A method to estimate crowd size and crowd count

Abstract

Crowd is known when people in large number gather together. When crowd gather for any reasons, it may cause any serious problem for administration. So authority must know about actual situation of crowd. In recent years need for crowd size estimation, has arisen to manage the crowd well. It can help in managing crowd before happening of huge event, traffic management, providing security to crowd and in many other areas where to manage crowd, some estimation on crowd size is needed. After reviewing existing applications, it can be concluded that there are many approaches for estimating density of a crowd and to count the individual’s, but most of the available approaches did not estimate both crowd density and crowd count together using single model. To overcome these restrictions, a single model is proposed to estimate crowd count and density in the crowded scenes. Here two models are prepared in four stages that work individually as well as forms steps for combined approach. These models firstly pre-process the crowd images extract features and then classify into different classes. Crowd density model classifies images into several density classes. Firstly, dataset of crowd images is prepared according to the requirement using crowd dataset. These crowd images are then sorted and provided to next models. Combined Crowd density and Count Estimation model classifies images into various densities and count classes.