Intelligent recognition platform video monitoring early warning system

Abstract

An intelligent recognition platform video monitoring early warning system comprises front-end equipment video-shooting, distribution service transfer, a video decoding program, a video analysis program, a resource server, program processing reception, a database, an FTP server, and service-end display. The video monitoring early warning system adopts Hoff linear detection, and is a joint system combining recognition of knives and sticks based on a classifier plus direction gradient histogram features and recognition of pedestrians based on direction gradient histogram features plus a classifier, and therefore, dangerous knives or sticks can be detected accurately in a slightly more complicated environment, and dangerous behaviors can be recognized; a mode of quick, convenient and efficient early warning is achieved through the functions of event detection, event early warning and event viewing; swarm behavior detection is realized by adoption of a social force model; and by combining a color model and a motion model, an intelligent analysis module developed for the video monitoring system can make a diversified and multi-rule intelligent analysis of multiple occasions at the same time.

Description

technical field [0001] The invention relates to a monitoring and early warning system, in particular to an intelligent identification platform video monitoring and early warning system that can realize real-time monitoring and immediate response. Background technique [0002] In recent years, the intelligent monitoring system has entered the stage of popularization in my country, and is widely used in public places such as banks, hotels, supermarkets, airports, stations, etc., and plays an increasingly important role in the field of public safety. [0003] The existing monitoring system is largely used as a "video query system" after the event, which not only wastes a lot of information data resources and storage equipment, but also wastes a lot of query time, but fails to fundamentally solve the problem of "prevention before it happens". In addition, the existing detection function is simple, and can only detect basic intrusion detection, tripwire detection, cross-border de...

AI Technical Summary

Problems solved by technology

[0003] The existing monitoring system is largely used as an after-the-fact "video query system", which not only wastes a lot of information data resources and storage devices, but also wastes a lot of query time, but fails to fundamentally solve the problem of "prevention before it happens". The purpose and function can no longer meet the needs, and it is impossible to truly realize real-time monitoring and instant response; and the existing detection functions are simple, and can only detect basic intrusion detection, trip wire detection, cross-border detection, left items, lost items, and personnel gathering , fast movement, etc. These technical detections assist staff to realize rapid discovery, rapid processing, and rapid evidence collection, but the accuracy of detection is affected by various external factors, and the accuracy rate is not high;

Images