Intelligent alarm convergence method suitable for audio and video fused platform

Abstract

The invention provides an intelligent alarm convergence method suitable for an audio and video fused platform. The method comprises the steps that a training set is constructed according to basic alarm record data generated by the existing audio and video fused platform; a learning model for alarm convergence which refers to a process of removing invalid alarms and reducing a false alarm probability is constructed; the learning model is trained by using training samples in the training set to obtain a trained learning model; the trained learning model is used to calculate and identify the alarm record input to the current audio and video fused platform, and a classification result including validity, neglect and false alarm is output; an alarm recognizer judges whether the alarm record ofthe current audio and video fusion platform effective through the trained learning model, and if YES, alarm processing is executed; and otherwise, the alarm record is pushed to a worker for checking,and the record is marked or directly deleted.

Description

technical field [0001] The invention relates to the field of audio and video communication, in particular to an intelligent alarm convergence method suitable for an audio and video fusion platform. Background technique [0002] With the trend of industry cloudification and microservices and their applications becoming more and more mature, many companies and enterprises have established a new generation of audio and video integration platforms based on the cloud federation + microservice architecture. Advanced technologies have brought a lot of application value, but at the same time they also face New problems and challenges have emerged, such as larger system clusters and more business components, all of which have brought new challenges to the operation and maintenance of the new-generation audio and video fusion platform. [0003] Multi-dimensional alarms are defined in the existing audio and video fusion platform operation monitoring business. For example, if a business...

AI Technical Summary

Problems solved by technology

Based on the above-mentioned problem of a large number of invalid alarms in the operation monitoring and alarm of the audio-video fusion platform, a mainstream method to solve many artificial intelligence problems is proposed based on the machine learning method to achieve intelligent identification of effective alarms on the audio-video fusion platform data, reducing the probability of false positives

In addition, there are many open source machine learning platforms today, such as Google, Baidu and other domestic and foreign companies have open source their own machine learning platforms, which also greatly reduces the entry threshold, making it no longer far away for small and medium-sized enterprises and traditional enterprises to realize intelligent operation and maintenance

Images