Monitoring the Environmental Impact of Majdanpek Open Mine Using Machine Learning and Gis

Potić, Ivan and Vakanjac, Boris and Petrović, Stefan (2022) Monitoring the Environmental Impact of Majdanpek Open Mine Using Machine Learning and Gis. In: 6th Scientific-Expert Conference with International Participation „Sustainable Development and Water Protection (Law, Economy and Management)”; Proceedings. Belgrade: Faculty of Busines Studies and Law, “Union – Nikola Tesla” University, Belgrade, pp. 233-244. ISBN 978-86-6102-090-2

[img] Text
2022_Potic_et-al._Monitoring-Uticaja-Otvorenog-Kopa-Majdanpek_Odrzrazvoj-I-Zastitavoda_M63.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.

Download (946kB)

Abstract

After several decades of open pit and underground mining near city of Majdanpek, the environment gave its response to the stresses caused by this human activity. The increasing need for minerals and profits leads to the expansion of mining, endangering not only the biomes that are in the immediate surrounding area of the mines, but also has an impact on the wider ecosystem, especially on surface and ground water. To quantify, comprehend, and mitigate the harmful impacts of the mineral’s extraction from the Earth’s crust, it is required to conduct continuous monitoring of the areas where these operations take place. Considering the large number of factors that obstruct the immediate monitoring of such areas, the solution is reflected in the use of remote detection products that provide relatively fast and precise data on remote areas that are difficult to access. Using medium-resolution satellite images (15 and 10 m spatial resolution) for different time periods (1999 and 2022), land cover maps are created. Prior to automatic classification of satellite images process, five classes were defined to be classified using a machine learning algorithm. After land cover classes determination, a map of the study area is created showing five classes: water, forest, pastures, agricultural areas, and urban environment. The position of the open part of the mine was determined which enables the possibility to monitor the expansion of the open mine pit and position of the pollutant. The second analysis considers the use of a medium-resolution (30m spatial resolution) digital terrain model (DTM). The Geographical Information System (GIS) made possible to perform analyzes that determine the watersheds around the detected open mine pit that is marked as a pollutant. After determining the watersheds, river network analysis is performed and zones with potentially polluted watercourses are marked. Thus, environmental data is collected, which makes the decision-making process much easier for decision-makers and determining the limits of economic and ecological sustainability for the area of interest.

Item Type: Book Section
Uncontrolled Keywords: Remote Sensing, Support Vector Machines (SVM), Sentinel-2, Landsat 7, ASTER DEM
Institutional centre: Centre for demographic research
Depositing User: D. Arsenijević
Date Deposited: 03 Dec 2024 09:45
Last Modified: 03 Dec 2024 09:45
URI: http://iriss.idn.org.rs/id/eprint/2489

Actions (login required)

View Item View Item