Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 08, 2016 - June 11, 2016
In this paper, we describe an obstacle detection system using Kinect V2. This method is using point clouds providing from the Kinect V2. As a reason to use the Kinect V2, it is very cheap compared to 3D LIDARs. Moreover, I can detect 3D shapes of obstacles with high sampling rates. The point clouds are provided by the Kinect, those plane regions are detected using the LMeDs method. Removing that from all point cloud, we extract point clouds with obstacles information. The experimental results show that robots can detect obstacles, because of extracting point clouds with obstacles information. Our proposed system can detect more than height of 3.0 [cm] obstacles in the outdoor environment.
