a calibrated grid on the paper surface. Specifically, select enough paper to cover the object platform and secure it to the object platform. The cutting control module controls the print cylinder to extend. The movable end of the print cylinder drives the printing unit downward, ensuring contact or a suitable gap between the printing unit and the paper. The cutting control module controls the movement of the blade holder, and the printing unit prints the grid lines on the paper.

[0033] Step 2: The camera control module controls the line scan camera to scan the gridded paper. The image data is received and stitched by the image acquisition and recognition module to complete the horizontal alignment of the camera images. Specifically, the camera control module activates the line scan camera to scan the paper. The collected image data is transmitted to the computer's image acquisition and recognition module. The image acquisition and recognition module adjusts the horizontally misaligned image. Using the first camera lens as a reference, the positions of subsequent camera lenses are adjusted in sequence to complete the camera stitching.

[0034] Step 3: Perform camera calibration. The image processing module generates a camera grid and compares it to the printed grid in the captured image. If there is any deviation, the X and Y offsets of the camera grid are automatically adjusted until the origin consistency calibration is completed. The calibrated image is saved to the data storage and management module. Specifically, the camera is calibrated, and the camera grid is generated by the image processing module. The painting image generated by the image acquisition and recognition module is transmitted to the image processing module. If the generated camera grid coincides with the collected printed grid, it means that the camera and the cutting knife origin are consistent. If there is any misalignment, the image processing module will automatically adjust the camera grid, adjusting the X and Y offsets until it is aligned with the grid obtained by scanning. The image processing module crops the image and saves it to the data storage and management module for subsequent automatic matching and call.

[0035] Step 4: Turn on the power of each device, start the discharge device, and transfer the object to be cut to the object platform through the discharge device. During this process, the smoothing mechanism of the discharge device smoothes the surface wrinkles of the object to be cut. The object platform drives the object to be cut forward, and the line scan camera collects the image of the object to be cut.

[0036] Step 5: Restart the line scan camera to capture an image of the object. The image acquisition and recognition module receives the image data and feeds it back to the camera control module to dynamically adjust the light source brightness to improve image contrast and edge clarity. Specifically, the computer's camera control module starts the line scan camera, controls the light source of each camera, and transmits the captured image to the image acquisition and recognition module. Based on the image display, the image acquisition and recognition module feeds back a signal to the camera control module. The camera control module controls the light source brightness in real time to improve image contrast and ensure that the captured image is clear and has distinct boundaries. With the assistance of the image acquisition and recognition module, the line scan camera completes image capture of the material surface.

[0037] Step 6: The captured image is transferred to the image processing module, where the design artwork is imported. After selecting a pattern, it is aligned with the physical image through scaling and rotation to generate a feature template. The FAST algorithm is then used to extract image features and generate a stable template, enabling multi-angle matching and object positioning. Specifically, the captured image data is transferred to the computer's image acquisition and recognition module, which synchronizes the captured image with the image processing module. The material design artwork is imported, and a single material graphic is selected from the scanned image. The design graphic and the material image are aligned proportionally by moving / rotating them. Click "Generate Template" and the SIFT algorithm is used to extract key image features such as corners, edges, and contours. These are converted into stable feature vectors, creating a visual template with green feature lines. The image processing module automatically optimizes the template quality to ensure clear feature lines and no redundant lines, improving subsequent matching speed and similarity. Based on the SIFT feature vectors and the template as a reference, a multi-angle (0-360°) feature search is performed on the full-width irregular material image. High-speed fitting calculations are performed by the computer's central processing unit (CPU) and graphics processing unit (GPU) to match the feature points of the template with the material image, automatically overlapping the visual templates and achieving precise positioning of multiple images.

[0038] Step 7: The image processing module sends the template matching result to the cutting path planning module, and the path planning module generates the corresponding optimal cutting path based on the template superimposed image.

[0039] Step 8: The cutting path is converted into control instructions and sent to the cutting control module. The cutting control module controls the cutter to execute the corresponding cutting path. The cutter head position sensor monitors and provides feedback on the position in real time to ensure accuracy. Specifically, the cutting path planning module converts the generated cutting path into a control signal and sends it to the cutting control module. Based on this control signal, the cutting control module precisely controls the cutter's gear motor, adjusting the cutter head's speed and direction to ensure the cutter follows the predetermined cutting path. Simultaneously, the cutter head position sensor monitors the cutter head's position in real time and provides feedback to the computer.

[0040] Step 9: When the object to be cut on the object platform is cut, the object platform drives the material to move to the material receiving device, and the cutting knife device continues to cut the subsequent objects to be cut. When the material is cut, the device is turned off and cleaned. When the device is used again, the operation can directly enter the scanning and cutting process without the need to recalibrate the camera.

[0041] The following further describes the computer algorithm in the system of this embodiment in more detail: In the cutting path generation system, cutting device and cutting method based on visual positioning of this embodiment, In the camera control module, the formula , the encoder pulse number is used to calculate the beam movement speed in real time, providing a basis for adjusting the camera scanning interval to ensure that the image acquisition ratio matches the material movement speed, where v is the beam movement speed (mm / s) and p is the encoder pulse increment. The change in encoder pulse increment is: L is the linear displacement corresponding to each motor revolution (mm / revolution), T is the sampling time interval (s), and R is the reduction ratio.

[0042] According to the formula ,Adjust the camera scanning interval to ensure the physical distance between adjacent scanning lines is constant and avoid image stretching or compression, where, is the camera line scanning interval (s), The physical length (mm) corresponding to each line scan.

[0043] According to the formula Control and adjust the camera scanning interval, compensate for the speed fluctuation during the acceleration and deceleration of the beam, and ensure the stability of image acquisition. is the scan interval adjustment amount, is the proportionality coefficient, is the integration coefficient, is the differential coefficient, is the speed error, , is the target speed, is the actual speed, is the integral term of the error, is the differential term of the error; de(t) is the small change in the error, and dt is a differential symbol in mathematics. The meaning is the same as , is to distinguish t in the two formulas.

[0044] In the image acquisition and recognition module, the formula Calculate the image grayscale distribution probability to analyze the image brightness characteristics and determine whether there are overexposed or underexposed areas. for The probability of a pixel appearing is is the kth gray level (0-255), The gray value is N is the total number of pixels in the image.

[0045] By formula , Evaluate the overall brightness and contrast of the image and provide quantitative indicators for parameter adjustment, where is the grayscale mean, is the grayscale variance.

[0046] By formula Iteratively optimize the exposure, gain and other parameters to make the image grayscale distribution concentrated in the target range and improve the accuracy of feature extraction. Stop when the change is less than the threshold or the maximum number of iterations is reached, where is the objective function, 、 is the iteration variable, For passing The exposure parameter value at the next moment after iterative adjustment, is the learning rate, and The objective function J is The partial derivative of guides the iteration direction.

[0047] By formula , Quantify the degree of overexposure / underexposure and trigger the parameter adjustment strategy. >5%, directly reduce exposure; if >5%, directly increase the gain to ensure that image details are visible. is the overexposure ratio, The underexposure ratio.

[0048] In the image processing module, when the color contrast of the cutting boundary is obvious, a threshold segmentation algorithm based on the RGB / HSV color space is used to convert the RGB color space into HSV, separate hue, saturation and brightness, and reduce the impact of lighting changes on color segmentation. By setting the threshold range of each HSV channel, the target color area is extracted and a binary mask is generated. The calculation formulas are as follows: ; ; ; ; Where H is hue, R, G, B are the values ​​of the pixel in the red, green, and blue channels, S is saturation, C is color difference, V is brightness, max=max(R,G,B), min=min(R,G,B), C=max-min; is a binary mask, (H(x,y),S(x,y),V(x,y)) is the HSV value of pixel (x,y), is the hue threshold interval, is the saturation threshold interval, is the brightness threshold range.

[0049] When the border is cut with the same color, the texture algorithm is used to remove noise based on Gaussian filtering (σ=1.0), retain the edge, and enhance the texture details based on the Laplace operator. Even if the pattern and the background color are the same, they can be distinguished by the texture difference, as shown in the Laplace algorithm formula (1).

[0050] in, is the Laplace transform result of the image at coordinate (x, y), I is the input image (grayscale value matrix), Refers to the second-order partial derivative of the image in the x direction, which is used to detect horizontal edge changes. Refers to the second-order partial derivative of the image in the y direction, which is used to detect vertical edge changes.

[0051] Regarding texture feature extraction, the grayscale difference between the center point and the neighborhood is calculated based on LBP, and the neighborhood texture of each pixel is converted into binary code to generate a texture histogram. The LBP features of different radii (R=1, 2, 3) are fused to enhance robustness. The grayscale difference distribution of pixel pairs in the image is quantified based on GLCM. High contrast indicates clear texture edges, and low contrast indicates smooth textures. It can distinguish subtle texture differences under the same color background, such as formula (2) of the LBP algorithm and formula (3) of the GLCM algorithm. ; ; in, Refers to the center pixel coordinates, P is the number of neighborhood pixels (such as 8 neighborhoods), R is the neighborhood radius (pixel distance), Refers to the gray value of the center pixel, Refers to the gray value of the pth pixel in the neighborhood, is a sign function used to compare the grayscale difference between the neighborhood and the central pixel, is the number of gray levels, is the co-occurrence probability of grayscale values ​​i and j in the specified direction and distance.

[0052] Regarding the calculation of weighted cross entropy + Dice loss, the edge pixel weights are strengthened. Even if the pattern and background color are the same, the boundary can be accurately divided by texture, shape and other features, such as formula (4) of the weighted cross entropy algorithm and formula (5) of the Dice loss function. ; ; Where N is the total number of image pixels, C is the number of categories (such as foreground / background), is the weight of category c (used to balance sample imbalance), is the true label (0 or 1) of pixel i in category c, Refers to the probability that the model predicts that pixel i belongs to category c; X refers to the set of foreground pixels predicted by the model, and Y refers to the set of true foreground pixels. Refers to the number of intersection pixels between the prediction and the true foreground, It refers to the product of the number of predicted foreground pixels and the number of actual foreground pixels.

[0053] When generating templates and matching, the scale-invariant feature points are detected by Gaussian difference pyramid, the gradient histogram of the feature point neighborhood is calculated, the main direction is determined, the SIFT feature descriptor is generated, the 128-dimensional vector, the gradient direction distribution of the feature point neighborhood is encoded, and the formula is used. Calculate the sum of the absolute values ​​of the grayscale differences between the central pixel and the neighboring pixels. If it exceeds the threshold, p is determined to be a corner point. For each pair of random pixels , compare its gray value, if The corresponding bit is set to 1, otherwise it is set to 0, thereby generating a BRIEF descriptor for quickly describing the local appearance characteristics of image feature points, where P is the center pixel coordinate, is the i-th pixel in the neighborhood, is the gray value of the center pixel, is the grayscale value of the neighborhood pixel, and thresh is the grayscale difference threshold, which is used to determine whether it is a corner point.

[0054] By formula Calculate the Euclidean distance between two floating-point descriptors. The smaller the distance, the higher the similarity. SIFT descriptors encode local features through gradient direction histograms. Euclidean distance can effectively measure the spatial differences of feature vectors. Refers to the i-th dimension component of two 128-dimensional SIFT descriptors.

[0055] By formula Count the number of different bits in two binary strings. The smaller the Hamming distance, the more similar the descriptors are. Refers to the i-th bit of two binary BRIEF descriptors.

[0056] Through the similarity transformation matrix formula , realize the rotation, scaling, and translation transformation of the template, so that the template matching can be performed when the objects are placed in any direction, randomly select 4 pairs of matching points, calculate the similarity transformation matrix, use the matrix to transform all matching points, count the number of "inliers" that meet the error threshold, iterate until the matrix with the most inliers is found, suppress the influence of outliers, eliminate mismatching points through random sampling consistency (RANSAC), use inliers to estimate the optimal transformation matrix, improve the robustness of template matching to noise and mismatching, and use the affine transformation matrix formula Complete the precise matching of cutting transformation and transparent transformation of the object to be inspected, through the normalized correlation coefficient formula Eliminate the influence of light, confirm the final matching position.

[0057] wherein, is the template point coordinate, is the target point coordinate after transformation, s is the scaling factor, and θ is the rotation angle, is the translation amount, controls the position of the template in the image, (x, y) is the template pixel coordinate, (u, v) is the corresponding coordinate of the target image, T is the pixel value of the template after transformation, and I is the pixel value of the corresponding region of the target image, and is the average pixel value of the template and the target region.

[0058] In the cutting path planning module, the pattern contour coordinates and the same color region mask output by the image processing module are received as the basis for path planning, and the contour optimization algorithm is used to mathematically fit the original contour, eliminate jagged edges caused by noise, and improve cutting quality. At the corner, the curvature k is calculated by the formula When k is greater than the threshold value, a circular arc transition is adopted to avoid material tearing caused by right-angle cutting.

[0059] wherein, is the control point on the contour, which determines the shape of the curve, is the k-order basis function, which satisfies the local support property, t is the parameter variable, and after traversal, a continuous and smooth curve is generated, t ∈ [0, 1]; k is the curvature at the corner of the pattern contour on the material to be cut, x is the horizontal coordinate at the corner of the pattern contour on the material to be cut, and y is the vertical coordinate at the corner of the pattern contour on the material to be cut; is the first derivative of x and y, is the second derivative of x and y, and is the mathematical derivative operator.

[0060] At the same time, according to the material properties, the complexity of the region, and other factors, the cutting order of each region is determined to avoid material deformation, and the weight function formula is , wherein is the cutting priority weight of the i-th region, and the area of the region accounts for a larger proportion and is given priority to avoid material deformation when cutting large regions; the contour complexity is greater, indicating that the shape is more complex, and the cutting is given priority to reduce the number of tool adjustments; the region overlap is higher, indicating that the intersection ratio with other regions is higher, and the cutting is given priority to avoid secondary damage; is the weight coefficient, which is adjusted according to the material. Based on graph theory and heuristic rules, the shortest path for tool movement is calculated to reduce idle travel time, and the optimized path is output as an instruction set that can be executed by the cutting tool.

[0061] In the cutting control module, the cutting knife action is controlled according to the instruction set. The PID controller calculates the control quantity through the proportion (P), integral (I), and differential (D). Dynamically adjust the motor voltage u(t) to make the cutter head speed accurately track the target speed, ensuring fluctuation ≤±2%. Calculate the actual moving speed based on adjacent position points and use it as the feedback value of the PID controller to ensure speed stability during the cutting process and avoid the degradation of cutting quality due to speed fluctuations. Monitor speed stability in real time and trigger PID parameter adjustment when the fluctuation exceeds the threshold.

[0062] in, is the proportional, integral, and differential coefficient, and the proportional term Directly respond to the current error. The larger the error, the larger the output. The integral term Accumulate historical errors, eliminate system static errors, calculus terms Predict error changing trends and suppress system oscillations. is the speed error, , is the target speed, is the actual speed, is the integral term of the error, is the differential term of the error, de(t) is the small change in the error, and dt is a differential symbol in mathematics. , dt have the same meaning, t is , is to distinguish t in the two formulas.

[0063] In summary, the technical solution of this embodiment uses an image processing algorithm to complete contour recognition in a single image within 500ms. Template matching technology achieves sub-pixel alignment of the product outline with the design template, with a matching accuracy exceeding 99%. Furthermore, a threshold segmentation algorithm based on the RGB / HSV color space can identify homochromatic regions on the material surface with a color difference of ≤5ΔE, enabling color-based zoning and cutting to improve material utilization. Furthermore, by extracting microtexture differences (such as fabric weave and leather pore distribution), pattern boundaries can be distinguished even when the color is the same. Texture features are less sensitive to lighting changes than color features, reducing positioning errors caused by light fluctuations.

[0064] The technical solution of this embodiment effectively combines a threshold segmentation algorithm in the RGB / HSV color space with a texture algorithm, overcoming the limitations of traditional single-feature segmentation and constructing a multimodal feature fusion framework that can adaptively adjust weights based on the different cutting objects. Furthermore, by combining material properties (such as hardness and ductility) with cutting requirements (such as jagged edges and smooth edges), a genetic algorithm or AI algorithm is used to generate a cutting route with the shortest path or the fewest turns. This improves cutting efficiency by over 30% while reducing tool wear (extending tool life by 20%). After cubic B-spline fitting of the original (serrated) contour, tool wear is reduced by 30%, and burrs on the cut edge are reduced by 50%.

[0065] In addition, the technical solution of this embodiment also has the following beneficial effects: The multiple groups of acquisition elements arranged along the X-axis in the line scan camera of the technical solution of this embodiment effectively expand the lateral field of view. Through the multi-lens collaborative acquisition mechanism, the distortion error of a single lens is significantly reduced to achieve high resolution in order to distinguish t in the two equations.

[0066] In summary, the technical solution of this embodiment uses an image processing algorithm to complete contour recognition in a single image within 500ms. Template matching technology achieves sub-pixel alignment of the product outline with the design template, with a matching accuracy exceeding 99%. Furthermore, a threshold segmentation algorithm based on the RGB / HSV color space can identify homochromatic regions on the material surface with a color difference of ≤5ΔE, enabling color-based zoning and cutting to improve material utilization. Furthermore, by extracting microtexture differences (such as fabric weave and leather pore distribution), pattern boundaries can be distinguished even when the color is the same. Texture features are less sensitive to lighting changes than color features, reducing positioning errors caused by light fluctuations.

[0067] The technical solution of this embodiment effectively combines a threshold segmentation algorithm in the RGB / HSV color space with a texture algorithm, overcoming the limitations of traditional single-feature segmentation and constructing a multimodal feature fusion framework that can adaptively adjust weights based on the different cutting objects. Furthermore, by combining material properties (such as hardness and ductility) with cutting requirements (such as jagged edges and smooth edges), a genetic algorithm or AI algorithm is used to generate a cutting route with the shortest path or the fewest turns. This improves cutting efficiency by over 30% while reducing tool wear (extending tool life by 20%). After cubic B-spline fitting of the original (serrated) contour, tool wear is reduced by 30%, and burrs on the cut edge are reduced by 50%.

[0068] In addition, the technical solution of this embodiment also has the following beneficial effects: The multiple groups of acquisition elements arranged along the X-axis in the line scan camera of the technical solution of this embodiment effectively expand the lateral field of view. Through the multi-lens collaborative acquisition mechanism, the distortion error of a single lens is significantly reduced, and a high-resolution, seamless image acquisition effect is achieved, which is particularly suitable for the panoramic scanning needs of large-format materials.

[0069] The three-dimensional sliding design of the crossbeam along the Y-axis, the connecting plate along the X-axis, and the knife seat along the Z-axis of the technical solution of this embodiment, combined with the rotation and vibration function of the knife head, enables the cutting knife to move freely in three-dimensional space, accurately adapt to complex cutting paths such as curves and oblique lines, and fully meet the high-precision cutting requirements of diversified patterns.

[0070] The inkjet printing device of the technical solution of this embodiment is equipped with various printing components such as spray valves or inkjet cartridges. It can complete operations such as pre-drawing contour lines, marking or numbering of same-color areas before cutting, realizing an integrated "visual recognition-marking-cutting" operation process, effectively improving process consistency and production efficiency. It automatically adjusts the position of subsequent lenses based on the first camera lens, eliminates parallax between lenses through software algorithms, forms a complete wide-format image, and has a stitching error of less than 0.5 pixels, reducing the time spent on manual calibration. Grid line calibration is used to make the camera and cutting knife share the same origin, ensuring a one-to-one correspondence between image coordinates and physical cutting positions, avoiding cutting offsets caused by coordinate deviations.

[0071] The cutter head position sensor in this embodiment provides real-time feedback on speed and position data. A computer dynamically adjusts motor parameters using a PID control algorithm, strictly controlling cutting speed fluctuations to within ±2%, effectively improving surface smoothness and ensuring consistent cutting quality. After initial calibration, the system memorizes the coordinate relationship between the camera and the cutting origin, eliminating the need for recalibration for subsequent cuts. This reduces single-cut preparation time from 30 minutes to 5 minutes, making it ideal for small-batch, multi-batch production.

[0072] The technical solution of this embodiment is not limited to the use of a vibrating knife, and other cutting tools can be used. It is not limited to the use of a material discharge device. When cutting non-flexible materials, the materials can be directly placed on the object platform.

[0073] The above are only preferred embodiments of the present invention and do not limit the implementation mode and protection scope of the present invention. For those skilled in the art, it should be aware that all solutions obtained by equivalent substitutions and obvious changes made using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A cutting path generation system based on visual positioning, applied to cutting equipment, characterized in that: include: The scanning camera control module is used to receive speed information fed back by the encoder set in the cutting equipment, adjust the scanning rhythm and exposure parameters of the object to be cut, ensure that the image acquisition ratio matches the movement speed, obtain the acquired image, and ensure the clarity and stability of the acquired image; An image acquisition and recognition module is used to receive the acquired image and automatically adjust acquisition parameters by analyzing the grayscale distribution and brightness characteristics of the acquired image to ensure that the details of the acquired image are clearly visible and generate a clear image; An image processing module, configured to receive the clear image and obtain pattern boundary contour data of the clear image through a threshold segmentation algorithm or a texture algorithm; When the edge color of the pattern in the clear image is similar to the background, the system first uses the Laplace algorithm to enhance and highlight the texture edge details, and then uses the LBP algorithm and the GLCM algorithm to jointly extract the local and global texture features of the clear image. Combined with the weighted cross entropy algorithm and the Dice loss function to optimize model training, the system can accurately identify the pattern boundary contour of the clear image based on texture and shape features even in the absence of obvious color contrast. A cutting path planning module is used to receive the pattern boundary contour data and generate optimal cutting path information by calculating the contour optimization algorithm formula, the arc compensation algorithm formula and the cutting sequence algorithm formula; A cutting control module, configured to control the cutting equipment to perform cutting operations according to the cutting path generated by the cutting path planning module; and a data storage and management module for storing the captured image, the clear image, the pattern boundary contour data and the cutting path information.

2. The system according to claim 1, wherein: In the texture algorithm of the image processing module, The Laplace algorithm formula is: ； in, is the Laplace transform result of the image at coordinate (x, y), I is the input image, Refers to the second-order partial derivative of the image in the x direction, which is used to detect horizontal edge changes. Refers to the second-order partial derivative of the image in the y direction, which is used to detect vertical edge changes; The formula of the LBP algorithm is: ； in, Refers to the center pixel coordinates, P is the number of neighborhood pixels, R is the neighborhood radius, Refers to the gray value of the center pixel, Refers to the gray value of the pth pixel in the neighborhood, is a sign function used to compare the grayscale difference between the neighborhood and the central pixel; The formula of the GLCM algorithm is: ； in, is the number of gray levels, is the co-occurrence probability of grayscale values ​​i and j in the specified direction and distance; The formula for the weighted cross entropy algorithm is: ； Where N is the total number of image pixels, C is the number of categories, is the weight of category c, refers to the true label of pixel i in category c, It refers to the probability that the model predicts that pixel i belongs to category c; The formula for the Dice loss function is: ； Among them, X refers to the foreground pixel set predicted by the model, and Y refers to the real foreground pixel set. Refers to the number of intersection pixels between the prediction and the true foreground, It refers to the product of the number of predicted foreground pixels and the number of actual foreground pixels.

3. The system according to claim 2, characterized in that The image processing module also includes finding the best match between the template pattern and the pattern at any angle, size, and lighting conditions in the current view, and achieving highly robust recognition and positioning to ensure accurate generation of the cutting path when the pattern in the clear image is irregularly placed, with random directions and scaling ratios, and there are occlusions or lighting changes.

4. The system according to claim 3, characterized in that According to the FAST algorithm formula: Detect corner points and combine SIFT and BRIEF descriptors to construct local features of the pattern, and use the Euclidean distance formula: And the Hamming distance formula: , judge the similarity between patterns and achieve accurate matching between template and target pattern. Even if the pattern has scale, angle, and illumination changes, it can maintain robust recognition accuracy. Where P is the center pixel coordinate, is the i-th pixel in the neighborhood, is the gray value of the center pixel, is the gray value of the neighborhood pixel, thresh is the gray difference threshold, which is used to determine whether it is a corner point. refers to the i-th dimension component of two 128-dimensional SIFT descriptors, Refers to the i-th bit of two binary BRIEF descriptors.

5. The system according to claim 3, wherein: Through the similarity transformation formula: , affine transformation formula: And the normalized cross-correlation coefficient formula: The system can accurately complete the alignment and matching of the template and the pattern in the view in complex scenarios such as pattern rotation, scaling, offset or lighting changes, ensuring the accuracy and stability of subsequent path generation. are the template point coordinates, is the coordinate of the target point after transformation, s is the scaling factor, θ is the rotation angle, is the translation amount, which controls the position of the template in the image, (x, y) is the template pixel coordinate, (u, v) is the corresponding coordinate of the target image, T is the pixel value of the transformed template, and I is the pixel value of the corresponding area of ​​the target image. and is the pixel mean of the template and target area.

6. The system according to claim 1, wherein: In the cutting path planning module, the contour optimization algorithm formula is used: And the arc compensation algorithm formula: Together they ensure the smoothness and safety of the cutting path, avoiding quality issues caused by jagged edges and sharp corners. They work together to achieve mathematical smoothing of the contour and optimization of physical cutting adaptability, ensuring high-quality and applicable cutting paths even for complex patterns. And through the cutting order algorithm formula: Combining material characteristics and regional information, scientifically formulate cutting sequences to reduce cutting risks and processing time; in, are the control points on the contour, which determine the shape of the curve. is a k-order basis function that satisfies local support, t is a parameter variable, and a continuous and smooth curve is generated after traversal, t∈[0,1]; k is the curvature of the corner of the pattern outline on the material to be cut, x is the abscissa of the corner of the pattern outline on the material to be cut, and y is the ordinate of the corner of the pattern outline on the material to be cut; is the first-order derivative of x and y, is the second-order derivative of x and y, which is the mathematical derivative operation symbol; and Is the mathematical derivative operator symbol; is the area of ​​the region, The larger the ratio, the better the priority, to avoid material deformation when cutting large areas; is the contour complexity, The larger the value, the more complex the shape, and the better it is to cut it to reduce tool adjustments. is the regional overlap, A higher value indicates a higher proportion of intersection with other areas, which should be cut first to avoid secondary damage; is the weight coefficient, which is adjusted according to the material.

7. The system according to claim 1, wherein: In the cutting control module, the PID control algorithm formula is also used: Calculate the motor voltage adjustment amount to make the cutter head speed accurately follow the target speed to avoid the reduction of cutting quality caused by speed fluctuation; is the proportional, integral, and differential coefficient, and the proportional term Directly respond to the current error. The larger the error, the larger the output. Accumulate historical errors, eliminate system static errors, calculus terms Predict error changing trends and suppress system oscillations.

8. A cutting device, characterized in that: include: An object platform, a crossbeam, a line scan camera, a cutting knife device, a material discharge device, a material receiving device and a computer, wherein the computer is electrically connected to and controls the operation of the crossbeam, the line scan camera and the cutting knife device; the computer is integrated with the cutting path generation system based on visual positioning according to claim 1; wherein, The object platform is used to carry the object to be cut, and a crossbeam is slidably connected to the object platform along the Y axis; A discharge device is provided behind the object platform for conveying the object to be cut, and the discharge device has a smoothing mechanism for smoothing wrinkles of the material; A material collecting device is provided in front of the object platform for collecting the materials after cutting; A line scan camera is fixedly connected to the lower end surface of the crossbeam, and a cutting knife device is slidably connected to the crossbeam along the X-axis; The cutting blade device includes a connecting plate, a blade holder, and a cutting blade. The connecting plate is slidably connected to the crossbeam along the X-axis, and the blade holder is slidably connected to the connecting plate along the Z-axis. The cutting blade is mounted on the blade holder. A blade position sensor is provided on the blade holder. The blade position sensor is used to monitor the status of the blade and feed back the position information to the computer. An inkjet printing device is also installed on the tool holder. The inkjet printing device includes a U-shaped bracket, a printing cylinder, a connecting block and a printing component. The printing component is a spray valve or an inkjet cartridge, which is used to mark or assist in positioning on the material surface.

9. A cutting method, characterized in that: The cutting method comprises the following steps: Step 1: Fix the paper on the object platform, and use the cutting control module to control the printing cylinder to drive the printing component to print the calibration grid lines on the paper surface; Step 2: The camera control module controls the line scan camera to scan the paper printed with the grid. The image data is received and stitched by the image acquisition and recognition module to complete the horizontal alignment of the camera image. Step 3: Perform camera calibration. The image processing module generates a camera grid and compares it with the printed grid in the captured image. If there is any deviation, the X and Y offsets of the camera grid are automatically adjusted until the origin consistency calibration is completed. The calibration image is saved to the data storage and management module. Step 4: Turn on the power of each device, start the discharge device, and transfer the object to be cut to the object platform through the discharge device. During this process, the smoothing mechanism of the discharge device smoothes the surface wrinkles of the object to be cut. The object platform drives the object to be cut forward, and the line scan camera captures the image of the object to be cut. Step 5: The image acquisition and recognition module receives the image data and feeds it back to the camera control module to dynamically adjust the light source brightness to improve image contrast and edge clarity; Step 6: The captured image is transferred to the image processing module, the design artwork is imported, and after the pattern is selected, it is aligned with the real image through scaling and rotation to generate a feature template. The FAST algorithm is used to extract image features and generate a stable template to achieve multi-angle matching and object positioning; Step 7: The image processing module sends the template matching results to the cutting path planning module, which generates the corresponding optimal cutting path based on the template superimposed image; Step 8: The cutting path is converted into a control instruction and sent to the cutting control module. The cutting control module controls the cutting knife to execute the corresponding path cutting. The cutter head position sensor monitors and feeds back the position information in real time to ensure accuracy. Step 9: When the object to be cut on the object platform is cut, the object platform drives the material to move to the material receiving device, and the cutting knife device continues to cut the subsequent objects to be cut. When the material is cut, the device is turned off and cleaned. When the device is used again, the operation can directly enter the scanning and cutting process without the need to recalibrate the camera.

Images