Automatic damage detection for railroad tracks by real-time DSP based dual analysis system

In this paper we explore a dual track approach for detecting sub critical defects in railroad lines on both the surface and subsurface regions of the rail. The surface defects that are usually detected by manual scanning are detected by applying canny edge detection algorithm to videos of rail surface on a real-time basis. Simultaneously using ground penetrating radar, substructure conditions were observed such as thickness of the ballast and sub ballast layers, variations in layer thickness along the track, pockets of water trapped in the ballast, and soft sub grade from high water content