Real time, dynamic target tracking using image motion

This paper describes the application of real time vision system for dynamic tracking of moving objects, using image motion. The study encompasses experimental development of an intelligent system that works in ordinary light and is capable of tracking moving objects with unpredictable motion. In this research the task of mobile target tracking using a pan-and-tilt camera is considered with motion-based dynamic target tracking to deal with complex targets (shape and motion pattern unknown). We propose a method based on retrieving the target position in the image from its estimated motion. Our proposed method is based on obtaining a desired 2-D motion field in the image sequence. In this study we link variations of visual features linearly to the camera velocity. Stereovision technique is used to calculate the maximum angular displacement to retrieve the object position in image and to get the direction of motion of the object. The results are obtained in a real time test environment and are discussed. We developed the program in C++ language for linking all the algorithms in one chain. The complete target tracking algorithm works in a real time mode and can be utilized for a number of applications.