| 吴一全,孟天亮,吴诗婳.基于NSST和人眼感知保真约束的图像自适应增强[J].光电子激光,2015,26(5):978~985 |
| 基于NSST和人眼感知保真约束的图像自适应增强 |
| Adaptive image enhancement based on NSST and constraint of human eye perception information fidelity |
| 投稿时间：2014-10-02 |
| DOI： |
| 中文关键词: 图像自适应增强 非下采样剪切波变换(NSST) 人眼感知保真约束 非线性增益函数 |
| 英文关键词:adaptive image enhancement non-subsampled shearlet transform (NSST) constrain t of human eye perception information fidelity nonli near gain function |
| 基金项目:国家自然科学基金(60872065)、江苏省社会安全图像与视频理解 重点实验室(南京理工大学)开放基金 (JSKL201302)、国土资源部地质信息技术重点实验室开放基金(217)、农业部淡水渔业与种质资源利用重点实验室开放基金(KF201313)、农业部东海海水健康养殖重 点实验室基金(2013ESHML06)和江苏高校优势学科建设工程资助项目 (1.南京航空航天大学 电子信息工程学院,江苏 南京 210016； 2.南京理工大学 江苏省社会安全图像与视频理解重点实验室,江苏 南京 210094； 3.国土资源部地质信息技术重点实验室,北京 100037； 4.农业部淡水渔业和种质资源利用重点实验室,中国水产科学研究院淡水渔业研究中心,江苏 无锡 214081； 5.农业部东海海水健康养殖重点实验室,福建 厦门 361021) |
|
| 摘要点击次数: 1314 |
| 全文下载次数: 5 |
| 中文摘要: |
| 鉴于现有的图像增强方法在提高图像对比度、清 晰度等方面仍存在不足,提出了基于非下采样剪切波变换 (NSST,non-subsampled shearlet transform)和人眼感知保真约束的自适应 增强方法。首先对输 入图像进行NSST,分解为一个低频子带图像和多个高频子带图像；然后利用非线性增益 函数增强高 频子带系数,同时对低频部分进行分块局部增强；考虑到传统分块局部增强存在局部图像块 间不连续进而 导致失真的情况,引入了人眼感知保真约束条件,并将其转化为求解一个典型的线性优化问 题,由此获取 增强参数,实现低频部分的增强；最后融合处理后的高低频子带系数,重构出期望的增强 图像。大量实验 结果表明,与近年提出的4种同类方法相比,本文方法所得增强图像的主观视觉效果更好 ,在清晰度、 局部对比度以及全局对比度等定量评价指标上平均高出50%,且实时性良好。 |
| 英文摘要: |
| In view of low gain in contrast and definition of image obtained by th e existing image enhancement methods,an adaptive enhancement method is proposed based on non-subsampled she arlet transform(NSST) and constraint of human eye perception information fidelity.Firstly,the input imag e is decomposed by NSST into a low-frequency sub-band and several high-frequency sub-bands.Then the high- frequency sub-band coefficients are enhanced by a nonlinear transform gain function,while the low-frequency sub-b and is processed through local blocking enhancement method.Because the traditional local blocking enhancement method has the problem of discontinuity between image blocks,which results in distortion,the constraint o f human eye perception information fidelity is introduced.And then a classical linear optimization model is establ ished according to the constraint to obtain the optimal enhancement parameters,thus the enhancement of low-frequenc y sub-band is accomplished. Finally,the processed low-frequency sub-band and high-frequency sub-bands a re fused and the desired enhanced image is reconstructed.Large numbers of experimental results demonstrate that c ompared with 4congener enhancement methods put forward in recent years,the proposed method can achieve images with superior subjective visual effects and has a 50% average increment in quantitative evalua tion indicators,such as definition, local contrast and global contrast.Moreover,its real-time performance is also satisfactory. |
| 查看全文 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|