The detection of human presence using a method outside camera surveillance demanded the incorporation of sensors and additional technology to the premise or area under observation.

An IoT-based project which involves image processing and video analysis, wherein we developed a fully automated module for obtaining person count using a thermal sensor unit. The thermal sensor constantly captures inputs or heat values from the surroundings and the software module transfers it to the processing unit for dynamic real-time people detection and headcount estimation.

• The thermal sensor constantly captures thermal readings from the surroundings and these are received as continuous time series data values by the processing unit.

• The processing unit uses these inputs to perform detection, calculation of the volume of connected components and removal of irrelevant components, etc. by employing various image processing techniques.

• The blob detection algorithm facilitates the detection of human presence and subsequently the count is processed and updated by applying the suitable tracking algorithms.

• The processing unit can efficiently detect human presence and update the count for versatile and random scenarios such as multiple person entry in the same direction or different directions, simultaneous entry and exit of persons etc.



The project will further scale up in terms of functionalities by employing deep learning models in its second phase.