Moving object detection based on blob analysis
Tao Jia, Nongliang Sun, Maoyong Cao · 2008
A single-mode background model based on blob analysis is proposed to segment foreground from image sequences in complex environment. Firstly, the symmetric difference is used to extract a rough moving object. Secondly, blob analysis is utilized to update background model .Finally, classification strategy (block-level and frame-level) is used to extract foreground accurately and avoid the affect of noise and illumination variance. Experimental results show that the presented approach works well in the presence of complex environment and illumination variance.