Passage passenger traffic density estimation method

Abstract

The present invention discloses a passage passenger traffic density estimation method, which comprises: establishing a multivariable linear model for estimation, acquiring a training sample set of the multivariable linear model in the class of scenes, and adopting multivariable linear regression analysis to calculate model parameters to obtain a complete multivariable linear regression equation; and extracting corresponding features from a monitoring video image to substitute into the established multivariable linear regression equation so as to real-timely carry out passenger traffic density estimation. According to the passage passenger traffic density estimation method, passenger traffic density estimation is performed based on multivariable linear relation among front view pixels, human face region pixels, front view outer / inner edge pixels, and passenger traffic density, such that the method is suitable for low density population, and estimation accuracy in the case of high density population is increased.

Description

technical field [0001] The invention relates to a technique for estimating the passenger flow density of one-way passages such as entrances, passages, and escalators in various public places, which can be widely used in the one-way passenger flow passages of subways, stations and other places for monitoring and estimating the flow density of people. Background technique [0002] With the rapid development of the economy and the improvement of people's living standards, more and more people are pouring into cities. For many public facilities in the city, such as shopping malls and subways, there will be a peak flow of people in a short period of time. In recent years, stampede casualties have also occurred in various places due to overcrowded flow of people. Therefore, it is very important to conduct real-time statistics and analysis of passenger flow density in these places. In addition, crowd density estimation technology plays an important role in the fields of public saf...

AI Technical Summary

Problems solved by technology

Among them, the method based on pixel analysis is based on the approximate linear relationship between the crowd density and the number of pixels, and the estimated result of the crowd density can be directly obtained, but it cannot solve the problem of overlapping of high-density crowds, because the larger the crowd density, the greater the overlap. The more serious, the crowd density and the number of pixels at this time are no longer the previous linear relationship

The method based on texture analysis is based on the fact that there is a corresponding relationship between the density of the crowd and the thickness of the texture pattern, which can well solve the problem of overlapping occlusion of high-density crowds, but there is a large error in low-density monitoring

The method based on individual target analysis is mainly to identify and track targets to distinguish different target individuals, but when the crowd density is too large, target segmentation is very difficult, and this method is very complicated and difficult to meet real-time requirements

Images