基于深度学习的微光条件下图像增强算法  

Image Enhancement Algorithm Based on Deep Learning in Dim Light Condition

作　　者：翁子寒[1] WENG Zihan(Shanghai Children’s Medical Center, Shanghai 200127, China)

机构地区：[1]上海儿童医学中心,上海200127

出　　处：《微型电脑应用》2021年第10期118-121,共4页Microcomputer Applications

摘　　要：由于噪声与光电子数等因素的干扰,在微光条件下的camera成像具有一定难度。因此为了提高微光条件下的图像质量,提出了一种基于深度学习的图像处理方法进行低照度下的camera图像增强。提出的方法可直接对原始的传感器raw图进行处理,以代替传统的多步骤的复杂的图像处理系统。最后,与多种方法进行对比并给出了对比的重建结果以显示方法的有效性。Due to factors such as noise and the photon number of the interference,therefore under low-lux camera imaging has the certain difficulty.In order to improve the camera image quality of dim light conditions,this paper proposes a method of image processing under low-lux based on deep learning.This method can directly deal with the raw sensor data,instead of the traditional multi-step complex image processing system.In the end,we compare with a variety of methods and presente a comparison of reconstruction results to show that the method is effective.

关 键 词：低照度 深度学习 camera图像增强 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]