This paper proposed a method of three dimensions (3D) reconstruction from a wide field of view(FoV) camera system. This camera system consists of two fisheye cameras each with 180° FoV. The fisheye cameras placed back to back to obtain a full 360° FoV. A stereo vision camera is placed to estimate the depth information of anterior view of the camera system. A novel calibration method using unified camera model representation has been proposed to calibrate the multiple camera systems. An effective fusion algorithm has been introduced to fuse multi-camera images by exploiting the overlapping area. Moreover, direct and fast 3D reconstruction of sparse feature matches based on the spherical representation are obtained using the proposed system.
CITATION STYLE
Jamaluddin, A. Z., Jiang, C., Morel, O., Seulin, R., & Fofi, D. (2017). 3D reconstruction from specialized wide field of view camera system using unified spherical model. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10484 LNCS, pp. 495–506). Springer Verlag. https://doi.org/10.1007/978-3-319-68560-1_44
Mendeley helps you to discover research relevant for your work.