Reaserch And Implementation Of Automatic-Detection Of Defects In The Weld Images

In the modern industry, welding plays a basic role. The quality of the welding is very important to the life of the products. So detecting welding defects efficiently and accurately has important practical significance and economic value. Now commonly used in industrial non-destructive testing (NDT) methods is the radiography. During the industrial working, widely usedartificial means to deal with assessment piece, but this method greatly influenced by subjective factors, but also inefficient, specially in the real-time detections.Now computer-aided film viewing system is doing online detection and analysis instead of people. It greatly improves the efficiency and andaccuracy of the film viewing system. Just to reduce labor intensity, improve the efficiency of detection, and complete real-time automatic detection, so we use digital detection and automatic defect recognition. In the computer-aided film viewing system, the edge detection of defects occupied a vital position.It is the premise of defect featureextraction, and the results of the film viewing play a decisive role.For these reasons, this paper designs and implements a set of 16-bit TIFF welding image automatic detection system, the main research are as follows:(1) Study the characteristics of 16-bit TIFF images, and implement preprocessing methods on the 16-bit TIFF images.(2) Study a variety of edge detection methods, do many experiments on different types of defects to prove a variety of different detection algorithms, and find out the right algorithm for each type of defects by improving some algorithms.(3) Study the classification and grading of weld defects, and discuss the methods of obtaineing and determining he information of defects. Achieving the automatic identificationof the defects, and meeting the requirements of tion.The result of this paper is that achieving a computer-aided assessment system of radiographic interpreta for the industrial X-ray non-destructive testing, and providing the right algorithm for typical defects (such as: round, strip). This system greatly improves the accuracy and efficiency of X-ray non-destructive testing, and makes some contributions to the development of industrial non-destructive testing.