灰度成像扩展目标跟踪直方图匹配修正方法

Gray Imaging Extended Target Tracking Histogram Matching Correction Method

针对灰度扩展目标色彩信息不丰富,目标和背景存在重叠这一系列特点,使跟踪点发生漂移的问题,提出了基于均值偏移直方图匹配修正扩展目标跟踪算法。以基于像素位置和像素灰度值的直方图模式作为目标模式;相似度度量采用Bhattacharyya系数定义的距离;用均值偏移迭代进行匹配区域搜索;采用直方图匹配系数进行修正。通过建立卡尔曼预测跟踪策略解决目标被遮挡时的跟踪问题。试验结果表明,采用该算法,跟踪点漂移由原来几十个像素的波动减少到偶有1个像素的抖动。

Due to the extended target gray color information is not rich,a series of features that target and background are overlaps make tracking point happen drift,correction extended target tracking method is put forward based on the mean shift histo-gram. Based on the pixel position and pixel value histogram mode as a target mode. Similarity adopts the distance defined by Bhat-tacharyya coefficient. By mean offset iterative matching area search. Adopt histogram matching factor amended. By establishing kalman prediction tracking strategy solving the target tracking problem when occluded. Experiments results show that,adopting the al-gorithm,tracking drift from originally dozens of pixels volatility reduced to accidentally have 1 pixel dithering.