Ballast 3D reconstruction by a matching pursuit based stereo matcher

In the last years the railway association spent much money to check the railway infrastructure. The railway infrastructure is a particular field in which the periodical surface inspection of rolling plane can help the operator to prevent critical situations. The maintenance and monitoring of this infrastructure is an important aspect for railway association. In this paper we describe a visual system able to detect defect where the ballast bed has an anomalous behaviour. The proposed system is a stereo rig, based on two high resolution line scanner TV cameras, installed under the train. To detect defect of the layer ballast we developed stereo vision techniques use matching pursuit method to extract features from images and the similarity function to execute the correspondence between left and right images. Visual inspection can help to increase the control quality and reduce costs to maintenance.