An improved moving target detection method and the analysis of influence factors

Abstract

As the core technology of intelligent monitoring system, the moving target detection method plays an important role in the intelligent monitoring system, but the current moving target detection method still has some shortages such as lower anti-jamming performance on the environment, low judge accuracy and so on. In this paper, the improved background subtraction based on mixed Gaussian probability background model is used, local optical flow algorithm is introduced to achieve the unconstrained stable initialization of background model, the added items which are sensitive to global illumination conditions are adopted to assist the adaptive updating of model parameters, eventually the improved algorithm framework are formed and combined with processing method after testing to extract the information of effective movement area and realize the judgment of break in of foreign matters. Finally, the overall realization of intelligent monitoring system is expounded theoretically in the paper, based on the improved moving target detection method, the factors of characteristics of moving targets which affect the detecting precision are analyzed systematically, so it shows that this moving target detection method can improve the detecting precision of intelligent monitoring system and has good real time, accuracy and reliability. © 2012 Springer-Verlag.