Image processing techniques for classification of linear welding defects

Detect and classify of welding defects is one of the most important factors in quality of welding. Researchers have done lots of attempts to develop an automatic (or semiautomatic) system for the detection and classification of weld defects in continuous welds using radiography. I have developed a new method for filtering and segmenting radiographic images of welding to describe an automatic system for classification of welding defects and compared with KNN and SVM classifiers. The classification used in this research is a new method as well. The linear defects such as lack of penetrations, incomplete fusion and external undercut were classified and recognized. Experimental results have shown this classification method is useful for lengthy defects and have been obtained through the mentioned method is better than the two classifiers methods.