Defect detection on unpainted car body

M. T. Mustafa

Department of Mathematical Sciences

King Fahd University of Petroleum & Minerals

Abstract

The talk presents results of an industrial project from automobile industry. The purpose is to develop a fast defect detection method which is accurate enough in detection of defects as well as is well-suited for online inspection process.

During the manufacturing process, car bodies can have small defects, for instance ripples or bumps. Such defects being very small in depth are not visible until cars are painted. However, it is expensive to repair the defects at this stage. A model is developed to detect ripples or bumps on unpainted car body by analyzing the scanned images, of unpainted car body, using smoothing splines.