| 摘 要: 图像去雾技术的目的是为了去掉图像中雾的影响，从而获得高质量的图像。本文主要从图像增强、图像复原和深度学习的角度归纳总结了图像去雾方法的研究状况，对暗通道先验等经典算法以及新活跃在去雾领域的几种深度学习去雾算法做了进一步的分析，并对各类算法的性能进行了总结，最后针对各类图像去雾方法指出了存在的问题及未来的展望。 |
| 关键词: 图像去雾；暗通道先验；神经网络；大气散射模型；透射率 |
|
中图分类号: TP391.4
文献标识码: A
|
|
| An Overview of Research on Image Dehazing Algorithms |
|
PU Hengfei, HUANG Zhiyong
|
(College of Big Data and Information Engineering, Guizhou University, Guiyang 550025, China)
18198542837@163.com; 1392599321@qq.com
|
| Abstract: Image dehazing technology aims to remove haze in image, so as to obtain high-quality images. This article summarizes the research status of image dehazing methods from perspectives of image enhancement, image restoration and deep learning. It makes further analysis of classic algorithms such as dark channel priors and several deep learning dehazing algorithms that are newly active in the field of dehazing. It also analyzes performance of each algorithm, and finally points out existing problems and future prospects for various image dehazing methods. |
| Keywords: image dehazing; dark channel prior; neural network; atmospheric scattering model; transmittance |
