基于分数阶微分的图像增强模板构造  

Construction of image enhancement mask based on fractional differential

作　　者：勾荣[1] 

机构地区：[1]江苏开放大学信息工程系,江苏南京210019

出　　处：《计算机工程与设计》2014年第10期3554-3557,3644,共5页Computer Engineering and Design

基　　金：江苏省高校自然科学研究基金项目(09KJD520010);校"十二五"规划课题基金项目(13SEW-Y-013)

摘　　要：为充分利用图像目标像素点周围8个方向及其邻近像素点的信息,构造一种新的分数阶微分图像增强模板,并根据分数阶微分的Riemann-Liouville定义,推导出模板系数。利用该模板及其它分数阶微分图像增强模板,分别对灰度图像和彩色图像进行图像增强处理实验;引入图像熵的概念和计算公式,对图像增强结果进行熵值比较。实验及熵值结果表明,该分数阶微分图像增强模板能有效保留图像的纹理细节和边缘信息,图像增强效果更为明显且熵值最大。To make full use of the image pixels ,which including the eight directions and the neighborhood around the image cen-tral pixel ,a kind of new image enhancement mask based on the fractional differential was constructed ,and the mask coefficients were based on the Riemann-Liouville definition of the fractional differential .Gray images and color images were experimented using the new and other fractional differential image enhancement masks .At last ,the definition of the image entropy and its cal-culation formula were introduced ,and the computing results of the image entropy were compared and analyzed .The results of image enhancement experiments and the image entropy calculation show that the new fractional differential mask can maintain the image texture feature and the image boundary information effectively .The image enhancement effect is better than other masks and the image entropy is the best .

关 键 词：图像增强 分数阶微分 R-L定义 图像增强模板 卷积运算 熵 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]