SONG Ya-ling, OU Cong-jie. Application of parameter of Tsallis entropy in image threshold segmentation[J]. 2015, 34(11): 147-149. DOI: 10.13873/J.1000-9787(2015)11-0147-03.
Tsallis熵的参数在图像阈值分割中的应用
摘要
Tsallis熵引入非广延参数q
用于描述集合元素间的长程关联
在图像分割中有着广泛应用。从数学原理与实证两方面分析了q参数的作用
并确定了它的合理取值范围。对于目标和背景之间不存在明显关联的图片
用两个具有不同q参数的Tsallis熵分别描述目标集合与背景集合
提出一种新的双q值算法
选择合适的q值确定这两个集合各自的像素灰度关联强度。通过对一系列红外图片的阈值分割
验证了该算法的有效性。
Abstract
Nonextensive parameter q is introduced by Tsallis entropy to illustrate long-range correlation between elements of set
and it is widely used in image segmentation. Function of parameter q is analyzed by mathematical principles as well as practical demonstrations
and reasonable range of q is then determined. For images that have no obvious correlations between target and background
two Tsallis entropies with different q parameters are adopted to describe sets of target and background
seperately
a new double q value algorithm is proposed to determine gray-level correlations intensity of pixels within these two sets is described by tuning the q value. The effectiveness of the proposed algorithm is verified by threshold segmentation of a series of infrared images.