摘要
针对各向异性扩散算法容易模糊图像细节和边缘以及去噪不彻底的问题,通过局部方差信息调整参数将LCC扩散函数与ECU扩散函数相结合,综合利用图像局部方差描述的局部区域信息和梯度信息,提出了一种结合局部方差信息的各向异性扩散模型。该模型不仅同时依赖局部方差信息和梯度信息,而且能够在不同性质区域根据局部方差信息调整参数适时地调整LCC扩散函数与ECU扩散函数扩散速度,充分利用LCC扩散函数和ECU扩散函数的优势。实验结果表明,该模型与其它几种经典算法相比,不仅能够有效地去除噪声、保持图像弱边缘,而且对图像的细节保持也有较好的效果。
In view of the problem that anisotropic diffusion algorithm easily blurred image details and edges and denoising of incomplete,combining the LCC diffusion function with ECU diffusion function by using local variance information adjustment parameters and comprehensively utilization of local information described by local variance and gradient information,this paper proposes an anisotropic diffusion algorithm combined with local variance information. The model not only depends on the local variance and gradient information at the same time,but also in different area can timely adjusts diffusion velocity of LCC diffusion function and ECU diffusion function according to the local variance information adjustment parameters,making full use of the advantage of LCC diffusion function and ECU diffusion function. The experimental results show that compared with several other classical algorithms,the model not only can effectively remove the noise and preserving the image edge,but also to keep the details of image also has a better effect.
作者
吴龙华
洪志强
闫晓天
WU Longhua;HONG Zhiqiang;YAN Xiaotian(Think Tank Surveying and Planning Company of Jiangxi,330013,Nanchang,PRC;Nanchang Institute of Science ＆ Technology,330108,Nanchang,PRC)
出处
《江西科学》
2018年第3期500-505,共6页
Jiangxi Science
基金
江西省自然科学基金项目(20161BAB203102)
江西省教育厅科学技术研究重点项目(GJJ150555)
关键词
各向异性扩散
局部方差信息
图像去噪
调整参数
扩散系数
anisotropic spread
local variance information
images denoising
adjustment parameter
diffusion coefficient
