Thresholding Segmentation and Classification in Automated Ultrasonic Testing Image of Electrical Contact

Abstract

This paper is concerned with the thresholding segmentation of ultrasonic C-scan image for defect with artificial intelligence and support vector machines techniques. Wavelet based method is used to remove the noise of image. Segmentation techniques based on thresholding were performed in order to segment the defect in ultrasonic testing image. Experimental results show that this segmentation technique performs well in finding defects in C-scan image of electrical contact.