This paper provides an insight on different classical methodology of weld defect detection from the industrial radiographic images. Three well accepted methods such as fuzzy, Artificial Neural Network (ANN) and Case Base Reasoning (CBR) were designed to individually observe the defect detection. Based on the result obtained through the attempted approaches, likelihood of defect detection is observed. A hybrid rule based approach is formulated to improve the accuracy on weld defect detection. © Research India Publications.