Privacy-Preserving Camera-based Monitoring and Tracking System for Parking Spaces

Camera-based tracking systems have been deployed in a wide range of applications. These systems usually aim to infer the temporal and spatial patterns of people and vehicles, rather than identifying them. Nonetheless, there is a substantial concern nowadays over user privacy - the image and video archives of pedestrians and vehicles may expose their identities and behaviors, which lead to unintended criminal consequences. Particularly, hackers may hijack the control of these camera-based tracking systems for malicious purposes. In this work, we explore a privacy-preserving approach by obscuring the camera by a physical blurry filter. We seek to develop an obscured camera-based tracking system that is capable of offering real-time monitoring of parking space vacancies using only low-cost embedded systems (Raspberry Pi). We evaluated the effectiveness of our system at various blurriness levels. Our system demonstrated high accuracy, despite the obstruction by blurry filters.