Detecting Shadow of Moving Object based on Phong Illumination Model
Feipeng Li, Zongxi Song, Bin Li, Meng Jie Wu, Chao Shen · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2015
When detecting moving objects using visible light image sequence, the shadows of moving objects always appear together with moving objects.Shadows usually change with light condition, which brings difficulty for recognizing objects after segmenting image.This text proposes an algorithm to detect shadows in foreground image according to the Phong illumination model.The algorithm combines the brightness information between foreground image and background image to calculate an array of brightness relative variation ratio, detecting shadows using the statistical feature of shadows in the array.The algorithm uses the statistical histogram to describe statistical feature of brightness relative variation ratio, and the Gaussian model to do curve fitting as a description for the shape of statistical histogram.And finally with the property of the fitted function could we find the threshold value of shadows adaptively.