NSCT变换域PCNN彩色图像增强算法  

An Enhanced Method of Color Image Combined PCNN Based on NSCT

作　　者：吴粉侠[1] 张小波[1] 

机构地区：[1]咸阳师范学院图形图像研究所,陕西咸阳712000

出　　处：《航空计算技术》2016年第5期21-25,30,共6页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(61401383);陕西省教育厅科学研究计划项目资助(16JK1823);咸阳师范学院科研基金项目资助(13XSYK058)

摘　　要：为了改善由于光照不均对真彩色图像的影响,根据人类对颜色的感知特性,利用YCb Cr彩色模型空间,提出了一种基于非抽样Contourlet变换(NSCT)的PCNN模型的图像增强算法。首先,将图像从RGB彩色空间转换到YCb Cr彩色空间;然后对亮度分量进行NSCT分解,得到低频子带系数和高频方向子带系数,对低频子带,利用PCNN增强,对高频子带采用非线性变换进行增强;最后,使用NSCT逆变换重构图像的亮度分量,并将图像从YCb Cr色彩空间模型还原到RGB空间得到增强后的图像。实验结果表明,算法增强效果明显优于同态滤波、空域PCNN及NSCT域的PCNN算法,不仅增加了彩色图像的明亮度,而且图像保真性好,纹理更清晰。A new enhanced method of non- uniform illuminaton color image based on NSCT was pro- posed. First, a color image is transformed from RGB color space to YChCr color space. Second, luminance channel is decomposed by NSCT, and We can obtain the low- frequency sub- image and a series high- frequency sub- images. Third the PCNN was applied to the low- frequency subimage, and nolined translation was applied to high- frequency sub- images. Forth, luminance component image is obtained by inverse NSCT. Last, the color image is transformed from YCbCr color space to RGB color space. Experiments il- lustrate that this algorithm can not only corrected non-uniform illumination in images, but also maintained a good image color and the local details.

关 键 词：非抽样CONTOURLET变换 脉冲耦合神经网络 YCbCr颜色模型 彩色图像增强 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]