Identifying Waste Burning Plumes Using High-Resolution Satellite Imagery and Machine Learning: A Case Study in the Maldives

A rapid increase in municipal solid waste generationhas far outpacedresources to manage waste in many developing countries, resultingin the burning of trash in designated landfills or public places,the release of harmful air pollutants, and exposure of nearby populations.While some governments have recently banned open burning at municipalfacilities, monitoring the success of mitigation strategies has beenchallenging due to the lack of adequate air pollution monitoring methodologies.To address this, we have developed a machine learning approach thatutilizes high-resolution (3 m/pixel) satellite imagery and appliedthe methodology to detect plumes of smoke from waste burning on Thilafushiin the Maldives. We employed an image classification and semanticsegmentation model based on a pretrained convolutional neural networkto identify and locate plumes within images. Our approach achievedan average intersection over union (overlap) of 0.70 between visuallyidentified plumes and the machine learning output as well as a pixel-levelclassification accuracy of 96.3% on our holdout testing data. Ourresults demonstrate the potential of machine learning models in detectingplumes from sources where measurements are not available, includingwildfires, coal-fired power plants, and industrial plumes, as wellas in tracking the progress of mitigation strategies.