基于时空二维目标描述的红外小目标跟踪

根据红外弱小目标的时空域特性,提出了一种基于时空二维直方图均值移动的红外弱小目标跟踪方法。本文构建了空域分量和时域分量来表示红外弱小目标,形成时空二维图像。当前帧图像包含目标的空域分布信息,因此将当前帧图像作为空域分量;差分图像包含目标的运动特性,反映了目标的时域特性,因此将前后两帧差分图像作为时域分量。利用时空二维目标表示方法,采用均值移动算法进行目标跟踪,同时根据Bhattacharyya系数来更新目标模板。采用实际拍摄的小目标视频来检测本算法,仿真结果显示本算法能稳健的跟踪红外弱小目标,均值移动算法的跟踪精度得到了较大的提高。

Infrared small target tracking based on spatial-temporal planar characterization

According to spatial-temporal characteristic of infrared small target,an infrared small target tracking method based on spatial-temporal planar histogram is presented.Spatial component and temporal component are established to describe infrared small target,which forms spatial-temporal planar image.Current frame image contains spatial distribution information of target,so it is as spatial component.Difference image contains moving characteristic of target,it illuminates temporal characteristic of target.Then difference image is as temporal component.At last mean shift algorithm is used to track target,meanwhile target template is updated according to Bhattacharyya coefficient.True images of infrared small target are adopted to check the algorithm.The experimental results show that infrared small target is tracked stably in complicated background by using this algorithm,and tracking accuracy of mean shift algorithm is improved greatly.