Vision based mura detection by using property of line scan camera for black resin-coated steel - Line scan algorithm

Abstract

This paper proposes vision based mura detection algorithm for the black resin-coated steel by using property of line scan camera. The proposed algorithm consists of three parts: preprocessing, selection of threshold value, and finally binarization and post processing. Preprocessing consists of moving average filtering, image partitioning and additional weight for black defects. Second, to distinguish between defect and background we must choose proper threshold value. Finally, we binarize original image by using threshold value and use the image opening and closing to eliminate small noise. The simulation results show detection accuracy of the proposed algorithm.