在捷联图像导引系统目标跟踪过程中,针对跟踪前期弹目距离较远、跟踪目标在背景图像中所占的区域较小、易遭受背景信息干扰,提出了一种新的基于Mean Shift算法的目标跟踪方法.该算法通过对目标模型做背景直方图加权变换和对目标候选模型做倒数加权变换,显著降低了Mean Shift目标跟踪算法迭代过程中背景信息所占的比重,有效抑制了跟踪区域中背景信息对目标定位的干扰,提高了目标的跟踪精度.实验结果表明,该算法与经典算法相比,迭代次数低、单帧处理时间短,可应用于捷联图像导引系统目标跟踪的前期阶段.%In view of the early stage target tracking guidance system in strapdown image homing system,because of the long distance between missile and target,the target occupies a small area and it is vulnerable to background interference,an new improved target tracking approach based on mean shift algorithm is proposed.By doing target model background histogram weighted while doing target candidate model background histogram countdown weighted,it significantly reduces proportion of background information in the iterations of mean shift target tracking algorithm,and the interference on the target location from track area background information is effectively inhibited,and also the tracking accuracy in the early stage is improved.Experimental results show that compared with the several classic tracking algorithms,the algorithm relatively has a lower average number of iterations,less single frame average processing time and can meet the conditions of application.
展开▼
