High-precision computer pattern machine processing method

Abstract

The application discloses a high-precision computer pattern machining method, and the method comprises the following steps: template automatic positioning: the positioning mark on the workpiece template is recognized through an infrared scanning sensor, and global machining origin calibration is automatically completed without manual positioning; image acquisition and preprocessing partition: a real shot pattern to be machined is collected through an industrial camera, and grayscale, Gaussian filter denoising and Canny edge detection operations are sequentially performed, the convolution kernel size of the Gaussian filter is set as 5*5, the high-to-low threshold ratio of the Canny edge detection is 3:1, after the preprocessing is completed, all independent pattern areas in the image are automatically recognized through a connected domain marking algorithm, and the number N of pattern types contained in the pattern is counted and output. The application can automatically offset the pattern deformation caused by material slip in the high-speed sewing scene, ensure that the position and shape of all sub-patterns of each product are completely consistent during batch machining, and greatly reduce the defective rate of batch production.

Description

Technical Field

[0001] This invention relates to the technical field of computer pattern making machine processing methods, specifically a high-precision computer pattern making machine processing method. Background Technology

[0002] In the sewing processing fields of clothing, bags, automotive interiors, and home decoration, computerized pattern sewing machines have become the core equipment for realizing automated sewing of complex patterns. As the consumer market continues to demand personalized customization, large-format continuous patterns, and high-precision complex pattern products, the processing mode of traditional computerized pattern sewing machines has gradually exposed many technical shortcomings that are difficult to adapt to, and the contradiction between processing accuracy and efficiency that the industry has long faced is becoming increasingly prominent.

[0003] In the current mainstream computer pattern making process, the pattern import stage generally relies on manual drawing and partitioning in the pattern-making software. This is not only time-consuming and labor-intensive, but also prone to problems such as missed identification and incorrect boundary drawing for micro-sized independent patterns smaller than 5mm. The partition recognition accuracy has long hovered below 90%, directly affecting the accuracy of subsequent stitch generation. At the same time, the workpiece alignment of traditional equipment relies entirely on operators manually adjusting the presser foot position and repeatedly calibrating the origin. The alignment time for a single workpiece is generally 3-5 minutes. Faced with customized orders of multiple varieties and small batches, the changeover efficiency is extremely low, making it difficult to adapt to the pace requirements of flexible production.

[0004] In the embroidery data generation stage, most existing technologies adopt a unified translation mode for the whole image, directly converting the entire pattern into continuous stitches. This cannot differentiate the processing of independent sub-patterns of different materials and styles in the image. Not only does it generate a large number of invalid empty stitches outside the boundaries of sub-patterns, significantly increasing the empty travel distance and increasing the probability of needle and thread breakage, but it also cannot automatically adapt the stitch density for fabrics of different thicknesses: in the case of thick materials, overly dense stitches can easily cause the fabric to wrinkle and the stitches to pile up, while in the case of thin materials, overly sparse stitches can cause the background to show and the pattern to become loose. Operators often need to repeatedly try and adjust the parameters, and novice operators find it difficult to quickly output stable and qualified processing results.

[0005] For large-format processing scenarios involving multiple patterns, traditional computerized pattern making machines cannot achieve fully automated, collaborative stitching of multiple sub-patterns. They can only adopt a step-by-step processing mode, where the workpiece is manually moved and repositioned after processing one sub-pattern before processing the next. This is not only cumbersome and inefficient, but also results in stitch misalignment errors exceeding 0.2 mm at the seams, easily leading to obvious stitch overlap and seam misalignment, significantly reducing the yield rate of large-format decorative fabrics and large bag cut pieces. Furthermore, the motion control strategy of traditional equipment is relatively crude, maintaining a fixed speed throughout the sewing process without targeted deceleration control at the inflection points of curves and polygons. At high speeds, stitches are prone to deviation, and the edge processing accuracy of complex patterns typically only reaches the 0.2 mm level, failing to meet the high-precision processing requirements of high-end leather goods and precision garment accessories.

[0006] Furthermore, most existing computerized pattern making machines lack real-time dynamic needle position compensation mechanisms, relying solely on preset paths before processing. Minor slippage caused by fabric tension during processing cannot be detected and corrected in real time. In batch processing, the pattern position deviation between different workpieces can reach several millimeters, resulting in poor pattern consistency within the same batch and a persistently high defect rate. This needle position deviation problem is further amplified, especially in high-speed sewing scenarios, severely restricting the industry's capacity for high-precision batch processing. In recent years, while the domestic computerized pattern making machine industry has achieved rapid expansion in hardware production capacity, with most basic models now domestically produced, significant gaps remain compared to international high-end models in core technologies such as intelligent image recognition, multi-pattern collaborative control, and dynamic precision maintenance. The technological bottlenecks of traditional processing methods have become a core obstacle hindering the sewing industry's upgrade towards higher precision, higher efficiency, and greater flexibility.

[0007] Therefore, a high-precision computer-aided pattern processing method is proposed to address the above problems. Summary of the Invention

[0008] The purpose of this invention is to provide a high-precision computer-aided pattern processing method, the steps of which are as follows:

[0009] S1. Automatic template positioning: The infrared scanning sensor identifies the positioning marks on the workpiece template and automatically completes the global machining origin calibration, eliminating the need for manual alignment.

[0010] S2. Image Acquisition and Preprocessing Partition: The actual image of the pattern to be processed is acquired by an industrial camera. Grayscale conversion, Gaussian filtering for noise reduction, and Canny edge detection are performed sequentially. The convolution kernel size of the Gaussian filter is set to 5×5, and the high-low threshold ratio of the Canny edge detection is 3:1. After preprocessing, all independent pattern regions in the image are automatically identified by the connected component labeling algorithm. The number of pattern types N contained in the pattern is counted and output. At the same time, a unique region number and independent coordinate bounding box are assigned to each independent pattern. A visual pattern partition preview image is automatically generated. The region boundary and number of each independent pattern are marked with different colors on the preview image. Users can manually correct misidentified partition boundaries.

[0011] S3. Separately convert embroidery data by pattern: The N independent pattern areas after preprocessing are input into the pre-trained embroidery adaptation AI model to generate executable embroidery basic data for each sub-pattern. The data of each sub-pattern is stored independently and does not interfere with each other. The training set of the embroidery adaptation AI model contains 120,000+ partitioned embroidery pattern samples of different materials and styles. The accuracy of stitch generation for a single sub-pattern is not less than 99.5%, and invalid empty stitches outside the sub-pattern boundary are automatically filtered.

[0012] S4. Material Adaptive Parameter Matching: Automatically matches stitch density and stitch direction parameters based on the processing material type corresponding to the current sub-pattern. The adaptive calculation formula for stitch density is as follows:

[0013] ;

[0014] in The target stitch density for the final output of the current sub-pattern, in needles per centimeter; The baseline trace density of the base material; The reference material thickness is in millimeters. The measured thickness of the material in the area where the current sub-pattern is located, in millimeters;

[0015] S5. Independent Parameter Adjustment for Individual Sub-patterns: Receives one-click scaling and rotation commands from the user for a single sub-pattern, performs local coordinate transformation on the basic embroidery data of that sub-pattern, and the transformation formula is:

[0016] ;

[0017] In the formula, The coordinates of the original stitches within the sub-pattern. The coordinates of the target needle mark after transformation. For the first Individual scaling factor for each pattern For the first Each pattern has its own independent rotation angle. For the first The center coordinates of each sub-pattern area; it also supports the adaptation of sub-regional parameters for a single sub-pattern, independently allocating stitch length and thread tension processing parameters without affecting the original data of other sub-patterns;

[0018] S6. Multi-pattern Global Collaborative Stitching: Automatically aligns all sub-pattern embroidery data with adjusted parameters based on the global coordinate system of the actual image, generating a continuous, ultra-long processing path without breaks. The formula for calculating the stitch fusion at the stitching points of adjacent sub-patterns is as follows:

[0019] ;

[0020] In the formula, For the coordinates of the splicing and merging points, The coordinates of the last stitch of the previous sub-pattern. The coordinates of the first stitch of the next sub-pattern. To integrate the weighting coefficient, the value range is 0.3-0.7, enabling continuous one-time stitching of multiple patterns on ultra-long sheets without the need for manual connection of sub-patterns in the middle, completely eliminating the problem of misalignment of seams in traditional step-by-step processing;

[0021] The aforementioned multi-pattern global collaborative splicing supports splicing up to 32 independent sub-patterns simultaneously, with a maximum splicing length of 6 meters. The stitch misalignment error at the splicing point is less than 0.03 millimeters, and there are no visible seam traces.

[0022] S7. Sewing Point Generation and Motion Optimization: A pattern sewing point generation algorithm is applied to the completed global processing path. For each non-closed shape, backstitch and shrink stitch paths are automatically generated at the beginning and end segments. For each closed shape, reinforcement seam points are automatically generated at the starting point of the contour. Simultaneously, single-sided deceleration control is implemented at all inflection points in the path. This single-sided deceleration control starts 3-5 stitches before the sewing needle reaches the inflection point, and automatically restores the baseline sewing speed 1-2 stitches after passing the inflection point to avoid stitch deviation during high-speed sewing. The target sewing speed at the inflection point is calculated using the following formula:

[0023] ;

[0024] in, The target sewing speed at the inflection point, measured in revolutions per minute; The reference sewing speed for a conventional straight path; The actual turning angle of the inflection point, in radians; the edge processing accuracy of complex curves and polygonal patterns is controlled to the level of 0.05 mm;

[0025] S8. Real-time needle position dynamic compensation: During processing, the actual coordinates of the current needle position are collected in real time by a servo encoder and compared with the theoretical coordinates of the preset global pattern path. When the displacement difference is... When the value exceeds the 0.02 mm threshold, a compensation command is immediately sent to the X / Y axis servo motors. The compensation formula is as follows:

[0026] ;

[0027] In the formula, This is the compensation displacement of the servo motor. The theoretical coordinates for the preset pattern path, The system collects the actual coordinates of the needle position in real time; it automatically compensates for the pattern deformation caused by material slippage, ensuring that the position and shape of all sub-patterns of each product are completely consistent during batch processing. The compensation response delay of the real-time needle position dynamic compensation is less than 1 millisecond, and it supports dynamic accuracy maintenance in high-speed sewing scenarios with a maximum speed of 2800 rpm.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0029] This invention utilizes an industrial camera combined with a preprocessing workflow of grayscale conversion, Gaussian filtering, and Canny edge detection, along with an eight-neighbor connected component labeling algorithm, to reliably identify miniature independent pattern regions as small as 2mm × 2mm. It achieves high accuracy in pattern partitioning and generates a visually appealing preview image with color annotations, allowing users to manually correct misidentified boundaries. This completely solves the pain points of traditional manual tracing partitioning, which is inefficient and prone to missing small patterns. It significantly shortens the initial pattern preparation cycle. Furthermore, the front-mounted infrared scanning sensor automatically identifies the workpiece template positioning marks and completes automatic global machining origin calibration, eliminating the need for manual alignment. The alignment time for a single workpiece is reduced from several minutes to several seconds, adapting to the rapid production changeover requirements of batch processing scenarios.

[0030] This invention establishes a pre-trained embroidery adaptation AI model, which generates executable embroidery basic data for each independent sub-pattern separately, and automatically filters out invalid empty stitches outside the boundary. Compared with the traditional method of uniformly translating the whole picture, it greatly reduces the empty stroke and lowers the probability of needle breakage and thread breakage. Moreover, based on the stitch density adaptive formula of the measured material thickness, it can automatically match and adapt the stitch density to the current fabric, avoiding the problem of wrinkles caused by overly dense stitches on thick materials and exposed stitches on thin materials. There is no need for operators to repeatedly manually adjust parameters, and even novices can quickly output qualified processing results.

[0031] This invention uses a stitch fusion formula to perform weighted smoothing on the connection points of adjacent sub-patterns, reducing stitch misalignment errors at the splicing points and eliminating the problems of seam misalignment and stitch overlap caused by the need for manual alignment of sub-patterns in traditional step-by-step processing. This significantly improves the yield rate of large-format decorative fabrics and large bag cut pieces. Furthermore, during processing, a servo encoder compares the actual and theoretical coordinates of the needle position in real time. When the displacement difference exceeds the 0.02 mm threshold, dynamic compensation is immediately triggered. Even in high-speed sewing scenarios, it can automatically offset the pattern deformation caused by material slippage, ensuring that the position and shape of all sub-patterns of each product are completely consistent during batch processing, and significantly reducing the defect rate of mass production. Attached Figure Description

[0032] Figure 1 is a flowchart of the method of the present invention. Detailed Implementation

[0033] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Example 1:

[0035] This embodiment uses a multi-pattern leather embroidery processing scenario for a brand of automobile interior door panel as an example to fully illustrate the high-precision computer pattern processing method of the present invention. The door panel requires continuous embroidery of 32 independently distributed brand logos, decorative lines, and subtle textured patterns on a 5.2-meter-long ultra-soft leather substrate. The finished product must be seamless and misaligned, with stitch accuracy controlled within 0.05mm, and batch processing consistency error not exceeding 0.03mm. Specific implementation steps are attached. Figure 1 As shown:

[0036] S1, Image Acquisition and Preprocessing Section:

[0037] Before processing, the 5-megapixel industrial camera on the equipment is used to vertically capture real-life images of the leather substrate pattern laid flat on the processing table. The camera lens is 35cm away from the substrate surface, and the acquisition resolution is set to 2560×1920. After acquisition, image preprocessing operations are performed in sequence: First, the RGB color image is converted into an 8-bit grayscale image. Then, a 5×5 Gaussian filter convolution kernel is used to complete image denoising, filtering out the fine texture noise on the leather surface. Subsequently, Canny edge detection is performed, with the high threshold set to 150 and the low threshold set to 50, strictly following the 3:1 high-low threshold ratio to accurately extract the contour edges of all patterns.

[0038] After preprocessing, the connected component labeling algorithm of the eight-neighborhood determination rule is called to automatically identify all independent pattern regions in the image. Finally, the number of independent patterns contained in the pattern is N=32, including 6 brand logo patterns, 18 decorative line patterns, and 8 texture patterns. Each pattern is assigned a unique area number from 1 to 32, and an independent coordinate bounding box corresponding to each pattern is generated. The system simultaneously generates a visual pattern partition preview image, using different colors such as red, green, blue, and yellow to mark the area boundary and number of each independent pattern. Operators can confirm through the preview image that there is no recognition error and no manual correction is required.

[0039] Before image acquisition, the device first uses its built-in infrared scanning sensor to identify three pre-set circular positioning marks on the corners of the leather substrate, automatically completing the calibration of the global processing origin without the need for manual alignment.

[0040] S2. Separately convert embroidery data by pattern:

[0041] The 32 independent pattern areas divided in step S1 are input one by one into the pre-trained embroidery adaptation AI model. The training set of this model contains 127,000 embroidery pattern samples covering different materials such as leather, fabric, and non-woven fabric, and different Chinese / industrial styles. The model generates executable embroidery basic data for each sub-pattern, including information such as needle point coordinates, stitch direction, and basic stitch length. The data of all sub-patterns are stored independently in 32 independent cache partitions of the system without interference.

[0042] For the 32 sub-patterns processed in this study, the model achieved a stitch generation accuracy of 99.82%, automatically filtering out all invalid empty stitches outside the boundaries of the sub-patterns, reducing the empty travel distance by 27% compared to the traditional manual plate-making method.

[0043] S3, Material Adaptive Parameter Matching:

[0044] The substrate used in this processing is ultra-soft automotive interior leather. The preset base material is ordinary PU leather with a thickness of 1.0mm. The corresponding baseline stitch density D0 is set to 45 stitches / cm, and the baseline material thickness T0=1.0mm. The actual thickness of the current leather substrate is measured to be 1.6mm using an online thickness sensor. Substituting these values ​​into the adaptive stitch density calculation formula:

[0045] needles per centimeter;

[0046] After automatic rounding, the system sets the basic stitch density of all current sub-patterns to 36 stitches / cm, avoiding fabric wrinkling and thread breakage caused by excessively dense stitches on thick leather. For three thin decorative areas with a thickness of only 0.8mm, the system individually identifies the measured thickness T=0.8mm for that area and calculates it using the formula. The stitch density is automatically adjusted to 50 stitches per centimeter to ensure that the embroidery texture in thin areas is full enough.

[0047] S4. Independent parameter adjustment for each pattern:

[0048] Based on the door panel design requirements, the operator performed independent parameter adjustments on the brand logo sub-pattern numbered 5: setting the independent scaling factor k5 of this sub-pattern to 1.2 and the independent rotation angle... Set to 15°, the center coordinates of the sub-pattern area. The coordinates are (125.6mm, 372.3mm). Select any original needle point coordinate within this sub-pattern. =(110.0mm, 360.0mm), substitute into the local coordinate transformation formula:

[0049] ;

[0050] Complete the overall enlargement and rotation operation of this sub-pattern.

[0051] Meanwhile, the operator adjusted the stitch length to 2.5mm and the thread tension to 12N for the decorative line sub-pattern numbered 12. This operation only modified the local parameters of this sub-pattern and did not affect the original embroidery data of the other 31 sub-patterns. The parameter adjustment took only 10 seconds.

[0052] S5, Multi-pattern Global Collaborative Stitching:

[0053] All 32 sub-pattern embroidery data with adjusted parameters were automatically aligned using the global coordinate system of the actual photograph as a reference, generating a continuous, uninterrupted processing path for an ultra-long format. The total processing length reached 5.2 meters, meeting the splicing capability requirement of a maximum 6-meter format. The splicing positions of two adjacent sub-patterns numbered 7 and 8 were selected, and the coordinates of the last stitch of the previous sub-pattern were used. =(215.7mm, 420.2mm), the coordinates of the first stitch of the next sub-pattern. =(216.5mm, 420.4mm), the fusion weighting coefficients will be used. Set it to 0.5 and substitute it into the stitch fusion calculation formula:

[0054] ;

[0055] The stitches of the two sub-patterns were smoothly transitioned at the fusion point. The measured stitch misalignment error at all 31 splicing positions was less than 0.025mm, far below the preset threshold of 0.03mm. There were no visible seam marks at the splicing point, achieving one-time continuous stitching of a 5.2-meter ultra-long pattern without the need for manual splicing of sub-patterns in the middle, completely eliminating the seam misalignment problem of traditional step-by-step processing.

[0056] Any technical solution that achieves the above-mentioned technical effects by utilizing the technical solutions described in this invention, or by designing similar technical solutions by those skilled in the art under the inspiration of the technical solutions described in this invention, falls within the protection scope of this invention.

Claims

1. A high-precision computer-aided pattern processing method, characterized in that: The steps of this method are as follows: S1. Image Acquisition and Preprocessing Partitioning: The actual image of the pattern to be processed is acquired by an industrial camera, and grayscale conversion, Gaussian filtering for noise reduction, and Canny edge detection are performed sequentially. After preprocessing, all independent pattern regions in the image are automatically identified by a connected component labeling algorithm. The number of pattern types N contained in the pattern is counted and output. At the same time, a unique region number and independent coordinate bounding box are assigned to each independent pattern. The connected component labeling algorithm adopts the eight-neighbor determination rule and can identify independent pattern regions with a minimum size of 2mm×2mm. The pattern partitioning recognition accuracy is not less than 99%. S2. Separately convert embroidery data by pattern: The N independent pattern areas after preprocessing are input into the pre-trained embroidery adaptation AI model to generate executable basic embroidery data for each sub-pattern. The data of each sub-pattern is stored independently and does not interfere with each other. S3. Material Adaptive Parameter Matching: Automatically matches stitch density and stitch direction parameters based on the processing material type corresponding to the current sub-pattern. The adaptive calculation formula for stitch density is as follows: ； in The target stitch density for the final output of the current sub-pattern, in needles per centimeter; The baseline trace density of the base material; The reference material thickness is in millimeters. The measured thickness of the material in the area where the current sub-pattern is located, in millimeters; S4. Independent Parameter Adjustment for Individual Sub-patterns: Receives one-click scaling and rotation commands from the user for a single sub-pattern, performs local coordinate transformation on the basic embroidery data of that sub-pattern, and the transformation formula is: ； In the formula, The coordinates of the original stitches within the sub-pattern. The coordinates of the target needle mark after transformation. For the first Individual scaling factor for each pattern For the first Each pattern has its own independent rotation angle. For the first The center coordinates of each sub-pattern area; it also supports the adaptation of sub-regional parameters for a single sub-pattern, independently allocating stitch length and thread tension processing parameters without affecting the original data of other sub-patterns; S5. Multi-pattern Global Collaborative Stitching: Automatically aligns all sub-pattern embroidery data with adjusted parameters based on the global coordinate system of the actual photograph, generating a continuous, ultra-long processing path without breaks. The stitch fusion calculation formula at the stitching points of adjacent sub-patterns is as follows: ； In the formula, For the coordinates of the splicing and merging points, The coordinates of the last stitch of the previous sub-pattern. The coordinates of the first stitch of the next sub-pattern. To integrate the weighting coefficient, the value range is 0.3-0.7, enabling continuous one-time stitching of multiple patterns on ultra-long sheets without the need for manual connection of sub-patterns in the middle, completely eliminating the problem of misalignment of seams in traditional step-by-step processing; S6. Sewing Point Generation and Motion Optimization: The completed global processing path is processed using a pattern sewing point generation algorithm. For each non-closed shape, backstitch and shrink stitch paths are automatically generated at the beginning and end segments. For each closed shape, reinforcement seam points are automatically generated at the starting point of the contour. Simultaneously, single-sided deceleration control is performed at all inflection points in the path. The target sewing speed at the inflection point is calculated using the following formula: ； in, The target sewing speed at the inflection point, measured in revolutions per minute; The reference sewing speed for a conventional straight path; The actual turning angle of the inflection point, in radians; the edge processing accuracy of complex curves and polygonal patterns is controlled to the level of 0.05 mm; S7. Real-time needle position dynamic compensation: During processing, the actual coordinates of the current needle position are collected in real time by a servo encoder and compared with the theoretical coordinates of the preset global pattern path. When the displacement difference is... When the value exceeds the 0.02 mm threshold, a compensation command is immediately sent to the X / Y axis servo motors. The compensation formula is as follows: ； In the formula, This is the compensation displacement of the servo motor. The theoretical coordinates for the preset pattern path, It collects the actual coordinates of the needle position in real time; it automatically cancels out the pattern deformation caused by material slippage, ensuring that the position and shape of all sub-patterns of each product are completely consistent during batch processing.

2. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: In step S1, the convolution kernel size of the Gaussian filter is set to 5×5, and the high-low threshold ratio of the Canny edge detection is 3:

1.

3. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The connected component labeling algorithm described in step S1 uses an eight-neighborhood determination rule and can identify independent pattern regions with a minimum size of 2mm×2mm. The pattern partitioning recognition accuracy is no less than 99%.

4. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: After the preprocessing in step S1 is completed, a visual preview image of the pattern partitions is automatically generated. The area boundaries and numbers of each independent pattern are marked with different colors on the preview image, and users can manually correct misidentified partition boundaries.

5. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The embroidery adaptation AI model mentioned in step S2 has a training set containing 120,000+ embroidery pattern samples of different materials and styles. The accuracy of stitch generation for a single sub-pattern is no less than 99.5%, and invalid empty stitches outside the sub-pattern boundary are automatically filtered out.

6. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The multi-pattern global collaborative splicing described in step S5 supports splicing up to 32 independent sub-patterns at the same time, with a maximum splicing area length of 6 meters, a stitch misalignment error of less than 0.03 mm at the splicing point, and no visible seam traces.

7. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The single-sided deceleration control described in step S6 starts deceleration 3-5 stitches before the sewing needle reaches the inflection point, and automatically restores the reference sewing speed 1-2 stitches after passing the inflection point, thus avoiding stitch deviation during high-speed sewing.

8. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The real-time needle position dynamic compensation described in step S7 has a compensation response delay of less than 1 millisecond and supports dynamic accuracy maintenance in high-speed sewing scenarios with a maximum speed of 2800 revolutions per minute.

9. The high-precision computer-aided pattern processing method according to claim 1, characterized in that: The method also includes an automatic template positioning step before image acquisition. The positioning marks on the workpiece template are identified by an infrared scanning sensor, and the global machining origin is automatically calibrated without the need for manual alignment.

Images