Effective algorithms of flare detection with analysis of the shape in real-time video surveillance systems

In this paper, the problem of false operation of a detector of abandoned objects in video surveillance systems during the appearance of sun flares is examined. Two algorithms of flare recognition are presented in this paper. The main feature of the algorithms is the possibility to analyze the shape of detected flares in order to raise the quality of the recognition. The first algorithm is based on application of the active contour technique with analysis of geometric characteristics. This algorithm is compared with the author’s technique, which uses nested windows and analysis of the fractal dimension. After the comparison of the performance and precision of algorithms, a decision about use of one of the algorithms in real-time video surveillance systems will be proposed.