An imaging target tracking software for a precision guided missile application

Detection and tracking are the critical problems of computer vision. It is essential pre-processing for other algorithms. In this paper, we present an efficient and robust detection and tracking algorithm which is based on morphological operation and dynamic filtering (HPDAF: Highest Probability Data Association Filter). In airborne IR imagery, objects are usually obscure and too small to recognize their shape. Therefore dynamic filtering approach can be the best solution for tracking. Several experimental results show that our proposed algorithm efficiently detects and tracks the object regardless of size and their shape consistency. Temporarily occluded targets can be handled track-termination and re-detect sequence of the proposed algorithm.