基于Guide滤波和NSCT—PCNN的图像增强
2016-05-03张雁翔祁育仙
张雁翔 祁育仙



摘要:针对图像不均匀的特点,提出了基于Guide滤波和非下采样Contourlet变换(NSCT)和脉冲耦合神经网络(PNCC)的图像增强方法。首先,采用直方图均衡化和中值滤波分别对图像进行了预处理;其次,采用NSCT-PCNN对预处理后的图像分别进行分解,提取出高频子带系数和低频子带系数;然后,Guide滤波对高频子带系数和低频子带系数进行处理;最后通过对所有子带系数进行NSCT逆变换,得到增强后的图像。实验结果表明该算法优于其他方法,有更好的图像增强效果和视觉效果。
关键词:图像增强;Guide滤波;NSCT-PCNN
中图分类号:TN911.73 文献标识码:A文章编号:2095-2163(2016)02-
Image enhancement based on Guide Filter and NSCT-PCNN
ZHANG Yanxiang, QI Yuxian
(Information Management and Construcion center,Taiyuan University of Technology,Taiyuan 030024,China)
Abstract: Image contains very significant informationwhich is one of the vital media. Image enhancement is a key technique in image processing. It is widely used in many aspects, also a very necessary technique in image processing. As the character of image is not uniform, a new algorithm based on Guide filtering and NSCT-PCNN is proposed to enhance image. First of all, histogram equalization and median filter are used respectively to preprocess the image. Secondly, image is decomposed by NSCT-PCNN to extract the high frequency subband coefficients and low frequency subband coefficients. Then, the coefficients are dealt with Guide filter. Finally the enhanced image is obtained by taking inverse NSCT transform. The experiments show that the algorithm is superior to other methods, which not only can obtain enhanced result , but also get better visual effect.
Key words:Image enhancement; Guide filter; NSCT-PCNN
0引言
在信息高速发展的时代,语言和图像已然成为信息传递的重要媒介。人们获取信息的75%来自于视觉系统,而图像是视觉系统获取信息的主要来源。图像增强[1],即提高图像目标识别能力,其目的在于改善图像的视觉效果,突显图像中的有用信息,并将图像转换成更适合计算机处理的形式。图像增强并不考虑图像质量变差的原因,而只是关注图像中感兴趣的识别[2]。目前,彩色图像处理已经渗透拓展至医学研究、印刷行业、智能交通和卫星图像等多个领域[3]。
图像增强方法可分为2类:频域法和空间域法[4~6]。前者是把图像看成一种二维信号,对其进行基于二维傅里叶变换的信号增强[6~7]。基于频域的算法是在图像的某种变换域内对图像的变换系数值加进某种修正,是一种间接的算法[5,8],基于空域的算法处理是直接对图像灰度级开展运算。……
