Motion removal from moving platforms: An RGB-D data-based motion detection, tracking and segmentation approach

Moving objects are the primary concern for most robot vision applications in dynamic environments. The existence of moving objects can lead to ambiguous decisions such as searching loop closure in visual mapping applications. Eliminating moving objects from the image sequences captured by moving camera is the key challenge. In this paper, a novel approach for moving objects removal using a hand-held RGB-D camera is proposed. Only the visual and depth data are used. No other sensor or prior information is needed in this paper. Our approach can be exploited as a pre-processing stage to filter out data that are associated with moving objects. We test our approach with various ego-motion patterns in different environments. The experimental results demonstrate that our approach can provide a practical solution for motion removal from moving platforms using an RGB-D camera.