Computer Vision-based Social Distancing Surveillance with Automated Camera Calibration for Large-scale Deployment

Abstract

Social distancing has been suggested as one of the effective measures to break the chain of viral transmission in the ongoing COVID-19 pandemic. We herein describe a computer vision-based AI-assisted solution to aid compliance with social distancing norms. The solution consists of modules to detect and track people, and to identify distance violations. It provides the flexibility to choose between a tool-based mode requiring user input or a fully automated mode of camera calibration (devised in-house), making the latter suitable for large-scale deployments. We also outline a strategy to estimate the number of video feeds which can be supported in parallel for scalability. Finally, we discuss different metrics to assess the risk associated with social distancing violations, including the use of “violation clusters”, and how we can differentiate between transient or persistent violations. Our proposed solution performs satisfactorily under different test scenarios, processes video feed at real-time speed, as well as addresses data privacy regulations by blurring faces of detected people, making it ideal for deployments.