Indoor Analytics based on Time-of-Flight (ToF) Cameras

Understanding human behavior within commercial buildings is critical for optimizing services for both workers and consumers. Accurate information about occupancy patterns, including the quantity, flow, and concentration of people in various areas, is essential for efficient space management, such as optimizing the positioning of amenities or managing queues. While several technologies, such as wireless signals, RFID, and conventional cameras, can be used for indoor tracking, time-of-flight (ToF) cameras offer a privacy-preserving and precise alternative for depth-based human tracking. 

For example, WiFi (IEEE 802.11) tracking can utilize signal strength (RSSI), angle of arrival (AoA), or time of flight (ToF) of Wi-Fi signals to estimate the location of users within indoor spaces. While WiFi tracking methods are useful since they leverage an existing infrastructure, they suffer from limited accuracy due to environmental factors (interferences, multipath effects, etc). Similarly, RFID (Radio Frequency Identification) systems can rely on passive or active tags to track individuals using RFID readers. These methods are suitable in controlled environments where users carry a given RFID tracker, which is not often possible in an open environment where people freely move. Relying on traditional cameras is the alternative that provides the highest level of accuracy, allowing to identify objects, people, poses, and trajectory, with the drawback that they raise ethical concerns regarding surveillance and potential misuse of data.

In this regard, Time-of-Flight (ToF) cameras and Event Cameras are an alternative "vision-based" approach that can cope with the privacy issues of traditional cameras. For example, ToF cameras measure the time it takes for an infrared light signal to travel to objects and return, constructing a depth map rather than capturing detailed color images. Event cameras, register only pixel-level changes in brightness, reducing data storage and processing needs while providing real-time tracking.

The goal of this thesis is to to develop and implement a multi-camera indoor tracking system based on ToF cameras with an edge-computing approach to provide real-time insights into occupancy and movement patterns in office environments. The thesis will evaluate the system's performance in terms of accuracy, processing efficiency, and resource consumption. Additionally, it must assess the privacy implications of using ToF cameras in indoor tracking applications. 

Key references (to be expanded in the thesis):

1. Ristani, E., Solera, F., Zou, R., Cucchiara, R., & Tomasi, C. (2016, October). Performance measures and a data set for multi-target, multi-camera tracking. In European conference on computer vision (pp. 17-35). Cham: Springer International Publishing.

2. Guomundsson, S. A., Larsen, R., Aanaes, H., Pardas, M., & Casas, J. R. (2008, June). ToF imaging in smart room environments towards improved people tracking. In 2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (pp. 1-6). IEEE.

3. Foix, S., Alenya, G., & Torras, C. (2011). Lock-in time-of-flight (ToF) cameras: A survey. IEEE Sensors Journal, 11(9), 1917-1926.

4. Nakagawa, M., & Kobayashi, T. (2016). Real-time floor recognition in indoor environments using tof camera. Asian Association on Remote Sensing, 399-402.

5. Kazmi, W., Foix, S., Alenyà, G., & Andersen, H. J. (2014). Indoor and outdoor depth imaging of leaves with time-of-flight and stereo vision sensors: Analysis and comparison. ISPRS journal of photogrammetry and remote sensing, 88, 128-146.