基于图像融合的双波段目标跟踪算法研究

针对红外和可见光图像各自的特点以及单一传感器在目标跟踪中的缺陷，提出了基于双波段融合图像的目标跟踪算法。该方法对原始图像进行小波分解后，为了满足目标跟踪的稳定性及实时性，重点考虑目标跟踪时需要的边缘等细节信息，采用对低频系数取零，高频系数基于小波系数绝对值取大的融合方法，然后对融合后的图像采用基于Mean shift算法进行目标跟踪。实验结果表明，此算法可以稳定并且实时跟踪目标，通过对单波段采用相同的跟踪算法进行比较，算法在性能上优于单波段的目标跟踪。

Research on dual-band target tracking algorithm based on image fusion

Aim at the limitation of target tracking with single sensor,a Dual-Band target tracking method based on image fusion is presented.The wavelet transform is adopted to decompose the input images.For the purpose of fusion which is to satisfy the stabilization and Real-time of target tracking,it mainly considers the detail information of target,such as the edge information.So the low frequency coefficients are set to be zero,the high frequency coefficients chooses larger absolute value of Wavelet coefficient.Then the Mean shift algorithm is used for fused image to track target.A group of experimental results demonstrate that the presented algorithm can track target steadily without delay.Compared with single sensor image,the proposed algorithm demonstrates better performance.