Active vision for surveillance and occupant detection applications using a high speed CMOS camera system
In this communication we address the problem of detection and tracking of moving objects for surveillance and occupant detection systems. The primary goal in this framework is the motion estimation of the extracted foreground. To overcome the drawbacks characteristically of classical block matching techniques, this contribution presents a feature-based hierarchical locally adaptive multigrid block matching motion estimation technique based on a foreground detection procedure using an adaptive recursive temporal lowpass filter. The usage of a high speed CMOS camera system with pulsed infrared illumination of the observed scene allows a successful execution of the foreground extraction and contributes to an operation under all-weather conditions. Moreover, the scheme leads to a robust and precise motion field estimation, close to the true motion in the scene. The simulation results show the superior performance of the proposed method. It yields better performance than the classical exhaustive search and the modified three-step search technique in terms of the peak signal-to-noise ratio.