基于图像多尺度熵的红外图像匹配跟踪算法

在对图像熵进行分析的基础上,引入图像多尺度熵的概念,定义了图像的多尺度熵及多尺度熵矢量,提出了一种基于区域的匹配跟踪算法—–基于图像多尺度熵的红外图像匹配跟踪算法.首先计算图像的多尺度熵,得到图像多尺度熵矢量;然后利用多尺度熵矢量间的绝对距离(AD)进行匹配跟踪.实验表明,该算法不仅具有稳定、精确的匹配跟踪性能,而且能在目标发生旋转时,较好地匹配跟踪目标,并具有良好的抗几何失真能力.

Infrared image matching tracking algorithm based on image multi-scaleentropy

After analyzing the entropy of the image, the conception of image multi-scale entropy is introduced, and image
multi-scale entropy and the vector of image multi-scale entropy are defined. A region matching tracking method-infrared
image matching tracking algorithm based on image multi-scale entropy is present. Firstly, the image multi-scale entropy and
the vector of the image multi-scale entropy are calculated. Then the absolute distance(AD) of the vector of image multi-scale
entropy is applied to matching and tracking. The experimental results show that the proposed algorithm has the robustness
matching tracking properties and is robust to the problems of the rotation and geometry distortion of targets.