A robust self-organizing algorithm to moving object detection

Liu Yong,Ke Zunhai,Yichun Xu,Renbin Xiao
DOI: https://doi.org/10.13245/j.hust.2013.07.012
2013-01-01
Abstract:A self-organizing algorithm to background subtraction for foreground extraction was proposed,which was robust in dynamic scenes,illumination change and ghost effects when visual computers were used.Firstly,the captured frame was converted from RGB(red green blue) color space to HSV(hue saturation valve) color one.In the phase of modeling background,the HSV components were mapped to a 2D flat grid through the method similar to the artificial networks.Then the foreground was extracted by chromaticity distortion and TOM(time out map) mechanism.In addition,the shadow cast and highlight were detected and removed through combining chromaticity distortion and brightness distortion.Finally,the background model was dynamically updated in a way,when the corresponding spatial relationship was considered.The experimental results demonstrate that this proposed approach can be well adapted to the changes of the dynamical scenes and the illumination and eliminate the ghosts.
What problem does this paper attempt to address?