Abstract: Image quality detection technology can automatically detection the image abnormality in order to replace manual inspection methods for the monitoring system. It can accurately analyze abnormalities of the video, and alarm the system in order to ensure normal running of the expanding network video surveillance system. Noise detection technology based on the spatial domain use the image information of the field characteristics, profile and orientation distributions of various kinds of abnormal noise in spatial domain, and take advantage of OpenCV-based image processing technology achieving detection of noise points, snowflakes and stripes. The noise detection algorithm in spatial domain proposed, is consistent with human visual perception and can be used to monitoring video for real-time detection.

Key words: video quality detection, image noise, spatial domain, noise points, snowflakes, stripes

摘要： 图像质量检测技术可以代替人工巡检的方式对视频质量进行自动检测，对监控系统中视频图像出现的异常进行准确分析、判断和报警，以保障规模不断扩大的网络视频监控系统的正常运行。基于空间域的图像噪声检测技术，利用图像的邻域信息特征与各类噪声异常在空间域上的轮廓和方向分布，并结合OpenCV图像处理技术，实现对噪点、雪花和条纹异常的检测。空间域噪声检测算法，与人的视觉感知相一致，可以用于监控视频的实时检测。

关键词: 视频质量检测, 图像噪声, 空间域, 噪点, 雪花, 条纹