计算机工程与设计
計算機工程與設計
계산궤공정여설계
COMPUTER ENGINEERING AND DESIGN
2014年
10期
3554-3557,3644
,共5页
图像增强%分数阶微分%R-L定义%图像增强模板%卷积运算%熵
圖像增彊%分數階微分%R-L定義%圖像增彊模闆%捲積運算%熵
도상증강%분수계미분%R-L정의%도상증강모판%권적운산%적
image enhancement%fractional differential%R-L definition%image enhancement mask%convolution operation%entropy
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。
為充分利用圖像目標像素點週圍8箇方嚮及其鄰近像素點的信息，構造一種新的分數階微分圖像增彊模闆，併根據分數階微分的Riemann-Liouville定義，推導齣模闆繫數。利用該模闆及其它分數階微分圖像增彊模闆，分彆對灰度圖像和綵色圖像進行圖像增彊處理實驗；引入圖像熵的概唸和計算公式，對圖像增彊結果進行熵值比較。實驗及熵值結果錶明，該分數階微分圖像增彊模闆能有效保留圖像的紋理細節和邊緣信息，圖像增彊效果更為明顯且熵值最大。
위충분이용도상목표상소점주위8개방향급기린근상소점적신식，구조일충신적분수계미분도상증강모판，병근거분수계미분적Riemann-Liouville정의，추도출모판계수。이용해모판급기타분수계미분도상증강모판，분별대회도도상화채색도상진행도상증강처리실험；인입도상적적개념화계산공식，대도상증강결과진행적치비교。실험급적치결과표명，해분수계미분도상증강모판능유효보류도상적문리세절화변연신식，도상증강효과경위명현차적치최대。
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 .
