基于单元信息熵矢量特征的图像匹配跟踪算法  

Image Matching Tracking Algorithm Based on the Unit-entropy Vector Feature

作　　者：江和平[1] 沈振康[1] 

机构地区：[1]国防科技大学ATR实验室,湖南长沙410073

出　　处：《信号处理》2006年第5期678-682,共5页Journal of Signal Processing

摘　　要：由于图像信息熵只与图像中的图像像素点值的数目有关,而与像素点的位置信息无关,即图像信息熵不能很好地反映图像间的形状差异。本文提出了基于单元信息熵矢量特征的图像匹配跟踪算法,解决了熵相同而形状不同的问题,利用单元信息熵矢量间的距离相关性来完成图像的匹配跟踪。因此该算法不仅具有抗噪能力,而且具有良好的抗辐射失真和抗几何失真的能力。仿真试验表明：在辐射失真情况下,该算法具有稳健的匹配跟踪能力。Because the image information entropy is only dependent on the numbers of pixels without the location information of the pixels in the image, the image information entropy cannot demonstrate the shape diversity between the images. In the paper, the image matching tracking algorithm based on the unit-entropy vector feature was proposed. The algorithm solved the problem that the image entropy was same ,but the shape was differ, and accomplished the image＇s matching tracking with the correlation between the unit-entropy vectors. Thus, the image matching tracking algorithm based on unit-entropy vector was possessed of the very good anti-noise properties, anti-distortion properties of the radiation and anti-distortion properties of the geometry. The simulation experimental result shows the image matching tracking algorithm based on unit-entropy vector has the robustness matching tracking properties.

关 键 词：图像匹配 图像信息熵 单元信息熵矢量 图像跟踪 

分 类 号：TN953[电子电信—信号与信息处理]