Peripheral Expansion of Depth Information via Layout Estimation with Fisheye Camera
Consumer RGB-D cameras have become very useful in the last years, but their field of view is too narrow for certain applications. We propose a new hybrid camera system composed by a conventional RGB-D and a fisheye camera to extend the field of view over
- PDF / 5,706,138 Bytes
- 17 Pages / 439.37 x 666.142 pts Page_size
- 103 Downloads / 178 Views
stract. Consumer RGB-D cameras have become very useful in the last years, but their field of view is too narrow for certain applications. We propose a new hybrid camera system composed by a conventional RGB-D and a fisheye camera to extend the field of view over 180◦ . With this system we have a region of the hemispherical image with depth certainty, and color data in the periphery that is used to extend the structural information of the scene. We have developed a new method to generate scaled layout hypotheses from relevant corners, combining the extraction of lines in the fisheye image and the depth information. Experiments with real images from different scenarios validate our layout recovery method and the advantages of this camera system, which is also able to overcome severe occlusions. As a result, we obtain a scaled 3D model expanding the original depth information with the wide scene reconstruction. Our proposal expands successfully the depth map more than eleven times in a single shot. Keywords: 3D layout estimation · RGB-D · Omnidirectional cameras · Multi-camera systems
1
Introduction
Recent low cost RGB-D cameras have caused a great impact in the fields of computer vision and robotics. These devices usually have a field of view (FoV) too narrow for certain applications, and it is necessary to move the camera in order to capture different views of the scene. However, that is often not easy to achieve or requires to use SLAM algorithms or additional sensors to maintain the system well localized. Recently, some alternatives to extend the FoV of depth cameras using additional elements have been proposed: [1] uses two planar mirrors as a catadioptric extension of the RGB-D device and [2] a consumer set of wide angle lens. Fernandez-Moral et al. [3] proposed a method to calibrate an Electronic supplementary material The online version of this chapter (doi:10. 1007/978-3-319-46484-8 24) contains supplementary material, which is available to authorized users. c Springer International Publishing AG 2016 B. Leibe et al. (Eds.): ECCV 2016, Part VIII, LNCS 9912, pp. 396–412, 2016. DOI: 10.1007/978-3-319-46484-8 24
Peripheral Expansion of Depth Information with Fisheye Camera
(a)
(b)
(c)
(d)
397
(e)
Fig. 1. (a) Image scene as view from a conventional RGB-D camera. (b) Proposed camera system. (c) Depth map projected to the fisheye camera. (d) Expansion of the depth map through spatial layout estimation. (e) From the wide field of view depth map of the scene, we compute a 3D model in a single shot.
omnidirectional RGB-D multi-camera rig. While these approaches are interesting, they are either complex to build and calibrate [1,3], or do not provide good enough depth maps [2]. Here, we propose a new hybrid camera system composed by a depth and a fisheye camera. The FoV of the fisheye is over 180◦ , in contrast with the usual FoV of 43◦ × 57◦ of consumer depth cameras (Fig. 1a). Once the cameras are calibrated, the system (Fig. 1b) is capable of viewing over a hemisphere of color information where the central part of t
Data Loading...
