HouseCraft: Building Houses from Rental Ads and Street Views
In this paper, we utilize rental ads to create realistic textured 3D models of building exteriors. In particular, we exploit the address of the property and its floorplan, which are typically available in the ad. The address allows us to extract Google St
- PDF / 6,192,745 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 42 Downloads / 176 Views
Abstract. In this paper, we utilize rental ads to create realistic textured 3D models of building exteriors. In particular, we exploit the address of the property and its floorplan, which are typically available in the ad. The address allows us to extract Google StreetView images around the building, while the building’s floorplan allows for an efficient parametrization of the building in 3D via a small set of random variables. We propose an energy minimization framework which jointly reasons about the height of each floor, the vertical positions of windows and doors, as well as the precise location of the building in the world’s map, by exploiting several geometric and semantic cues from the StreetView imagery. To demonstrate the effectiveness of our approach, we collected a new dataset with 174 houses by crawling a popular rental website. Our experiments show that our approach is able to precisely estimate the geometry and location of the property, and can create realistic 3D building models. Keywords: 3D reconstruction
1
· 3D scene understanding · Localization
Introduction
Significant effort is being invested into creating accurate 3D models of cities. For example, Google and other map providers such as OpenStreetMap are augmenting their maps with 3D buildings. Architects craft such models for urban/property planning, and visualization for their clients. This process typically involves expensive 3D sensors and/or humans in the loop. Automatically creating accurate 3D models of building exteriors has thus become an important area of research with applications in 3D city modeling, virtual tours of cities and urban planning [1–3]. The problem entails estimating detailed 3D geometry of the building, parsing semantically its important facade elements such as windows and doors, and precisely registering the building with the world’s map. Most existing approaches to 3D building estimation typically require LIDAR scans from either aerial [4,5] or ground-level views [2], or video scans [1]. While these approaches have shown impressive results [1,2,6], their use is inherently Electronic supplementary material The online version of this chapter (doi:10. 1007/978-3-319-46466-4 30) contains supplementary material, which is available to authorized users. c Springer International Publishing AG 2016 B. Leibe et al. (Eds.): ECCV 2016, Part VI, LNCS 9910, pp. 500–516, 2016. DOI: 10.1007/978-3-319-46466-4 30
HouseCraft: Building Houses from Rental Ads and Street Views
501
Fig. 1. We exploit rental ads and a small set of (wide-baseline) Google’s StreetView images to build realistic textured 3D models of the buildings’ exterior. In particular, our method creates models by using only the (approximate) address and a floorplan extracted from the rental ad, in addition to StreetView.
limited to the availability of such sensors. In this paper, our goal is to enable a wider use, where the user can easily obtain a realistic model of her/his house by providing only an approximate address and a floorplan of the building. Towards this goal,
Data Loading...
