摘要
针对图像噪声过多以及模糊度过高所造成的多光谱图像视觉效果较差、图像细节难以分辨等问题,提出了一种模糊核聚类的线性滤波多光谱图像增强算法。该算法采用模糊核聚类的去噪方法,对分解图像得到的模糊系数进行了阈值处理,并引入去噪增益因子,可以有效地去除多光谱图像的噪声。在多光谱图像亮度增强上,采用了多向聚类亮度增强公式来将图像的模糊像素亮度提升至标准亮度,对图像边缘部分的亮度则采用边缘化增益方法来进行增强,最后采用线性滤波的方法来保护多光谱图像的结构张量,防止多光谱图像的结构信息发生扩散变化。实验结果表明,采用模糊核聚类的方法能够有效地去除多光谱图像噪声,在图像亮度增强上相比对比算法取得了较好的效果。
For issues such as excessive image noise and blur too high to cause less effective visual images and image detail is difficult to distinguish,this paper proposed a linear filtering multispectral image enhancement algorithms with fuzzy kernel clustering. The algorithm used de-noising method based on fuzzy kernel clustering,obtained fuzzy coefficients by decomposition of image thresholding,and introduced the de-noising gain factor,could effectively remove the noise multispectral images.In the multi-spectral image brightness enhancement,it used a multi-directional clustering fuzzy brightness enhancement formula to enhance the brightness of the pixels of the image to the standard brightness. It adopted the marginalized gain method to enhance the brightness of the image on the edge portion. And finally it used linear filtering way to protect the structure tensor of multispectral images,to prevent structural information of multispectral image from diffusing and changing. Experimental results show that the fuzzy kernel clustering can effectively remove the multispectral image noise,brightness enhancement in image contrast has achieved good results compared to the previous algorithms.
出处
《计算机应用研究》
CSCD
北大核心
2015年第5期1536-1539,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(71210003)
关键词
图像增强
模糊核聚类
多向聚类亮度增强
线性滤波
多光谱图像
image enhancement
fuzzy clustering
multi clustering brightness enhancement
linear filtering
multi-spectral image
