数字图像的局部分数阶微分增强  

Local Fractional Differential Algorithm for Image Enhancement

作　　者：陈庆利[1] 黄果[1] 门涛[1] 张秀琼[1] 秦洪英[1] 王明蓉[1] 

机构地区：[1]乐山师范学院互联网自然语言智能处理四川省高等学校重点实验室,四川乐山614004

出　　处：《四川大学学报（工程科学版）》2016年第4期115-122,共8页Journal of Sichuan University (Engineering Science Edition)

基　　金：国家自然科学基金资助项目(61201438);四川省科技厅资助项目(2014JY0036;2016JY0238);四川省教育厅资助项目(13ZB0104;13TD0014);乐山师范学院资助项目(Z1272)

摘　　要：为了扩展传统分数阶微分在图像增强处理中微分阶次的范围,改善传统分数阶微分对图像亮度增强不甚理想的问题,提出一种局部分数阶微分增强图像算法。根据局部分数阶微分理论,建立了数字图像的分数阶微分增强分算法。然后构造了新的数字图像分数阶微分增强模板,在该模板中增加亮度控制函数得到边缘、纹理和对比度同时增强的分数阶微分增强算法。实验表明,该方法能扩大分数阶微分在图像增强处理中阶次的范围;不但能很好地增强图像的边缘、纹理和轮廓等信息,又能明显改善图像的对比度和亮度,增强图像的视觉效果优于传统的分数阶微分增强方法的视觉效果。In order to extend range of fractional differential order and to improve brightness when fractional differential methods are applied to image enhancement, a local fractional differential algorithm for image enhancement was proposed. According to theory of local fractional differential, a local fractional differential algorithm for image enhancement was deduced. Then, on basis of a new constructed fractional differential template for image enhancement, a new template for texture, edge and brightness enhancement was gotten when a set of brightness control functions were added. Experiments and results showed that the proposed method can extend the range of frac- tional differential order. It can not only enhance edges, textures and other details of the image, but also significantly improve the contrast and lightness of the image. Experiments also showed that the visual effect of enhanced images is better than that of traditional fractional differential-based methods.

关 键 词：图像增强 分数阶微分 局部分数阶微分 分数阶微分模板 对比度增强 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]