Defect detection algorithm of plastic 3D paint surface based on image fusion

Abstract

Abstract In this paper, we proposed a new novel algorithm to detect defects cataloged as dings, dents and scratch on plastic 3D paint surface using patterned illumination. This algorithm first fuses the continuous images through different ways to obtain one fused image, which is obtained by patterned light illumination. Then, a series of morphological operations are used to obtain high-contrast image for subsequent feature statistics and defect screening. Experiments show that real-time inspection can be achieved, and the detection of plastic 3D paint surface defects can be completed within 13s, with an accuracy of 86%.