摘要
针对直方图均衡图像去雾方法存在噪声被放大和图像失真较严重的问题,提出了一种基于直方图均衡化的自适应图像去雾算法.该算法首先将图像变换到HIS颜色空间,将图像分为k×k个子图像,然后对每个子图像在确保色调H分量不变的基础上,对饱和度S分量和亮度分量I分别做自适应直方图均衡处理.处理后的图像不但从视觉上看出对比度有了很大的改善,边缘细节保持较好,而且几种图像质量评估方法的计算结果也表明,改进后的算法比用直方图均衡化去雾后的图像质量从对比度、均方误差、信息熵、峰值信噪比以及协方差都有明显的提高.实验用了250组图像,得出的结果是PSNR提高18%以上,均方误差降低15%以上.
Aiming at the problems of noise amplification and image distortion seriously for image defogging using histogram equalization,an adaptive image defogging algorithm based on histogram equalization is proposed. The image is transformed into HIS color space and divided into k×k sub images firstly,then for each sub image,on the basis of keeping the hue H component unchanged,the saturation S component and brightness component I are processed by adaptive histogram equalization respectively. The images are processed with the improved algorithm,the contrast is improved greatly,and the edge details are maintained well. The calculation results of several image quality assessment methods also show that the image quality of the improved algorithm is obviously improved than the histogram equalization method. The 250 groups of images are used in the experiment,and the results show that PSNR is increased by more than 18%,while the MSE is reduced by more than 15%.
作者
李竹林
车雯雯
钱梦杭
徐雪丽
LI Zhulin;CHE Wenwen;QIAN Menghang;XU Xueli(Institute of Mathematics and Computer Science,Yan’an University,Yan’an 716000,Shaanxi China)
出处
《河南科学》
2021年第1期1-6,共6页
Henan Science
基金
国家自然科学基金(61761042)
延安大学科研项目(YDQ2019-10)
省级大创项目(S202010719085)。
