Image processing algorithms for video-based real-time railroad track inspection

The automated railroad track inspection described in this paper meets the requirements of real-time capability and cost-effective implementation. It is based on the analysis of images provided by a high-speed line scan camera. The boundary detection is performed by a simply-structured edge detection algorithm where sensitivity can be adapted to the regional brightness of a scene.