Alleviating Cavity Problems in Moving Object Detection based on Background Modeling

• Main Authors: Wei-Wen Chang, 張偉文
• Other Authors: Chih-Yang Lin
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/66811274325803982450

碩士 === 亞洲大學 === 資訊工程學系碩士班 === 101 === Traditional background modeling not only requires a lot of computing, it also faces challenges such as noise and illumination changes, which prevent complete modeling of moving objects. In this paper, we propose a background subtraction method that strengthens the foreground object and adapts to environmental changes. This method consists of three parts. The first is to build a color-based model and a texture-based model through color and texture information respectively. In the second part, we introduce the concept of suppression & relaxation to repair broken foreground regions. Finally, we apply motion compensation, which uses the previous output, the history, to repair current output. Our proposed scheme is the first to apply the concepts of suppression & relaxation and motion compensation to background modeling. Our experimental results show that the proposed scheme is significantly better than other methods in terms of computation complexity, precision, recall, similarity and F-measure.