Rigid registration of 3D indoor scenes is a fundamental yet vital task in various fields that include remote sensing (e.g., 3D reconstruction of indoor scenes), photogrammetry measurement, geometry modeling, etc. Nevertheless, state-of-the-art registration approaches still have defects when dealing with low-quality indoor scene point clouds derived from consumer-grade RGB-D sensors. The major challenge is accurately extracting correspondences between a pair of low-quality point clouds when they contain considerable noise, outliers, or weak texture features. To solve the problem, we present a point cloud registration framework in view of RGB-D information. First, we propose a point normal filter for effectively removing noise and simultaneously maintaining sharp geometric features and smooth transition regions. Second, we design a correspondence extraction scheme based on a novel descriptor encoding textural and geometry information, which can robustly establish dense correspondences between a pair of low-quality point clouds. Finally, we propose a point-to-plane registration technology via a nonconvex regularizer, which can further diminish the influence of those false correspondences and produce an exact rigid transformation between a pair of point clouds. Compared to existing state-of-the-art techniques, intensive experimental results demonstrate that our registration framework is excellent visually and numerically, especially for dealing with low-quality indoor scenes.

In this paper, we propose a novel guided normal filtering followed by vertex updating for mesh denoising. We introduce a two-stage scheme to construct adaptive consistent neighborhoods for guided normal filtering. In the first stage, we newly design a consistency measurement to select a coarse consistent neighborhood for each face in a patch-shift manner. In this step, the selected consistent neighborhoods may still contain some features. Then, a graph-cut based scheme is iteratively performed for constructing different adaptive neighborhoods to match the corresponding local shapes of the mesh. The constructed local neighborhoods in this step, known as the adaptive consistent neighborhoods, can avoid containing any geometric features. By using the constructed adaptive consistent neighborhoods, we compute a more accurate guide normal field to match the underlying surface, which will improve the results of the guide normal filtering. With the help of the adaptive consistent neighborhoods, our guided normal filtering can preserve geometric features well, and is robust against complex shapes of surfaces. Intensive experiments on various meshes show the superiority of our method visually and quantitatively.