Abstract:

Rail inspection is a very important task in railway maintenance for traffic safety issues and in preventing dangerous situations. Monitoring railway infrastructure is an important aspect in which the periodical inspection of rail rolling plane is required. Up to the present days the inspection of the rail is operated manually by trained personnel. A human operator walks along the rail track searching for rail anomalies. This described monitoring way is not more acceptable for its slowness and subjectivity. The results are constrained to the ability of the observer to catch the critical situations. The aim of this paper is to present a vision based technique to detect automatically the presence or absence of the fastening elements that fix the rail to the sleepers. We process the images acquired by a digital line scan camera installed under a train. The images are pre-processed by using wavelet transform with Haar and Daubechies approximation coefficients. We have used two types of pre-processing techniques in order to reduce the computational time and speed up the bolt recognition phase. These coefficients are fed as input to two different neural networks: the first one identify the bolts candidates and the second one validates the bolt recognition process. The final system applied over a long sequence shows a high reliability robustness and good performances.

Key words: Visual Object recognition, Image processing, Wavelet Transform, Neural Network.

	Pier Luigi Mazzeo was born in Casarano –Lecce - (Italy) in 1975. He received a degree in Computer Science Engineering from the University of Lecce in 2001. Since then, he is a collaborator of research at the Italian National Research Council (C.N.R.), Institute of Study on Intelligent Systems for Automation (ISSIA) in Bari (Italy). His current interests are in fields of image and signal processing, wavelet transform, neural networks, pattern recognition and Machine Learning.
	Ettore Stella Received the degree in Computer Science in 1984 at University of Bari (Italy). In 1987 he was researcher of A.S.I. (Italian Space Agency) at C.G.S. (Space Geodesy Center) in Matera (Italy). Since 1990 he is researcher at Institute on Intelligent System for Automation (early Signal and Image Processing Institute). Since 2001 he became Senior Researcher at issia. His main interests are in fields of Artificial Intelligence, Pattern Recognition, Robotics and Machine Learning.
	Massimo Nitti Was born in Bari (Italy) on 1968 and received the degree in Computer Science from Bari University in 1996. From 1997 to 2000 he was employee in the Research Division of MerMec S.p.a. developing railway diagnostic system. Since 2000 he is technician at Institute of Intelligent Systems for Automation (I.S.S.I.A) of Italian National Research Council (C.N.R.) in Bari (Italy). His current interests are in railway application of computer vision, prototype developing, real time application, neural networks.
	Arcangelo DISTANTE received the degree in Computer Science from the University of Bari (Italy) in 1976. He joined the National Nuclear Physics Institute until 1983 where he worked on various theoretical and computation aspects of 3D reconstruction and pattern recognition of nuclear events. Since 1984 he has been working with the Institute for Signal and Image Processing (IESI) of Italian National Research Council (CNR). Currently he is the coordinator of the Robot Vision Group and the Director of the Institute of Intelligent System for Automation (ISSIA - CNR). In 1996 he joined the University of Lecce where he is associate professor in Theory and Practice of Image Processing at the Faculty of Engineering. His current research deals with computer vision, pattern recognition, machine learning, neural computation, robot navigation and architectures for computer vision. Dr. Arcangelo DISTANTE is a member of IAPR and SPIE.