提出了一种新的前视红外序列图像目标跟踪方法.利用基于区域分割和模板匹配的全局补偿算法进行运动估计,有效消除了运动平台中摄像机运动对跟踪性能带来的不利影响.同时,采用二元联合变换相关器进行红外目标的检测与跟踪.通过将振幅调制的功率谱转化为脉冲调制的功率谱,可以获得更清晰尖锐的相关峰,提高了算法的鲁棒性.实验结果表明,该方法是有效的和稳健的.
A novel target tracking algorithm for tracking targets in forward- looking infrared (FLIR) image sequences is proposed. Motion estimation in image sequences is realized to overcome the detrimental effects of the camera motion in airborne by global motion compensation via region segmentation and template matching. Then, a robust target tracking system has been developed using binary joint transform correlation. The sharper correlation peak can be achieved by converting the amplitude modulated power spectrum into pulse modulated power spectrum. Experimental results using infrared image sequences are shown to validate the robusmess of the propose technique.