Large-format Cloth Based On Image Recognition Research On Automatic Generation Of Cutting System Trajectory

In the field of traditional clothing and apparel,there are various pattern cutting processing requirements for fabric products.The traditional manual cutting method is relatively inefficient,and due to the difference in the subjective understanding of the product pattern by the operators,or the manual operation proficiency It is difficult to guarantee the quality of the final product.To this end,this paper has developed a large-format cloth cutting system based on image recognition.The research contents include:Based on the analysis of the requirements of the large-format cloth cutting system,the overall software and hardware scheme was designed;in the process of system calibration,an optimized camera calibration algorithm was proposed based on Zhang's calibration principle,and an automatic image reading method based on bmp format file traversal was designed.,The method of reconstructing the camera internal parameters and adjusting the zoom factor retains more effective pixel information.The Lanczos interpolation method is used to optimize the image correction and restoration effect,and the correction accuracy and efficiency are verified through experiments;in the cloth pattern stitching process,based on the ORB feature The matching algorithm proposes an improved image stitching algorithm,which improves the scale robustness by constructing image pyramids,completes rough matching based on Hamming distance,eliminates mismatches based on RANSAC,and uses linear weight fusion to optimize the stitching effect,and through experiments The matching efficiency and splicing effect are verified;in the cloth pattern contour detection process,an improved contour detection algorithm is proposed based on the Canny detection algorithm,which uses bilateral filtering instead of Gaussian filtering to obtain smooth cloth pattern images,and uses the improved Sobel gradient template to obtain the gradient matrix,Designed interactive threshold detection based on Otsu and hysteresis boundary tracking based on DFS to optimize the effect of cloth pattern contour detection,and verified the anti-noise ability and detail detection performance through experiments.The software and hardware platform of the large-format cloth cutting system was integrated and built,and the overall performance of the system was tested.The results show that the designed large-format cloth cutting system meets the needs of users and can be applied to the actual cutting and production of large-format cloth.