基于小波变换的彩色图像去雾方法  

A Method of Color Image Defogging Based on Wavelet Transform

作　　者：贺欢 吐尔洪江·阿布都克力木[1] 何笑 HE Huan;Turghunjian·ABDUKIRIM;HE Xiao(School of Mathematical Sciences,Xinjiang Normal University,Urumqi 830017,China)

机构地区：[1]新疆师范大学数学科学学院,新疆乌鲁木齐830017

出　　处：《计算机技术与发展》2020年第9期60-64,共5页Computer Technology and Development

基　　金：国家自然科学基金(11261061,61362039,10661010);新疆维吾尔自治区自然科学基金(200721104);新疆师范大学数学教学资源开发重点实验室招标课题(XJNUSY082017B03)。

摘　　要：针对雾天拍摄的图像模糊、对比度低、图像失真严重、获取重要信息困难等一系列问题,提出了一种基于小波变换的彩色图像去雾方法。首先在RGB颜色空间对图像的RGB三个颜色通道进行直方图均衡处理;同时在YCbCr彩色空间提取Y分量,并对其进行二维离散小波变换,得到一个低频分量和三个高频分量,对低频分量进行同态滤波处理,而对三个高频分量进行限制对比度直方图均衡处理,然后进行二维离散小波逆变换重构高低频部分,最后转换回RGB颜色空间,并将两幅图像进行线性组合,得到最终的去雾图像。仿真实验表明,通过结合主观分析和客观评价标准,该方法与其他去雾方法相比,对比度较高,视觉感较好,亮度较高,颜色恢复较逼真,去雾效果较好。Aiming at a series of problems such as blurred image,low contrast,serious image distortion and difficulty in obtaining important information,we propose a color image defogging method based on wavelet transform.Firstly,histogram equalization is performed for the three color channels of the image in RGB color space.Meanwhile,Y component is extracted from YCbCr color space,and two-dimensional discrete wavelet transform is carried out to obtain one low-frequency component and three high-frequency components.Homomorphic filtering is carried out for the low-frequency components,and restricted contrast histogram equalization is carried out for the three high-frequency components.Then,the high and low frequency parts are reconstructed by two-dimensional discrete wavelet inverse transform,and finally converted back to RGB color space,and the two images are combined linearly to obtain the final defogging image.Simulation results show that combined with the subjective analysis and objective evaluation standard,compared with other defogging methods,the proposed method has higher contrast,better visual perception,higher brightness,more realistic color recovery and better defogging effect.

关 键 词：小波变换 同态滤波 MALLAT算法 YCbCr彩色空间 RGB颜色空间 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]