9523123 Anand This research addresses the problem of using an integrated machine vision system in the optimization of the cut and minimization of the scrap losses in the leather industry. This is a cutting stock problem in which machine vision technology is applied to optimize the cuts and minimize scrap. The operation involves cutting smaller irregular leather pieces from a larger piece in the shoe industry. Based on input from shop schedule and the prices of leather parts, a systems approach is applied for planning the cutting operations. The machine vision is used in acquiring the image of the leather pieces and the shoe parts for which the leather pieces are to be attached, polygonizing them, and storing the polygonized images in separate image databases. Using the polygonal images of the leather pieces, the shoe parts, and the manufacturing priorities as input, algorithms are developed for solving the irregular shape cutting stock problem from irregular sheets. The layout information for each leather piece is stored in a layout database and communicated to the cutting operator as a graphic display on a terminal. Alternatively, the layout information is converted to a numerical control (NC) code and fed to a computer numerical control (CNC) controller for automated cutting (laser or waterjet cutting of leather). The general cutting stock problem has applications in a number of industries, including construction, sheet metal, apparel, textile, and glass. Most of the parts in these industries are irregular in shape. Therefore, any methodology or algorithm that improves the utilization of raw materials in these industries can substantially reduce scrap losses and make them more competitive. This project is focused at developing such a tool. The outcome of this research will also help automate some of the critical operations in these industries and thus, reduce overdependency on human labor and experience.
