Monitoring of Plastic Debris in the Marine Ecosystem with Remote Sensing

Sofiya Tumanova

Bakalářská práce

Abstract:

Marine plastic pollution is a major global environmental problem, having numerous negative effects on marine ecosystems and organisms. The use of remote sensing technologies has the capabilities to detect, analyze and monitor plastic from space on a large spatial scale. This study explored different approaches and methodologies from previous studies and compared their results and effectivities in detecting marine plastic debris. To better understand the characteristics of plastic materials, this study inspected plastic’s spectral properties by using a  spectroradiometer and generated a reflectance graph of different plastic samples. Additionally, with the use of freely available remote sensing data, and an open-source atmospheric correction processor ((Case 2 Regional Coast Color) (C2RCC)), this study utilized the Random Forest classification algorithm to automatically detect plastic pixels in a remotely sensed image. Various combinations of Sentinel-2 satellite’s bands and indices were tested to see which combination will produce an optimal result of plastic detection. The algorithm used in this study was able to detect plastic pixels with a 91.5% accuracy. This study concluded that when used as variables in the Random Forest algorithm, bands 2 (blue), 3 (green), 4 (red), 8 (near infrared), of the Sentinel-2 satellite and the “Plastic Index” are able to produce the best results in terms of plastic detection. Moreover, this study highlighted the importance of the C2RCC atmospheric correction processor in producing vital bands known as “inherent optical properties”, which also served as an input to the algorithm and were able to contribute to the optimal plastic detection result.

Keywords

plastic pollution, plastic detection, remote sensing, marine litter

Jazyk práce: angličtina

Datum vytvoření / odevzdání či podání práce: 1. 4. 2023

Obhajoba závěrečné práce

• Obhajoba proběhla 23. 5. 2023
• Vedoucí: Ing. David Moravec, Ph.D.
• Oponent: Michela Perrone

Česká zemědělská univerzita v Praze

Fakulta životního prostředí

Bakalářský studijní program:
Geographic Information Systems and Remote Rensing in Environmental Sciences