Radioscopic and Ultrasonic Data Fusion Via the Evidence Theory
Fusion de Données Radioscopiques et Ultrasonores via la Théoriede L'Évidence
OPEN ACCESS
A data fusion method, based on the the Dempster-Shafer evidence theory, is presented that aims at improving the reliability of taking a decision by simultaneously exploiting complementary information from two different sources . Within this framework, both discrete and continuous hypotheses are studied in a systematic manner. The data fusion method is applied to problems of improving the reliability of Non Destructive Testing (NDT) using real-time x-ray (RX)radioscopy and ultrasounds (US).It is illustrated with the aid of radioscopic and ultrasonic data corresponding to the same test objects . The results of detection, identification and quantification of defects are discussed and compared in detail, in both monomodality and multimodality.
Résumé
A data fusion method, based on the the Dempster-Shafer evidence theory, is presented that aims at improving the reliability of taking a decision by simultaneously exploiting complementary information from two different sources . Within this framework, both discrete and continuous hypotheses are studied in a systematic manner. The data fusion method is applied to problems of improving the reliability of Non Destructive Testing (NDT) using real-time x-ray (RX)radioscopy and ultrasounds (US).It is illustrated with the aid of radioscopic and ultrasonic data corresponding to the same test objects . The results of detection, identification and quantification of defects are discussed and compared in detail, in both monomodality and multimodality.
Data fusion, Dempster-Shafer evidence theory, non destructive testing, real-time x-ray radioscopy, ultrasonic imaging .
Mots clés
Fusion de données, théorie de l'évidence de Dempster-Shafer, contrôle non destructif, radioscopie numérique, contrôle ultrasonore.
[Appr-91] A. Appriou, « Probabilités et incertitudes en fusion multicapteurs », Revue Scientifique et Technique de la Défense, Vol. 11,No. 1, p. 27-40, 1991.
[Appr-93] A. Appriou. « Formulation et traitement de l'incertain en analyse multisenseurs », 14eme Colloque GRETSI,p. 951-954, Sept 1993.
[Bloc-96] I. Bloch, « Some aspects of Dempster-Shafer evidence theory for classification of multi-modality medical imaging taking partial volume effect into account », Pattern Recognition Letters, Vol. 17, No. 8, p. 905-919, July 1996.
[Bogl-87] P. Bugler, « Shafer-Dempster Reasoning with applications to Multisensor Target Identification Systems », IEEE Transactions on Systems, Man and Cybernetics, Vol. 17, No. 6, p. 968-977, Dec 1987.
[Case-92] W.F. Caselton, W. Luo, « Decision making with imprecise probabilities: Dempster-Shafer Theory and Application », Water RessourcesResearch, Vol.28, No. 12, p. 3071-3083, Dec 1992.
[Clar-90] J.J. Clark, A.L. Yville, « Data Fusion for Sensory Information Processing Systems », Kluwer Academic Publishers, Boston, 1990.
[Demp-68] A.P. Dempster, « A Generalization of Bayesian Inference », Journal of the Royal Statistical Society, Vol. 30, p. 205-247, 1968.
[Deno-95] T. Denoeux, « A k-nearest neigbour classification rule based on Dempster-Shafer theory >>, IEEE Transactions on Systems, Man and Cybernetics, Vol. 25, No. 05, 1995.
[Drom-94] A. Dromigny, «Etude des différentestechniques de fusion d'images - Applicationaux images cardiaques parrésonance magnétique », Mémoire de DEA, INSA- Lyon, 50 p., Juillet 1994.
[Geor-94] B. Georgel, B. Lavayssiere,« Fusion de données; un nouveau concept en CND », 6e Conf, Européenne sur les CND, Nice, 24-28 octobre, p. 31-35, 1994.
[Gros-94] X.E.Gros, P . Strachan, D. Lowden, I. Edwards, «NDT Data Fusion", 6e Conf.Européenne sur les CND, Nice, 24-28octobre,p. 355-359, 1994.
[Gros-95] X.E. Gros,P. Strachan, D. Lowden, « Theory and Implementation of NDT Data Fusion », Results on Non Destructive Evaluation, Vol.6,Springer Verlag,p. 227- 236, 1995.
[Henk-881 S. Henkind, M.C. Harrison, « An Ananlysis of Four Uncertainty Calculi », IEEE Transactions on Systems, Man and Cybernetics, Vol. 18, No. 5, p. 700-714, Sept1988.
[Joha-94] K. Johannsen, S. Heine, C. Nockermann, « New Data Fusion Techniques for the Reliability Enhancement of NDT », 6e Conf. Européenne sur les CND, Nice, 24-28 octobre,p. 361-365, 1994.
[Kaft-95] V. Kaftandjian, « Reconnaissance automatique des défauts dans les produits métalliques par radioscopie numérique », Thèse de l'INSA Lyon, 1995.
[Lee-90] R.H. Lee, R. Leahy, «Multi-SpectralClassification of MR Images Using Sensor Fusion Approaches », SPIEMedical Imaging IV : Image Processing, Vol. 1233, p. 149-157, 1990.
[Luo-89] R.C. Luo, M.G. Kay, «Multisensor Integration and Fusion in Intelligent Systems », IEEE Transactions on Systems, Man and Cybernetics, Vol. 19, No. 5, p. 901-931, Oct 1989.
[Ray-92] C. Ray Smith, « A Bayesian Approach to multisensor fusion », SPIE, Vol. 1699, p. 285-299, 1992.
[Shaf-76] G. Shafer, « A mathematical theory of evidence », Princeton University Press. 1976. [Smet-94] P. Smets, R. Kennes, « The Transferable Belief Model », Artificial Intelligence,Vol.66, p. 191-243, 1994.
[Sure-941 F. Sureda, I. Bloch, C. Pellot, A. Herment, « Reconstruction 2D de vaisseaux sanguins par fusion de données à partir d'images angiographiques et échographiques », Traitementdu Signal,Vol. 11, No. 6, p. 525-540, 1994.
[Zade-88] L.A. Zadeh, « Fuzzy logic », IEEE Computer Magazine, p . 88-93. avril 1988.
[Zhu-901 Y.M. Zhu, D. Babot, G. Peix, « A quantitative comparison between linear x-ray sensitive array and image-intensifier systems », NDT International, Vol. 23, No. 4, p. 214-220, Aug 1990.