Motion recovery using the image interpolation algorithm and an RGB-D camera

Abstract

In this paper, a simple and robust method to recover ego-motion from a RGB-D camera is proposed. For our algorithm, no assumptions about the moving object or dynamic scenes are needed. Using the optical flow fields computed from the image sequence and depth information obtained from the RGB-D camera we estimate the ego-motion parameters. Our experiments have shown that using the intensity image and corresponding depth value can be very helpful for ego-motion estimation. We also show the experimental results of different scenarios, such as translation only and combination of translation and rotation.