PATTERNS NONLINEAR CLASSIFIERS OF WELD DEFECTS IN INDUSTRIAL RADIOGRAPHIES

The progresses in researches for the development of an automatic system to analyze weld defects in radiographic images have been evident in the last years. Such research refers to a detailed study of nonlinear classifiers of patterns implemented by artificial neural networks, in order to classify existing weld defects in radiographed weld joints. To increase the reliability of the obtained results, radiographic patterns of the IIW (International Institute of Welding) were used. Geometrical features of the defect classes are used as inputs of the classifiers. The relevance of such features to classify the studied classes was also evaluated. Techniques of analysis of the principal components of discrimination, which were also developed by neural networks, are described in order to visualize in two dimensions the classification problem, as well as the obtained classification performance with the obtained components. The results proved the efficiency of such technique for the data used in the study.

[1]  Dieter Filbert,et al.  CLASSIFICATION OF POTENTIAL DEFECTS IN THE AUTOMATIC INSPECTION OF ALUMINIUM CASTINGS USING STATISTICAL PATTERN RECOGNITION , 2002 .

[2]  T. Warren Liao,et al.  Automated Extraction of Welds from Digitized Radiographic Images Based on MLP Neural Networks , 1997, Appl. Artif. Intell..

[3]  T. Warren Liao,et al.  An automated radiographic NDT system for weld inspection: Part I — Weld extraction , 1996 .

[4]  Redouane Drai,et al.  X-rays image segmentation , 1999, NSIP.

[5]  Yueming Li,et al.  Extraction of welds from radiographic images using fuzzy classifiers , 2000, Inf. Sci..

[6]  A. Kehoe,et al.  Image processing for industrial radiographic inspection : image enhancement , 1990 .

[7]  Simon Haykin,et al.  Neural Networks: A Comprehensive Foundation , 1998 .

[8]  Yasuo Suga,et al.  Application of artificial neural network to discrimination of defect type in automatic radiographic testing of welds , 1999 .

[9]  S. Matsui,et al.  DEVELOPMENT OF AN AUTOMATIC WELD DEFECT IDENTIFICATION SYSTEM FOR RADIOGRAPHIC TESTING , 1993 .

[10]  George V. Lashkia.  Defect detection in X-ray images using fuzzy reasoning , 2001, Image Vis. Comput..

[11]  Yasuo Suga,et al.  Intelligent image processing for abstraction and discrimination of defect image in radiographic film , 1997 .

[12]  P. Muller.  A review of “ Digital Image Processing” (Signal Processing Series)By Kenneth R. Castleman. ( New Jersey: Prentice-Hall Inc., 1979.) Price U.S. $ 29 95. , 1980 .

[13]  Yueming Li,et al.  An automated radiographic NDT system for weld inspection: Part II—Flaw detection , 1998 .

[14]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[15]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[16]  Luis Volnei Sudati Sagrilo,et al.  Evaluation of the relevant characteristic parameters of welding defects and probability of correct classification using linear classifiers , 2002 .

[17]  T. McAvoy,et al.  Nonlinear principal component analysis—Based on principal curves and neural networks , 1996 .

[18]  T. W. Liao,et al.  Detection of welding flaws from radiographic images with fuzzy clustering methods , 1999, Fuzzy Sets Syst..

[19]  R.P. Kraft,et al.  Automated inspection of solder joints-a neural network approach , 1995, Seventeenth IEEE/CPMT International Electronics Manufacturing Technology Symposium. 'Manufacturing Technologies - Present and Future'.

[20]  Wilson H. Tang,et al.  Probability concepts in engineering planning and design , 1984 .