Crowd management using image processing

ABSTRACT:

      At present, we live in a place where chaos /improper management in areas prone to overcrowding are not dealt properly. This project aims at managing crowd in an efficient method so as to avoid stampede. The idea is to have stationary cameras mounted at every pole in the areas prone to overcrowding to monitor the movement of people. By having an image processing equipment, we may properly monitor the amount of people present by using the frame rate so that when the number of people are above a particular limit, the device can sense the excess in threshold, thereby alerting the authorities that a stampede/overcrowding is about to happen. In special cases where crowd monitoring is required within closed spaces, cameras can be used over entrances and exits to get an idea of the number of people present within the enclosure and have alert systems that will send an alarm to the authorities concerned.

METHODOLOGY:

The idea is to install cameras with poles, 4 meters above the floor at the locations where a crowd is possibly expected or to make use of the existing cameras at various locations. Then, a picture is taken from the video cameras after every 20 second and then it is to be processed. Since the system gets only images from the cameras, another component is required to analyze the images and allow to count the number of people. This device is placed close to each camera in order to avoid sending too much data through the company's network. With the usage of image processing techniques, the number of people in that particular location is identified. In case of a sudden emergency, the system can send data to the main control room which has manual operations and they can alert the possibility of any misfortunes. This system is also the most accurate and can be used and improved to other applications. Whenever it exceeds a threshold count (no. of people in every 5 * 5 m² location), then an alarm is sent to the control room so as to ensure proper monitoring of the crowd. Hence, by this process, the possibility of stampede to occur and the loss of people’s life’s can be reduced.          

PROCESS:
This works by having a raspberry pi controller and having cameras on the top, there by  using texture analysis and image subtraction the number of people who are present can be found out and thus when it exceeds a particular level, it can send an alert to the authorities that over-crowding is occurring and proper management must be done.