第 42 卷 第 5 期
2015 年 5 月
Vol. 42, No. 5
June, 2015
中 国 激 光
CHINESE JOURNAL OF LASERS
0512001-
基于子带分解多尺度 Retinex 的红外图像
自适应细节增强
李 毅
1,2
张云峰
1
*
李 宁
1
方艳超
1
吕春雷
1
于国权
1
陈 娟
1
1
中国科学院长春光学精密机械与物理研究所, 吉林 长春 130033
2
中国科学院大学, 北京 100049
摘要 为实现高动态范围红外图像压缩和高亮区与阴影区细节增强,提出一种基于子带分解多尺度 Retinex 自适应
细节增强方法 。 利 用子带分解 多 尺 度 Retinex 获 取 三 个 独立光谱子 带 ;利用引导滤波 将 各 子带分为细 节 层 和 基础
层;之后依据子带特性设计细节增强权值基函数,自适应实现红外图像细节增强;针对输出图像平滑区灰度不均匀
特点,自适应求取 Gamma 曲 线实现灰 度映射。实验结果表 明:经本文算 法处理后 图像阴影区与高亮区 细节得到 明
显增强,全局视觉效果良好。客观测评结果表明:本文算法有效增强图像细节信息,并且与经典基于双边滤波的细
节增强算法比较,本文算法耗时没有增加。
关键词 图像处理; 红外图像; 细节增强; 子带分解; 引导滤波
中图分类号 TP391 文献标识码 A
doi: 10.3788/CJL201542.0512001
Adaptive Detail Enhancement for Infrared Image Based on Subband-
Decomposed Multi-Scale Retinex
Li Yi
1,2
Zhang Yunfeng
1
Li Ning
1
Fang Yanchao
1
Lü Chunlei
1
Yu Guoquan
1
Chen Juan
1
1
Changchun Institute of Optics, Fine Mechanics and Physics, Chinese Academy of Sciences,
Changchun, Jilin 130033, China
2
University of Chinese Academy of Sciences, Beijing 100049, China
Abstract An adaptive detail enhancement method based on subband-decomposed multi-scale Retinex is proposed
to deal with high dynamic range compression of infrared images and detail enhancement in both high light regions
and dim regions. Three independent spectrum subbands using subband- decomposed multi- scale Retinex are
gained. Then guided image filter is applied to get detail layer and base layer from each subband. Later the basis
weight function for detail enhancement is proposed according to characteristic of separate spectrum subband.
Adaptive detail enhancement is achieved with basis weight function. In order to eliminate the nonuniformity of gray
intensity in the outcome image, a new adaptive way to get Gamma curve for gray value remapping is put forward.
Experimental results show that the detail of the enhanced images is upgraded greatly in both high light regions and
dim regions, and have a satisfied visual effect. Objective evaluation parameters illustrate that the proposed algorithm
can effectively enhance detail of infrared images. In addition, the time consuming is not lengthened compared to
other algorithms in the experiment.
Key words image processing; infrared images; detail enhancement; sub-band decomposition; guided image filter
OCIS codes 100.2000; 110.3080; 100.2980; 110.1085
收稿日期: 2014-11-14; 收到修改稿日期: 2014-12-28
基金项目: 国家自然科学基金(61205143)、吉林省科技厅重点项目(20110329)
作者简介: 李 毅(1988—),男,博士研究生,主要从事实时图像处理方面的研究。E-mail: leey2009@qq.com
导师简介: 陈 娟(1962—),女,研究员,博士生导师,主要从事光电测量、伺服控制等方面的研究。
E-mail: chenjuan@mail.ccut.edu.cn
*通信联系人。E-mail: zyfciomp@sohu.com
1