Rail Identification Using Camera and Millimeter-Wave Radar Data

It is essential for the railway active obstacle detection system to identify rail before judging whether the obstacle occupies the area. A good identification method has to face harsh environment including foggy, rainstorm and dim condition and identify various kinds of rails. What’s more, algorithms need to respond rapidly and not take up too much computing resources because obstacle identification consumes the most. A new approach based on camera and millimeter wave radar data is given in this paper. The radar point cloud data can be obtained by the millimeter wave reflecting on the trend of the rail. A Clustering algorithm is used to process radar data and Haar features are used to identify rail in camera data. We use Kalman filter to fuse the two data results and predict the current rail position. The proposed method is verified on the main line of the subway and the experimental results indicate that the algorithm is valid.