The tracking of multiple moving objects in image sequences is an important task for vehicle guidance, following moving objects, robot control and so on. In order to implement these tracking system, in most systems, a necessary step to be done in tracking of object is to separate the foreground from the background and detect it's motion. But current methods for separating the foreground from background are dependent on illumination(daylight, clouds, shadows) change, processing time and speed, direction or texture of object. In this paper, we deal with tracking of multiple moving objects by Kalman filter for the foreground prediction and the background adaptation. We proposed an algorithm for optimal foreground detection and background adaptation based on Kalman filter. It is composed of three modules such as Block Checking Module (BCM), Object Movement Prediction Module( OMPM) and Adaptive Background Estimation Module(ABEM). BCM detects a new object from image sequences and OMPM predicts the position of moving object by Kalman filter. ABEM takes into account that changing illumination should be considered in the background estimation and should not be detected as foreground. In experiment, we consider that a rigid-body moves continuously in image sequences and a CCD camera with a fixed focal length immovable. The presented approach shows through experimental results that it can well track the multiple objects and adapt the illumination change of background simultaneously.
展开▼
