Motion segmentation based robust RGB-D SLAM

A sparse feature-based motion segmentation algorithm for RGB-D data is proposed which offers us a unified way to handle outliers and dynamic scenarios. Together with the pose-graph SLAM framework, they constitute an effective and robust solution that enable us to do RGB-D SLAM in wide range of situations, although traditionally they have been divided into different categories and treated separately using different kinds of methods. Through comparisons with RANSAC using simulated data and testing with different benchmark RGB-D datasets against the state-of-the-art method in RGB-D SLAM, we show that our solution is efficient and effective in handling general static and dynamic scenarios, some of which have not be achieved before.