DeepDist: A Deep-Learning-Based IoV Framework for Real-Time Objects and Distance Violation Detection

Abstract

Crowd management systems play a vital role in today's smart cities and rely on several Internet of Things (IoT) solutions to build prevention mechanisms for widespread viral diseases such as Coronavirus 2019 (COVID-19). In this article, we propose a framework to aid in preventing widespread viral diseases. The proposed framework consists of a physical distancing notification system by leveraging some existing futuristic technologies, including deep learning and the Internet of Vehicles. Each vehicle is equipped with a switching camera system through thermal and vision imaging. Afterward, using the Faster R-CNN algorithm, we measure and detect physical distancing violation between objects of the same class. We evaluate the performance of our proposed architecture with vehicle-to-infrastructure communication. The obtained results show the applicability and efficiency of our proposal in providing timely notification of social distancing violations.