A fast background update mechanism for vehicle detection in urban roads
Fei Liu, Zhiyuan Zeng, Zhongyi Li · 2017
This paper proposes a fast and real-time background update mechanism for vehicle detection in urban city. The roads of urban city are relatively clean from the perspective of digital image which is helpful for developing a specialized fast algorithm of background update. However, frequent congestion in urban roads poses a great challenge on constructing a reliable real-time background model. In this paper, a new indicator is proposed to help avoid redundant background update and a novel ghost management technology is developed to help recover the reliable background. The effectiveness of method is validated through experiments and results show that the state-of-art method is highly fast and effective.