Identification and Classification of Defects at Different Leather processing Stages using Vision Based System

Abstract

Leather is a natural material with varying texture as in case of raw leather as well as processed one. The leather at any stage of processing may prone to various defects. Leather inspection has been recognized as a very complex problem on the area of texture classification. Like most of the natural surface, values have a high variation, framing a pseudo-arbitrary structure, with an analytical distribution. Leather deformities can have numerous shapes and sizes relying upon the imperfection. Identifying all these defects at various leather processing stages were a major challenge in leather industries. In this proposed method, a novel methodology has been developed that automatically and efficiently inspects various leather defects that occur at various leather processing stages. Further based on some of the first and second order statistical features, the common type of defects occurring in leather such as defect due to insects, scar, stain, cut and crack and can also be classified by using the proposed method and the results obtained were plotted in Graphical User Interface (GUI).