Research On The Crowd Density Estimation Algorithm For Urban Rail Transit Based On Texture Analysis

Along with the progress of the society,the process of urbanization accelerated,which led the urban population density became more and more dense,urban rail transit often has a brief flow of people peak.As the urban rail transit system has the characteristics of passenger flow concentration,closed operation,once highly crowded crowd can produce a variety of emergencies,and may even cause significant casualties and property damage.So it is important to make full use of the intelligent monitoring system to conduct real-time monitoring of the importance of analysis has become increasingly obvious.The estimation of population density in the monitoring scene is of great significance to the maintain the safety order of the place.In this paper,the population density estimation algorithm is further studied.For urban rail transit application scenarios,a population density estimation method based on texture analysis is used to realize the population density estimation in urban rail transit.In terms of population foreground extraction,by analyzing the background difference method to extract the crowd image prospects in urban rail transit,the background modeling is carried out by using the improved mixed Gaussian model.The ideas of the improved algorithm is to give different segmentation area different update rates.The median filter method is used to eliminate the noise.In order to further improve the accuracy of the estimation,it is necessary to correct the perspective effect.In the population density feature extraction,The method based on texture feature of crowd density estimation.From gray level co-occurrence matrix to extract texture features is currently the most commonly used,but the limitations of the texture characteristics are used only crowd image gray level information.In order to fix this defect,a method which combines the gray level co-occurrence matrix and the gray gradient co-occurrence matrix is proposed,the parameters of the gray level co-occurrence matrix are chosen as eight characteristics,the gray gradient characteristics are four,the parameters combined into a 12-dimensional feature vector as a follow-up classification parameters.In terms of classification,the support vector machine is used as the classifier of population density estimation.The optimal combination of kernel function and penalty parameter C of polynomial function is studied experimentally.The population is divided into low density,medium density,high density and very high density by using support vector machine.And experimental verified.In order to further optimization,the series correction idea is added to the population density estimation algorithm.Effectively reduce the false positives.,it is important to be able to alert the crowd suddenly crowded events.So it has important application value in population density estimation and early alarm system.At last,in order to verify the effectiveness of this algorithm.People choose urban rail transit video sequence as the experimental sample.After experiment,the accuracy of the experiment is obviously improved.It shows that the algorithm works well and can be applied in the actual scene of urban rail transit.