An Optical Surface Inspection and Automatic Classification Technique Using the Rotated Wavelet Transform

Abstract

This paper discusses a novel optical nondestructive testing and automatic classification technique using customized image processing techniques. In contrast to conventional spatial analyses, which are highly susceptible to noise and human perception, our proposed transform domain approach provides a high degree of robustness and flexibility in feature selection, and hence a better classification efficiency. The presented algorithm classifies the part-under-test (PUT) into two bins of either acceptable or faulty using transform-domain techniques in conjunction with a classifier. Since the classification is critically dependent on the features extracted from stored images, a sophisticated scalable database was created. The initial database contains 100 parts with various degrees of reflectivity and surface conditions. The transformation algorithm relies on the discrete wavelet transform (DWT) and rotated wavelet transform (RWT) for feature extraction, while a K-nearest neighbor (KNN) classifier is employed for PUT classification. The maximum accurate classification efficiency achieved is 80% and more than 93% by DWT and RWT, respectively.