Ding Ying^{1, 2}, Li Wen-hui¹, Fan Jing-tao² and Yang Hua-min²

¹ School of Computer Science and Technology, Jilin University,
Changchun 130012, China
liwenhui2050@163.com

² School of Computer Science and Technology, Changchun University of Science and Technology,
Changchun 130022, China
{dingying, fanjingtao, yhm}@cust.edu.cn
 

Abstract. We present a novel method to robustly and efficiently detect moving object, even under the complexity background, such as illumination changes, long shadows etc. This work is distinguished by three key contributions. The first is the integration of the Local Binary Pattern texture measure which extends the moving object detection work for light illumination changing. The second is the introduction of HSI color space measure which removes shadows for the background subtraction. The third contribution is a novel fuzzy way using the Choquet integral which improves detection accuracy. The experiment results using several dataset videos show the robustness and effectiveness of the proposed method.