Crowd Size Estimation: Smart Gathering Management

Linear increase in population which results in overcrowding has become an unavoidable element in any public gathering. Public safety under such condition has become a very vital problem in areas like streets, malls and railway stations during weekends, festive seasons, holidays, concerts, etc., normally or in any pandemic situation. The massive disasters that can occur includes numerous instances of fatality where people gather in form of throng. In present time, surveillance cameras are deployed to maintain peace, security and manage crowd, as surveillance videos for proper analysis of crowd activities is an important issue for communal harmony and security; however, some major limitations in video surveillance system are that includes picture getting blurred, peculiarities among person cannot be identified automatically with respect to surroundings during live video streaming, along with that to save the information a lot of storage spaces is also required and hence it becomes costly to run and maintain. The present study proposes a method that is based on principle of Histogram of Oriented Gradients (HOG) and OpenCV that efficiently keeps in track count of the people in the scene which helps in efficient crowd management. OpenCV-based method used for crowd estimation written in Python used in this study in order to count the number of heads in live streaming and helps in crowd management according to requirement in an economical way.