Field intelligent video monitoring method with sound acquisition and recognition fused, device and system

Abstract

The embodiment of the invention discloses a field intelligent video monitoring method with sound acquisition and recognition fused, device and system. Sound data acquired by microphones arranged at different positions in the field are acquired; the sound data are preprocessed and stored; video data acquired by cameras arranged at different positions in the field are acquired; the video data are stored; a sound whose amplitude exceeds a set threshold is subjected to sound source positioning, the sound whose amplitude exceeds the set threshold is subjected to feature extraction, the sound is recognized according to the extracted features, and whether to be an abnormal sound in the field is judged; if yes, according to the sound source position of the abnormal sound, the photographing angle of the camera closest to the sound source position is adjusted; and the video data photographed by the camera with the angle adjusted are acquired, the abnormal sound and the video data at the positioncorresponding to the abnormal sound are fused and are recorded and stored, and an alarm signal is given at the same time.

Description

technical field [0001] The embodiment of the present application relates to the field of video surveillance and the automatic control of the acquisition end of the video surveillance system, and in particular to an intelligent video surveillance method, device, and system for venues that integrate sound collection and recognition. Background technique [0002] The traditional video surveillance system integrates distributed video capture cameras, transmits them to the control host through the network for storage, manually controls, checks and compares them, feeds back to the front-end control of the camera, or retrieves the stored files in the control host Delete operations such as playback. [0003] A camera with a rotating pan / tilt, currently some cameras can support automatic cruising, automatic line scanning, and rotation at a certain speed and angle, that is, automatic rotation video collection according to the preset rules of the video surveillance system. [0004] In...

AI Technical Summary

Problems solved by technology

[0005] Problems in traditional video surveillance: the security personnel directly monitor the screen, which is inefficient. If you watch the surveillance for a long time, you will be easily fatigued and distracted

Unable to meet the requirements of continuous monitoring, unable to distinguish in time, easy to generate a large amount of useless video information, and the identification of useful information becomes difficult

The camera is mainly used for real-time mechanical recording of events that occur, mainly to deter offenders or obtain evidence after the event. It cannot perform real-time intervention and does not play a real-time role in monitoring.

Traditional video surveillance relying on manpower has become increasingly difficult to meet the requirements of real-time monitoring and intelligent traceability in today's society and industrial sites

[0006] Camera pan / tilt control problem: The current camera pan / tilt control performs regular rotation video collection according to preset rules, which easily makes important abnormal situations negligent, and offenders can commit purposeful crimes based on the blind spots brought about by rotation

Due to the complexity of the surveillance scene, only relying on the intelligent video surveillance system to process the image information has certain limitations. Simply relying on the single information of the video image as the analysis source for analysis and judgment may cause information blind spots, resulting in false alarms in the automatic detection and alarm system. or false report

Images