Exploring The Use of Convolutional Neural Networks in UAV Based Rescue Missions

Abstract

Multi-rotor based flight platforms, especially quadcopters and unmanned aerial vehicles (UAVs), and their embedded avionics have become emerging technologies in rescue operations. Due to the simplicity of mass production, as well as air superiority, rapid deployment and vertical take-off and landing (VTOL) capabilities, autonomous driving will be the next step in rescue operations. This project aims to explore the application of UAV combined with deep learning algorithm in drowning scenario. With the help of Betaflight, create an autonomous flight program that implements specific drowning rescue functions. Deep learning algorithms are critical to developing neural recognition networks for drowning individuals, with pre-programmed flight parameters eliminating the need for human interaction, human error to be precise. The core goal of the project is to improve the success rate of lifesaving solutions through the availability of this technology, a longer operating range than lifeguards, and shorter response times.