摘要
将分数阶微积分理论和偏微分方程有效结合,提出了一种基于分数阶变分逆扩散的图像增强模型。该模型在传统的分数阶微分图像增强模型的基础上,引入了演化时间因素,从而使得图像增强模型能够由单次增强处理拓展为多次迭代增强处理,并通过边缘增强算子实现了在逆扩散的过程中根据图像的局部信息量来自适应地控制图像增强的程度,由此获得了更好的图像增强效果。实验仿真结果表明,基于分数阶变分逆扩散方程的图像增强模型较目前经典的图像增强模型,能够使处理后的图像具有相对更高的信息量和更大对比度,边缘更加突出、纹理更加清晰并保留图像平滑区域信息。
The image enhancement model based on fractional -order Variation inverse diffusion was constructed by effective combination of theory of fractional calculus and methods of partial differential e -quations. On the base of traditional fractional differential image enhancement model,the image enhancement model proposed in this paper could extend the single enhancement processing to iterative enhancement pro-cessing by introduction of evolution time, and the edge enhancement operator realized adaptive control the degree of image enhancement in the process of inverse diffusion, so it could get the better effect of image enhancement. The experimental results show that： compared with the traditional image enhancement model, the image enhancement model based on fractional-order variation inverse diffusion equation can strengthen the contrast of image, make the edge and texture significantly clearer, and preserve more smoothing regional information.
作者
黄果
陈庆利
门涛
秦洪英
许黎
HUANG Guo;CHEN QingliD;MEN Tao;QIN Hongying;XU Li(University Key Laboratory of Internet Natural Language Intelligent Processing of Sichuan Province;School of Physics and Electronics,Leshan Normal University,Leshan Sichuan 614000, China)
出处
《乐山师范学院学报》
2018年第4期5-14,37,共11页
Journal of Leshan Normal University
基金
国家自然科学基金"分数阶偏微分方程在图像去噪中的应用研究"(61201438)
四川省科技厅应用基础项目"基于人眼视觉系统的分数阶微分图像增强关键技术研究"(2016JY0238)
四川省教育厅重点项目"分数阶微积分理论在图像去噪中的关键技术研究"(18ZA0363)
四川省教育厅一般项目"基于非相干光反馈半导体激光器的混沌保密通讯关键技术研究"(18ZB0268)
关键词
分数阶变分
逆扩散
图像增强
图像熵
Fractional-order Variation
Inverse Diffusion
Image Enhancement
Image Entropy
