The Research And Implementation Of The Improve The Crowded Density Estimation In Important Area

With the development of society and the acceleration of urbanization, the social activities isfrequent, such as sports game and lantern exhibition. It is imminent problem about urban securitythat how to deal with security problem caused by high crowd density in social activities. The paperresearch on the estimation of the crowd density in the important area, which could be applied to The2014Youth Olympic Games in Nanking. According to the different real-time level of crowd densitygiven by the system, the police can take different methods to secure the important area. The level ofcrowd density is vital factor of the population surveillance and the basis of the populationmanagement.The paper introduce the theory about the development of the population surveillance indomestic and international situation, then have a description of the advantage and disadvantage ofthe density estimation based on pixel-method and texture analysis method.This paper use SupportVector Machine to analysize the eigenvectors about population density. On the basis of the above,the paper combine the pixel-based method and texture analysis method. For low density we usepixel-based method,and texture analysis method for high density. We divide the density into fivelevels: very low,low, normal, crowd,very crowd, and utilize the estimation of population densitybased on Support Vector Machine. The experiments witness the accuration and instantaneity of thesystem.