Automatic tracking of an approaching target in IR image

Abstract

In this thesis, a new target tracking algorithm is proposed for real-time target tracking system. Target tracking process has much computation complexity. Since fast computation is a critical point for real-time system, we concentrate on the ability to track the target on complex background with simple calculation For real-time target tracking, the proposed method uses template matching algorithm with CAPS (Correlation-based Adaptive Predictive Search). We modified the CAPS algorithm for faster computation and more accuracy. Furthermore, we verified the proposed algorithm implemented on DSP (TMS320C6416 DSK) Experiments were performed with real tank images and simulated images. The experimental results show that the proposed method has enough accuracy and simple calculation for real-time target tracking system.