Obstacle Detection Of Indoor Vehicle by Estimating 3D Wall from Feature Points
Details
It is necessary to avoid obstacles so that the indoor vehicle may safely reach the destination. A monocular camera is installed in an indoor vehicle, the feature points is obtained from a time series of images, and the self movement of a vehicle is estimated by eight point algorithm. Then, feature points inside both walls is detected as obstacles by reconstructing 3D scene of walls. It is shown that the wall can presume by the experiment.
