基于众值理论的微光图像帧积分算法  

A Multi-Frame Integration Technique Based on Mode Theory for Low Light Level Image

作　　者：陈立[1,2] 李野[1] 秦旭磊[1] 田景全[1] 

机构地区：[1]长春理工大学理学院,吉林长春130022 [2]长春大学计算机科学技术学院,吉林长春130022

出　　处：《电子学报》2011年第10期2217-2220,共4页Acta Electronica Sinica

摘　　要：帧积分算法是图像增强技术中常用的一种方法,它在处理微光图像时有一定的局限性.本文对此做了进一步的实验与讨论,提出了一种改进的算法———众值帧积分.众值帧积分是对邻帧内对应像素点的灰度值作数学统计,并取出现几率最大的像素灰度值作为滤波结果,因此处理过的图像更接近真实.实验结果表明,众值帧积分能有效地抑制微光图像中的亮点噪声和暗点噪声,对图像边缘的保护特性比帧积分好,具有重要实用价值.Multi-frame integration is commonly used in image enhancement technology as a method,which has some limitations when dealing with low light level image.Further discusses and experiments are done,and an approach based on the mode theory is carried out.It is called mode multi-frame integration.Mode multi-frame integration is making the grey scale value statistics of all the pixels inside the neighborhood frames and taking the grey scale mode as the filtering result,so the processed image is closer to the truth.The experimental results show that it can suppress the light and the dark noises of low light level image effectively,and as for the protection of the image edge,the property of it is better than that of multi-frame integration.It has important practical value.

关 键 词：噪声 帧积分 众值帧积分 微光图像 峰值信噪比 

分 类 号：TL82[核科学技术—核技术及应用]