Abstract
As the world advances it becomes increasingly technology-dependent, bringing together infrastructure and technology to improve the quality of life for the citizens. Smart cities have become the future of urbanization. Since the priority of a city is to protect its citizens, a video surveillance system is required to ensure their safety. This paper proposes a multi-camera cloud-Edge surveillance system for smart cities and homes. Multiple units of Raspberry Pi act as the Edge Computing device that streams and summarizes the processed video footage. After summarizing the video to reduce its length and size, it sends the videos to the cloud (virtual machine). The cloud applies resource-intensive computer vision algorithms such as detecting motion, objects including humans, weapons, and fire. Furthermore, it manages the recorded surveillance videos, stores them in the database, and alerts the user if a threat occurs. The experimental results show that the time taken to perform these tasks was reduced by an average of 83% for the object detection models.