Smart Surveillance Video Stream Processing at the Edge for Real‐Time Human Objects Tracking

This chapter introduces an edge computing based smart surveillance system. It discusses and compares the computations and algorithms used at the edge and fog levels to create such automated surveillance system. The chapter briefly introduces the human object identification algorithms that are potentially feasible in the edge computing environment, followed by the object tracking algorithms. Object tracking plays an important role in human behavior analysis in smart surveillance systems. It discusses the design issues of a lightweight human object detection scheme. Due to the constraints on resources, lightweight algorithms are required for the edge devices. There are two important components to building a good object detector: the feature extractor and the classifier. The chapter presents a case study using Raspberry Pi as the edge device. The case study provides more information about the algorithms that are applied to process sample surveillance video streams.