Font Size: a A A

Design And Implementation Of Crowd Density Monitoring System In Outdoor Scenes

Posted on:2018-12-06Degree:MasterType:Thesis
Country:ChinaCandidate:Y Z ZhangFull Text:PDF
GTID:2428330566495774Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the development of economy and the improvement of people's life,people's social activities are increasing constantly,and the density of people in outdoor scenes is increasing day by day.As a result of the crowded density,the occurrence of security incidents is not uncommon and can easily cause bad social influence and serious loss of life and property.At present,although the public places will be equipped with cameras for monitoring,most of them only play a monitoring role but do not have the capability of data analysis.The current monitoring is still based on manual monitoring,which has the characters of high costs and can not do real-time monitoring for 24 hours.Therefore,a monitoring system which can analyze population density in video is urgently needed to be developed.Based on the digital image processing technology and combined with convolution neural network,the design and implementation of crowd density monitoring system in outdoor scenes is completed.The system is designed to detect the crowd density level and the number of population in the video,and output the corresponding population density distribution.When the crowd density level is too high,a timely warning message is sent out.The system uses the pixel-based method to calculate the crowd density value in the video.First,it provides a function of selecting the region of interest to detect the moving objects by hand.Then a method based on gaussian mixture background model method is applied to foreground extraction.The method of combining open and close operations is used to denoising.Finally,according to the total number of foreground pixels and the size of selected,current population density is to be calculated.For the estimation of the number of people,the system uses a trained population density estimation model to test and obtains the population density distribution of the input image,and then sums the crowd density matrix,which is the number of people.Finally,the outdoor crowd monitoring system was tested.The result shows that the system is fast and accurate,and can get the density of people and the total number of people within the monitoring area.For abnormal situations,an warning message will be sent in time.
Keywords/Search Tags:Intelligent monitoring, Crowd density detection, Convolution neural network, OpenCV technology
PDF Full Text Request
Related items