摘要
低照度图像增强作为图像处理领域的一个重要分支,其目的是为了提高输入图像的亮度、对比度,同时抑制图像中的噪声,增强细节区域,以提升图像整体的视觉效果。针对低照度图像增强问题,梳理了代表性方法,从基于域处理、基于Retinex理论、基于图像去雾算法和基于深度学习等4种类型,详细介绍了各类算法的基本原理及步骤流程,并对各算法的实际增强效果进行了评价。据此,分析指出了目前低照度图像增强的瓶颈问题。最后,结合各种算法的特点,对低照度图像增强算法未来的发展趋势进行了总结和展望。
As an important branch of image processing,low-illumination image enhancement aims to improve the brightness and contrast of an input image,suppress the noise in the image and enhance detail areas so as to improve the overall visual effect of the image.Aiming at the problem of low-illumination image enhancement,representative methods were sorted out,the basic principles and step flow of various algorithms were introduced in detail from four types,i.e.domain-based processing,Retinex-based theory,image defogging algorithm and deep learning,and the actual enhancement effect of each algorithm was evaluated,based on which the bottleneck of low-illumination image enhancement was analyzed and pointed out.Finally,combined with the characteristics of various algorithms,the future development trend of low illumination image enhancement algorithms was summarized and prospected.
作者
汤浩
朱泓宇
谢超
TANG Hao;ZHU Hong-yu;XIE Chao(College of Mechanical and Electronic Engineering,Nanjing Forestry University,Nanjing Jiangsu 210037,China)
出处
《林业机械与木工设备》
2022年第7期4-10,共7页
Forestry Machinery & Woodworking Equipment
基金
国家自然科学基金资助项目(61901221)
江苏省研究生科研与实践创新计划资助项目(KYCX21_0872)
国家重点研发计划资助项目(2019YFD1100404)。
