Identification of weld defects using magneto-optical imaging

The weld cracks of the high-strength steel are identified by magneto-optical imaging. The background and basic principle of micro-crack inspection after welding by magneto-optical imaging (MOI) are discussed. The key point is to adopt continuous fuzzy enhancement on the basis of fuzzy set theory, to improve the degree of separation of welding crack and weld and solve the problem of uneven magnetic surface of high-strength steel. The experiment of restoring the magneto-optical image is carried out by using the algorithm of unevenness of crack magneto-optical imaging of high strength steel. After restoration, the PSNR data of magneto-optical image is large, indicating that image quality is greatly improved. According to the characteristics of magneto-optical imaging method, an array crack identification model of laser welding is established by using principal component analysis (PCA) method and support vector machine (SVM). The test results validate that our proposed method can efficiently extract the features of welding cracks and improve the precision of detecting welding cracks.