Image-based Automatic Tracking Control System

Auto-tracking system could be widely used in everyday life and engineering practice. This paper aims at combining control theory with digital video processing technology to design a camera direction auto-tracking system based on images. In the camera direction close-loop control part, a pure angle tracking concept has been put forward which set up a movement model on projection surface of image and control object has been included in image processing and analysis part; a two cushion-area model and prediction algorithm has been advanced to solve sampling delay; a step-by-step method is proposed to compute pan-tilt control instructions. In image capturing and processing part, some novel methods like image part-thresholding, hide-binarizing, part-searching have been used to improve image processing and object searching speed. On Windows OS platform, VC++6.0 and LogiTech QCSDK tools have been used to program system software which employed asynchronism multi-thread serial communication technique. In order to determine the algorithm parameters, a experiment equipment has been set up which used CCD camera, pan-tilt, PC. A regular circle moving object has been studied to test the system capacity.