Audio and video abnormal state detection method and system based on internet television service

Abstract

The invention provides an audio and video abnormal state detection method based on an internet television service. The method comprises the following steps: S1, collecting historical video stream dataof the internet television service; S2, decoding, cleaning and converting the historical video stream data, and carrying out abnormal phenomenon classification; S3, extracting features of the classified abnormal phenomenon data to obtain an abnormal feature library; S4, acquiring real-time video stream data, and performing feature extraction on the real-time video stream data; and S5, judging anabnormal phenomenon of the real-time video stream data by utilizing the abnormal feature library, and generating an abnormal alarm. The system comprises a video stream acquisition module, a video analysis module, an abnormal characteristic analysis module and an abnormal phenomenon detection module. The system and the method provided by the invention have the advantages that the calculation amountis small, the calculation accuracy is high, a comprehensive and accurate abnormal information base can be established, and comprehensive and accurate abnormal detection is provided for the internet television service.

Description

technical field [0001] The invention relates to a method and system for detecting anomalous phenomena of audio and video data in Internet television services, and belongs to the technical field of audio and video anomalous detection. Background technique [0002] With the development of network technology, the Internet TV business has gradually become the mainstream business of audio and video transmission. Affected by various factors, the audio and video of the Internet TV business will have some abnormalities. Common audio and video abnormalities in the Internet TV business include: : Video black field, monochrome field, still frame, color field, color bar, mosaic, etc., as well as audio mute, left (right) channel volume is too high (low), etc. These abnormal phenomena are all visible and audible to the user, which directly affects the user's viewing perception. [0003] For these anomalous phenomena, the existing solutions mainly include: 1. Utilize the characteristics o...

AI Technical Summary

Problems solved by technology

There are following technical defects and deficiencies in the above-mentioned detection method: 1. The method according to the protocol feature is mainly used to analyze the QoS index of the Internet TV service, that is, network delay, packet loss rate, throughput rate, TCP retransmission rate, MDI-LR, MDI- DF, PAT / PMT table errors, CC errors, etc. QoS indicators can reflect the overall service network quality status, but cannot judge specific fault phenomena; 2. Some abnormal phenomena are abnormal states caused by video content sources, through protocol feature analysis Such problems cannot be found in IP stream data, but the user's viewing perception has been affected; 3. The abnormality analysis based on the image comparison method is very computationally intensive, and requires high hardware configuration and software computing capabilities.

Images