Human Detection for Flood Rescue: Application of YOLOv5 Algorithm and DeepSort Object Tracking

Abstract

This thesis proposes a method of human detection using high-resolution surveillance cameras to monitor sections of the Chattahoochee River that require frequent search and rescue efforts due to flooding. The areas of interest are located in the city of Columbus, Georgia. The goals of this study are to evaluate the effectiveness of the YOLO (You Only Look Once) algorithm for human detection on the river as well as to propose future improvements to the city’s existing alert methods in the event of a flood.