Machine Vision in Flexible Cutting of Irregular Leather Workpieces

Abstract

With the increasing demand for high-quality and customized leather products in industries such as automotive, fashion, and furniture, traditional manual cutting methods struggle to meet precision and efficiency requirements. This paper proposes a machine vision-based flexible cutting system for irregular leather workpieces, integrating image processing, contour recognition, and path planning to optimize material utilization and cutting accuracy. A multi-stage algorithm combining edge detection, adaptive thresholding, and deep learning-based defect detection is employed to enhance cutting precision. Experimental results demonstrate that the proposed system achieves a cutting accuracy of ±0.3 mm and reduces material waste by 15% compared to conventional methods. The system's adaptability to varying leather textures and defects makes it suitable for industrial applications requiring high flexibility and automation.