INDOOR CROWD DETECTION USING RGB CAMERA AND IoT

In the age of automation the ability to detect the crowd in an indoor environment is impossible without being physically present in the area. IoT (Internet of Things) has moved the field forward by automating anything anywhere. Monitoring the crowd is a critical task in a civilian pl ace. Ubiquitous vision based techniques mostly involve in counting techniques. Counting techniques are prone to error in a dense crowded environment. This paper proposes a technique to detect the density of the crowd with the help of heat map representatio n. The indoor environment which is to be monitored has to be fitted with RGB cameras. Initially the area is captured using the camera without any crowd/people and stored in the database. Once the model is deployed the camera captures the image of the area every second and sends it to the database. A comparison of the already stored map and the newly captured map are performed using the SURF algorithm. The difference recorded is then sent to the database and retrieved by the application and displayed in the form of a heat map. The heat map is regenerated every one second and the density of the crowd is displayed in RGB colors on a heat map. The results show that the heat map gives accurate density of the crowd.