Planar clustering algorithm based on RGB-D data fusion

Plane detection is one of the frame works to the intelligent navigation of unmanned vehicles. In this article, a Kinect camera is used in the smart navigation field. We segment the RGB image first using a mean-shift algorithm and project three-dimensional point cloud, which is based on depth image, to segmented image. Then we cluster the whole plane area according the characteristics of the three-dimensional points set. The experiments show that the method can cluster plane from complex environments, and has a significant meaning to obstacle detection and intelligent avoidance of the next step.