Weld Defect Detection based on Completed Local Ternary Patterns

Contemporarily, the artificial way to review the X-ray film is a common manner to the Quality Examination for Weld. However, this manner has much subjectivity, which may greatly affect the detection efficiency and accuracy, especially after doing a great deal of repetitive mental work. The automatic welding defect inspection system based on X-ray could overcome the shortcomings of artificial marking. Worldwide researchers have made extensive and in-depth research on defect extraction and recognition, and have achieved a great number of effective research results. However, there are still some issues, such as the accurate detection of small defects in uneven background, and effective classification of various defects and automatic identification. For the issues of the weld image based on X-ray, this paper aims to use common texture features to make feature extraction and improved local binary patterns(LBP) as the foundations to propose the completed local ternary patterns (CLTP) to detect weld defects and use SVM classifier based on binary tree to classify and recognize the weld defects to solve the issues on inaccurate detection of small defects and lack of valid classification.