Crowd density estimation recognition and crowd behavior analysis algorithm

Abstract

The invention relates to the technical field of computer vision, and discloses a crowd density estimation recognition and crowd behavior analysis algorithm. The algorithm has the advantages that a system can automatically remind each point location abnormal condition, a passenger flow peak period can be predicted to make a deployment scheme in advance, and after a dangerous condition occurs, the system automatically makes a chain reaction to minimize the life and property safety loss of people. The crowd density estimation and recognition system uses current advanced machine vision, graphic image video analysis technique, human body pattern recognition technology, and artificial intelligence machine learning technology, to detect the mobile people flow density of a preset area at a second-level speed. The method has the advantages that the influence of light, shadow and heat can be effectively eliminated through node passenger flow quantity and queuing length statistics, interference objects in a detection area can be self-calibrated and eliminated every two minutes, when the monitored passenger flow quantity reaches a set threshold value, passenger flow alarming is carried out, acorresponding early warning scheme is started, and safety accidents are prevented.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a crowd density estimation recognition and crowd behavior analysis algorithm. Background technique [0002] With the continuous improvement of people's living standards in our country, the phenomenon of "blowout" of crowds often occurs in comprehensive commercial areas of large cities, famous tourist attractions and other areas, especially during holidays. The long stay of personnel increases the number of potential accident sites and factors that cause accidents, further increases the risk of crowd gathering, and increases the difficulty of crowd safety management. The early warning system for crowded places is collected by sensors and cameras at various points (crowd density can be divided into three levels: red, orange, and blue), and the collected data and images are transmitted to the server through the router, and the client can see the location of each point. Dens...

AI Technical Summary

Problems solved by technology

[0003] The traditional way is to rely on manpower to keep an eye on dozens of images, and rely purely on experience to determine the security plan for the day. When danger comes, it is impossible to make a reasonable and correct response quickly; after applying this system, the system will automatically remind each point of abnormal conditions The deployment plan is prepared in advance during the peak period of passenger flow. After a dangerous situation occurs, the system will automatically make a chain reaction to minimize the loss of people's lives and property.

Images