Crowd Detection And Tracking In Surveillance Video Sequences
Sohail Salim, Othman Omran Khalifa, Farah Abdul Rahman, Adidah Lajis · 2019
The importance for video-based monitoring systems is on the rise leading to the growth of interest in the field of computer vision. With the increase of human population, crowd needs to be monitored, be it in a public place or in a building. Human monitoring can be quite tiresome and expensive, making way for the upcoming of automated crowd monitoring systems. Crowd analysis comprises of detection, tracking, behavioral analysis, etc. In this paper a framework for the detection of crowd, tracking and counting is proposed. The goal is to create a robust system with utmost accuracy in its results. Contrasted with sensor-based arrangements and human-based, the video-based ones take into account more adaptable functionalities, enhanced execution with lower costs. In this work, the dataset PETS2009 was used. The results showed that the proposed system has the capability to count all the people passing through the field of view of the surveillance camera. The system was tested for different types of crowd and the average efficiency of various scenarios is 83.14%.