Title :
A simple and real-time moving object detection invariant to cast shadow
Author :
Tae-Hyun Oh ; In So Kweon
Author_Institution :
Dept. of Electr. Eng., KAIST, Daejeon, South Korea
Abstract :
We present a moving objection detection method robust to shadow effects and illumination changes on a static camera. This study addresses the effect of the declining sun at evening, which produces elongated shadow of objects. The illumination condition of the evening time is more rapidly changed. These two factors can affect the quality of the moving object detection. We deal with these artifacts by fusing the several illumination or shadow invariant features. The preliminary results of this study are validated qualitatively on two real datasets. Also, the proposed algorithm achieves the real-time performance.
Keywords :
image motion analysis; image sensors; lighting; object detection; cast shadow; declining sun; elongated shadow; illumination changes; illumination condition; real-time moving object detection; shadow effects; shadow invariant features; static camera; Feature extraction; Lighting; Object detection; Probabilistic logic; Real-time systems; Robustness; Sun; background subtraction; moving object detection; shadow; surveillance;
Conference_Titel :
Ubiquitous Robots and Ambient Intelligence (URAI), 2014 11th International Conference on
DOI :
10.1109/URAI.2014.7057496