CO2FISHEYE Dataset: CO2 Measurements versus Fisheye Cameras

Motivation

Occupancy estimation from CO₂ measurements is currently the only method used in practice to save HVAC energy in commercial spaces.  However, new approaches to occupancy sensing are emerging, such as surveillance cameras, depth sensors, thermal imagers. We have compared occupancy-sensing performance of CO₂ sensors with that of two overhead fisheye cameras supported by an advanced computational algorithm (J. Konrad, M. Cokbas, P. Ishwar, T.D.C.Little, and M. Gevelber, “High-accuracy people counting in large spaces using overhead fisheye cameras,” Energy Build., vol. 307, p. 113936, Mar. 2024.). Details of our study can be found here. To facilitate this comparison, we collected CO₂ measurements over 3 days in a large classroom with highly-dynamic occupancy, as well as ground-truth people counts and people counts produced by the above algorithm using two fisheye cameras.

Description

CO2FISHEYE dataset has been collected by the Visual Information Processing (VIP) Laboratory at Boston University over 3 days (Wednesday-Friday) of a typical semester week in October 2022 in a large classroom with highly-varying occupancy. The dataset consists of 3 sets of data files:

• CO₂ measurements: 3 data files (Day 1, Day 2, Day 3) with precise time stamps (8,265 time instants in total, each with 6 CO₂ measurements),
• ground-truth people counts and people counts estimated from two overhead fisheye cameras: 3 data files (Day 1, Day 2, Day 3) with precise time stamps (50,338 time instants, each with 2 counts of people),
• air inflow into the room: 3 data files (Day 1, Day 2, Day 3) with precise time stamps (626 time instants, each with one air inflow value).

Data Format

All data are provided in CSV format (comma separated values). For details regarding the data, please see the README.txt file included with the CSV files.

Dataset Download

You may use this dataset for non-commercial purposes. If you publish any work reporting results using this dataset, please cite the following paper:

M. Cokbas, V. Pyltsov, J. Zolkos, M. Gevelber, and J. Konrad, “A comparison of occupancy-sensing and energy-saving performance: CO₂ sensors versus fisheye cameras,” Energy Build., vol. 321, p. 114652, Aug. 2024.

To access the download page, please complete the form below (tested only in Chrome).

Contact

Please contact [mcokbas] at [bu] dot [edu] if you have any questions.

Acknowledgements

The development of this dataset was supported in part by the Advanced Research Projects Agency – Energy (ARPA-E), within the Department of Energy, under agreement DE-AR0000944, and by the Boston University Campus Climate Lab Program.

We would also like to thank Boston University students for their help in the recording and annotation of our dataset.