Monitoring video data real-time screening method for anomaly detection

The invention discloses a monitoring video data real-time screening method for anomaly detection, and the method comprises the steps: adding mixed distortion on a monitoring image, analyzing the quality difference between data of different distortion levels, and determining a screening threshold value corresponding to each distortion type; judging the distortion type of the input video frame through the established distortion classification network module; obtaining a quality score of the input video frame through the established image quality evaluation module; and screening and shunting the input video frames, excluding and storing the video frames of which the quality scores are higher than a screening threshold, and inputting the video frames of which the quality scores are lower than the screening threshold into an anomaly detection network for anomaly detection. According to the method, video frames with low quality can be accurately screened and stored, the screening method is applied to anomaly detect