Title :
Background modeling methods based on two-stage strategy
Author :
Xiaohua Shu ; Xiyu Xiao ; Yonghong Long ; Guangsheng Zeng ; Yunfei Zhong
Author_Institution :
Dept. of Electr. Eng., Hunan Univ. of Technol., Zhuzhou, China
Abstract :
To solve the problem of background image during traffic jam, the background modeling method based on two-stage selection strategy is presented. The first selection of image frame is performed according to the degree of traffic congestion which is measured by the variance value of the image frame, then the secondary pixel-wise selection is performed by frame difference method to exclude the motion pixels of the candidate image frame. To estimate the background image value of each pixel, the median method is adopted, therefore, a better background image could be obtained in a short time even in the traffic jam.
Keywords :
feature selection; image motion analysis; intelligent transportation systems; road traffic; statistical analysis; background image value estimation; background modeling method; frame difference method; image frame selection; image frame variance value; median method; motion pixels; secondary pixel-wise selection; traffic congestion degree; traffic jam; two-stage selection strategy; background modeling; frame difference; two-stage selection strategy; variance of frame;
Conference_Titel :
Information Science and Control Engineering 2012 (ICISCE 2012), IET International Conference on
Conference_Location :
Shenzhen
Electronic_ISBN :
978-1-84919-641-3
DOI :
10.1049/cp.2012.2328