摘要
基于非线性扩散在保留图像重要特征方面表现出的良好性能,提出将曲率作为一个控制传导率因素的曲率驱动与边缘停止相结合的非线性扩散模型(C&E模型)应用于图像放大。该模型在实现热扩散时不仅能够连接等照度线和强化边缘,而且强调保护小曲率和大梯度。针对这一模型进一步讨论了其应用于图像放大的初始化条件和数值实现方案。实验结果证明,具有曲率运动、边缘冲击特性和平滑去噪性能的C&E模型较线性扩散方法能更好地放大图像。
Based on the good performances of nonlinear diffusion in preserving important image features,the Curvature-driven with Edge-stopping(CE model) was proposed and applied into image zooming.The CE model has the protection of little curvature and big gradient is emphasized,at same time,it has two effects of connecting broken isophotes and strengthening edge.The condition of the initial and boundary value of image zooming,as well as numerical schemes,were further discussed.For image zooming,the CE model exhibits better performance than the nonlinear diffusion model.
出处
《计算机科学》
CSCD
北大核心
2011年第3期290-291,299,共3页
Computer Science
基金
陕西省自然科学基金(2010JM8026)
陕西省教育厅科学研究计划项目(09JK796)资助
关键词
图像放大
非线性扩散
曲率驱动
边缘增强
Image zooming
Nonlinear diffusion
Curvature-driven
Edge-stopping