This project was commissioned by RQM Engineering to address the challenge of efficiently managing public toilet facilities in Rajshahi through advanced computer vision technology. The primary objective was to develop a sophisticated system capable of accurately counting the crowd within these facilities, thereby enabling informed decision-making for maintenance and resource allocation.

The project posed significant challenges, notably stringent accuracy requirements and hardware limitations, necessitating extensive optimizations to enable operation on edge devices. Despite these challenges, our team successfully delivered on the requirements, marking our inaugural engagement with a government contract.

Fig: People Counting System Working Procedure

Our solution entailed the implementation of a comprehensive human detection, tracking, and counting system within designated areas of the public toilet, including men's and women's zones, as well as the shower room. Leveraging cutting-edge technologies such as WebSockets, Pytorch, and OpenCV, we engineered a robust framework capable of accurately discerning the flow of individuals entering and exiting these spaces.

Fig: Web Dashboard

Furthermore, to facilitate streamlined decision-making, we developed a user-friendly web dashboard. This dashboard provides maintenance authorities with actionable insights derived from the crowd count data, empowering them to make informed decisions regarding cleaning schedules and resource allocation

Our collaboration with RQM Engineering resulted in the successful implementation of an intelligent computer vision system tailored to the unique needs of public toilet management. Through innovation, expertise, and meticulous attention to detail, we have contributed to enhancing the efficiency and effectiveness of public infrastructure maintenance in Rajshahi.