COVOCHA: COmputer Vision-based Occupancy CHAracterization

To improve the indoor environmental quality and energy efficiency of buildings, it is important to provide accurate occupancy information to the automation and management system regulating the different HVAC and lighting devices. For instance, the mechanical ventilation rate of an office room can be lowered when there are no or a few occupants, thus saving energy. On the other hand, high occupancy of a meeting room would require an increase in the mechanical ventilation rate to avoid an excessive indoor CO2 concentration and airborne pollutant build-up deteriorating indoor air quality and occupants’ comfort, well-being and productivity. In addition, acquiring insight on typical occupancy level, activities, clothing level, and other occupancy characteristics for given types of buildings can greatly improve the design of the latter.

This Bridging project aims at developing and benchmarking a computer vision-based tool for the detailed characterization of occupants in the built environment: accurate presence detection, accurate counting of occupants, occupants’ distribution in the indoor space, clothing level, metabolic rate, activity type and distance between individuals. These algorithms are of great interest for the current and future research activities conducted at Aalborg University in the fields of advanced control of smart buildings and the study of occupants’ practices in the built environment.

Participants:
Department of the Built Environment (BUILD), Department of Architecture, Design and Media Technology (CREATE)