Design And Implementation Of Automatic-Detection System Of Circular Defects In The Weld Images

In modern industry, weld technology plays a irreplaceable role. With the developing of Computer Science and Technology, it’s more mature that the new technology applied to recognize the weld defects. Now, some computer intelligent assessment have been appeared in the industry detection. The intelligent assessment system improved the efficiency and accuracy greatly, and it overcome the artificial instability.The defect image recognition technology takes an important position in the assist system, as for the defect image recognition technology. Weld defects in the image included cracks, incomplete penetration, and other types of circular defects, the different defects have their own different characteristics, according to their characteristic values. The characteristic values could identify the defects and determine the types of defects. Although it had been a variety of defects in weld image recognition method, but it could cause the edge disappeared.For these reasons, the paper designs and implements a detecting system about circular weld defects.The main research as follows:(1) Analyzing the characteristics of circular weld defects, and implementing the method of reading and displaying circular weld defects.(2) Studying the classification and grading of circular weld defects, and implementation the wav of identification of circular defects.(3) Implementation the common edge detection,such as Roberts edge detection,Sobel edge detection, Prewitt edge detection, Krisch edge detection and Canny edge detection operator.In this paper, there is a improved Canny operator.(4) It added the image pre-processing in order to improve and enhance the effect of edge detection.The paper achieved a computer-aided assessment system of circular weld defects. It improved the accuracy and efficiency greatly.As well as,it played a role in the development of non-destructive testing.