3D Terrain Reconstruction for Patrol Robot Using Point Grey Research Stereo Vision Cameras

This paper uses Bumblebee stereo vision system to reconstruct the 3D terrain for patrol robot. Firstly, with the selected points in the disparity image, the ground plane equation of the current frame is calculated by the RANSAC (Random Sample Consensus) algorithm, and then estimated and predicted by the Kalman filter. Secondly, the elevation image is obtained through the distance between the obstacle and the ground plane, which is calculated on basis of the formula from the point to the plane. Finally, the noise points in the elevation image are removed through erosion and dilation operations, and the location and size of obstacles are determined via the connected component analysis method. The actual test in substation environment verified the reliability of the system.