引用本文: | 遆晓光,曲悠杨.一种改进的色彩保持低照度图像增强方法[J].哈尔滨工业大学学报,2014,46(3):1.DOI:10.11918/j.issn.0367-6234.2014.03.001 |
| DI Xiaoguang,QU Youyang.An improved low illumination image enhancement algorithm with color preserving[J].Journal of Harbin Institute of Technology,2014,46(3):1.DOI:10.11918/j.issn.0367-6234.2014.03.001 |
|
摘要: |
针对现有低照度图像增强方法恢复结果存在的色彩失真问题,采用原用于图像去雾领域的暗通道先验方法,提出了一种基于暗通道先验的低照度图像增强方法.通过提取图像的暗通道信息来获得对应场景的光照强度分布图;针对暗通道先验原理导致的白色像素区域暗通道数值计算错误的问题,提出了一种通过计算像素的RGB三通道数值差值来判定错误区域方法,并给出了具体的错误修正方法;为了消除暗通道先验产生的块效应,对修改后的光照强度分布图进行平滑;基于图像每个像素点光照强度与标准光照强度的比例关系对原图像的RGB三通道进行放大,实现色彩保持的图像增强;与现有方法进行实验对比分析,从恢复结果的色彩失真情况以及峰值信噪比等方面验证了所提方法的有效性. |
关键词: 低照度图像增强 暗通道先验 色彩保持 图像平滑 |
DOI:10.11918/j.issn.0367-6234.2014.03.001 |
分类号: |
基金项目:国家自然科学基金资助项目(61174202);中央高校基本科研业务费专项资金资助项目(HIT. NSRIF 201163);黑龙江省博士后科研启动金资助项目(LBH-Q10078). |
|
An improved low illumination image enhancement algorithm with color preserving |
DI Xiaoguang, QU Youyang
|
(Control and Simulation Center,Harbin Institute of Technology, 150080 Harbin, China)
|
Abstract: |
To solve the problem of color distortion existing in low illumination image enhancement method, a novel low illumination image enhancement method based on dark-channel prior with the capability of preserving scene color is presented. First, the intensity of every image pixel with dark-channel prior is estimated directly, and for dealing with wrong dark-channel values in white region, the regions with wrong grey values by calculating the minimum difference value of the RGB channels and correct the dark-channel values of the regions are founded. Then, a linear smoothing method is used to refine the block effect caused by dark-channel prior, and the enhanced image is achieved by scaling up the pixel intensity according to its value in dark-channel image. Finally, by comparing with the existing low illumination image enhancement method, the effectiveness of our method with better color preserving capability and higher PSNR is proved. |
Key words: Low illumination image enhancement Dark-channel Prior Color preserving Image smoothing |