Using Drone Technology and Deep Learning to Map Plastic Pollution and Waste Disposal Locations in a Terrestrial Environment

Keywords: drones, plastic pollution, Africa, Waste management
Session Type: Virtual Paper Abstract
Day: Friday
Session Start / End Time: 2/25/2022 05:20 PM (Eastern Time (US & Canada)) - 2/25/2022 06:40 PM (Eastern Time (US & Canada))
Room: Virtual 30

Abstract

There exists heightened concern about the threats posed by plastics to natural ecosystems, with growing evidence of the harm presented to economies, public health, and society. Although plastic pollution is an issue of great concern, there are still challenges to monitoring the disposal of plastics into the environment. The traditional approach for identifying the location of plastics involves field surveys, including attempts to map plastics at beaches and in oceans using drones and satellites. In this study, I explore the potential for using drone technology to map plastic pollution in actual terrestrial environments by determining optimal parameters for mapping plastics in a controlled community. Targeting Ndirande neighborhood in Southern Malawi (Africa), I assess the performance of different deep learning algorithms at detecting plastics and waste dumpsites, to generate vital information that can help in design and implementation of policies and programs aimed at curbing plastic disposal.