无人值守变电站视频监控系统中摄像头的维护依赖于人工巡查，工作量大，工作效率低。为此，提出一种基于图像处理技术的视频质量检测方法。首先，通过颜色空间转换和偏色因子计算检测视频中偏色故障；其次，通过高斯滤波和峰值信噪比检测视频中雪花噪声；然后，通过帧间方差法检测视频画面冻结故障；最后，通过相似度计算检测视频画面黑屏故障。实验中，视频质量检测准确率均达到98%以上，算法耗时在110 ms 以内。测试结果表明，该方法能实现监控视频质量的实时、准确检测。%Maintenance for camera in video monitoring system in unattended substation mostly depends on human inspection which means heavy work and low working efficiency. Therefore,this paper presents a kind of detection method for video quality based on image processing technology. Firstly,it uses color space conversion and partial color factor to calculate par-tial color fault in detected video. Secondly,it is able to detect snowflake noise in video by Gaussian filter and peak signal to noise ratio. Then,by means of variance between frames method it is able to detect video screen freeze fault and finally to detect black screen fault of video by similarity calculation. In the experiment,accuracy rate of video quality detection rea-ches over 98% and time consumption for the algorithm is within 110 ms.Testing result indicates that this method can realize real-time and accurate detection on monitoring video quality.
展开▼
