Robust Vergence with Concurrent Identification of Occlusion and Specular Highlights

摘要

The early detection, classification, and elimination of occlusion and specular highlights which occur frequently in real scenes are of utmost importance. However, most of vergence algorithms which have been proposed in recent years do not take into consideration the effect of occlusion and specular highlights present in real scenes. In this paper, we propose an active exploratory vergence method based on concurrent cross-correlation of multiscale stereo images and an adaptive feedback strategy using weighted normalized cross-correlation to detect and distinguish occlusion and specular highlights concurrently during the vergence process. It not only improves computational efficiency but also leads to the exploration of a better viewing direction and position that overcome the problem created by occlusion and specular highlights. The robustness of the proposed method against occlusion and specular highlights is demonstrated with real scenes.