3D Line Segments Extraction from Semi-dense SLAM

Despite the development of Simultaneous Localization and Mapping (SLAM), there lacks efficient methods for representing and processing their large scale point clouds. In this paper, we propose to simplify the point clouds generated by the semi-dense SLAM using three-dimensional (3D) line segments. Specifically, we present a novel approach for 3D line segments extraction. This approach reduces a 3D line segments fitting problem into two two-dimensional (2D) line segments fitting problems, which takes advantage of both image edge segments and depth maps. We first detect edge segments, which are one-pixel-width pixel chains, from keyframes. We then search 3D line segments of each keyframe along their detected edge pixel chains by minimizing the fitting error on both image plane and depth plane. By incrementally clustering the detected line segments, we show that the resulting 3D representation for the scene achieves a good balance between compactness and completeness.