Reconstructing illumination environment by omnidirectional camera calibration

  • Authors:
  • Yong-Ho Hwang;Hyun-Ki Hong

  • Affiliations:
  • Dept. of Image Eng., Graduate School of Advanced Imaging Science, Multimedia and Film, Chung-Ang Univ., Seoul, Korea;Dept. of Image Eng., Graduate School of Advanced Imaging Science, Multimedia and Film, Chung-Ang Univ., Seoul, Korea

  • Venue:
  • AI'06 Proceedings of the 19th Australian joint conference on Artificial Intelligence: advances in Artificial Intelligence
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a novel approach to reconstruct illumination environment by omnidirectional camera calibration. The camera positions are estimated by our method considering the inlier distribution. The light positions are computed with the intersection points of the rays starting from the camera positions toward the corresponding points between two images. In addition, our method can integrate various synthetic objects in the real photographs by using the distributed ray tracing and HDR (High Dynamic Range) radiance map. Simulation results showed that we can generate photo-realistic image synthesis in the reconstructed illumination environment.