摘要
针对图像纹理复杂且难以分析定义的问题,提出一种基于模糊K-S熵的图像纹理分析方法.该方法基于无序理论,延伸了柯尔莫哥洛夫-西奈(K-S)熵的概念.首先,估算所有可能路径的序列模糊隶属度级别;然后,绘出熵差点迹并估计熵差的近似值;最后,使用队列元素等级来测定不精确系统的熵率.实验采用了不同类型的纹理图像,结果表明:在图像模糊集中,或在建立纹理图像像素分布不确定性模型时,估测K-S熵值都是可行的,模糊K-S熵可以作为图像纹理分析的有效特征.
Aiming at the problem that image texture is difficult to be analyzed and defined, a method based on fuzzy K-S entropy is proposed. This method extends the concept of Karl Region-Sinai(K-S) entropy based on chaos theory. Firstly, the sequence fuzzy membership grades of all possible paths are estimated. Then the entropy difference trace is drawn and the approximate value of the entropy difference is estimated. Finally, the rank of the queue element is used to measure the entropy rate of the imprecise system. Different types of texture images are adopted in the experiment. The results show that it is feasible to estimate K-S entropy in image fuzzy concentration, and it is also effective when establishing the uncertain model of pixel distribution of texture images. Therefore, the fuzzy K-S entropy can be used as an effective feature of the image texture analysis.
作者
范伟
江昕
FAN Wei;JIANG Xin(Department of Automation and Information Engineering,Anhui Electrical Engineering Professional Technique College,Hefei,Anhui 230051,China;The Dean’s Office,Anhui Electrical Engineering Professional Technique College,Hefei,Anhui 230051,China)
出处
《湖南城市学院学报（自然科学版）》
CAS
2020年第2期62-67,共6页
Journal of Hunan City University:Natural Science
基金
安徽省高等学校质量工程项目(2019cxtd103)。
关键词
图像纹理
无序理论
K-S熵
模糊集
模糊隶属度
image texture
chaos theory
K-S entropy
fuzzy set
fuzzy membership grade
