A series of anomalies that affect video quality may occur in surveillance videos,such as sharpness and brightness anomalies.And there may be the failures of noises,colour cast,picture freezing,water ripples,or horizontal noise waves,etc.,existing in video pictures as well.In view of this,we propose an anomaly detection and quality diagnosis algorithm for surveillance videos.This algorithm analyses the a-nomalies mentioned above,uses grayscale statistical information and gradient information of the image,combines with image’s median filtering and inter-frame difference analysis,and completes the corresponding anomaly detection and quality diagnosis.Experimental results show that the method proposed can effectively do anomaly detection and quality diagnosis on the surveillance video.%监控视频可能会出现一系列影响视频质量的异常情况，如清晰度、亮度异常或者视频画面存在噪声、偏色、画面冻结、水波纹、以及水平杂波等故障的情况。针对这些情况，提出监控视频异常检测与质量诊断算法。算法分析上述各种异常情况，利用图像的灰度统计信息、梯度信息，结合图像的中值滤波和帧间差异分析，完成相应的异常检测与质量诊断。实验结果表明，该方法可以有效地对监控视频进行异常检测与质量诊断。
展开▼
