摘要
提出了基于运动检测的自适应降噪算法.通过计算相邻图像的相同背景区域噪声方差得到噪声强度,根据噪声强度大小设置运动阈值;通过计算图像局部相似度自适应地区分图像运动区域和静止区域,对静止区域采用时域加权均值滤波,对运动区域采用均值空域滤波.Matlab仿真实验结果表明:采用该降噪算法在去除视频序列噪声的同时,也很好地保护了图像的细节.
An auto-adapter noise reduction algorithm based on motion detection was proposed.The noise intensity was obtained by calculating noise variance of same background regions for adjacent images and motion threshold was set according to noise intensity.The still regions were distinguished from motion regions of video image by local image similarity computation.The temporal weighted filter to the still regions and spatial average filter to the motion were adapted.The experimental result in Matlab simulation environment shows that the noise reduction algorithm can effectively reduce the noise in video signals and protect image details.
出处
《上海工程技术大学学报》
CAS
2015年第2期134-137,共4页
Journal of Shanghai University of Engineering Science
基金
上海市大学生创新训练计划资助项目(cs1421003)
关键词
高斯噪声
噪声方差
局部相似
时域
空域
Gauss noise
noise variance
local similarity
temporal
spatial
