Manhattan-World Urban Reconstruction from Point Clouds
Manhattan-world urban scenes are common in the real world. We propose a fully automatic approach for reconstructing such scenes from 3D point samples. Our key idea is to represent the geometry of the buildings in the scene using a set of well-aligned boxe
- PDF / 4,073,133 Bytes
- 16 Pages / 439.37 x 666.142 pts Page_size
- 72 Downloads / 339 Views
Visual Computing Center, KAUST, Thuwal, Saudi Arabia [email protected], [email protected], [email protected] 2 College of Electronic and Information Engineering, NUAA, Nanjing, China
Abstract. Manhattan-world urban scenes are common in the real world. We propose a fully automatic approach for reconstructing such scenes from 3D point samples. Our key idea is to represent the geometry of the buildings in the scene using a set of well-aligned boxes. We first extract plane hypothesis from the points followed by an iterative refinement step. Then, candidate boxes are obtained by partitioning the space of the point cloud into a non-uniform grid. After that, we choose an optimal subset of the candidate boxes to approximate the geometry of the buildings. The contribution of our work is that we transform scene reconstruction into a labeling problem that is solved based on a novel Markov Random Field formulation. Unlike previous methods designed for particular types of input point clouds, our method can obtain faithful reconstructions from a variety of data sources. Experiments demonstrate that our method is superior to state-of-the-art methods. Keywords: Urban reconstruction struction · Box fitting
1
·
Manhattan-world scenes
·
Recon-
Introduction
Obtaining faithful reconstructions of urban scenes is an important problem in computer vision. Many methods have been proposed for reconstructing accurate and dense 3D point clouds from images [4,5,26,31]. Besides these point clouds computed from images, there exist an increasing amount of other types of point clouds, e.g., airborne LiDAR data and laser scans. Although these point clouds can be rendered in an impressive manner, many applications (e.g., navigation, simulation, virtual reality) still require polygonal models as a basis. However, few works have addressed the problem of converting these point clouds into surface models. In fact, reconstructing polygonal models from these point clouds still remains an open problem [17,23]. The main difficulty for urban reconstruction from point clouds is the low quality of the data. For example, the obtained point clouds of urban scenes Electronic supplementary material The online version of this chapter (doi:10. 1007/978-3-319-46493-0 4) contains supplementary material, which is available to authorized users. c Springer International Publishing AG 2016 B. Leibe et al. (Eds.): ECCV 2016, Part IV, LNCS 9908, pp. 54–69, 2016. DOI: 10.1007/978-3-319-46493-0 4
Manhattan-World Urban Reconstruction from Point Clouds
55
typically exhibit significant missing regions, as well as uneven point density. This is because the data acquisition process unavoidably suffers from occlusions. Therefore, incorporation of prior knowledge about the structure of the urban scenes into the reconstruction process becomes necessary. In this work, we aim to tackle the problem of reconstructing Manhattan-world urban scenes from the above mentioned point clouds. Such scenes are common in the real world [4]. Existing methods on urban reconstruction fro
Data Loading...
