Open AccessUsing remote sensing data for environmental monitoring of water objects using GIS and machine learning
Author(s) -
Денис Кривогуз,
Anna Semenova,
Sergei Mal’ko
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/937/2/022051
Subject(s) - remote sensing , dependency (uml) , computer science , environmental monitoring , field (mathematics) , environmental data , satellite , environmental science , data mining , artificial intelligence , geography , engineering , mathematics , environmental engineering , aerospace engineering , political science , pure mathematics , law
The main way to understand variability of any spatial data using remote sensing is calculating spectral indices. For now, some difficulties have receiving water surface temperature due to specific properties for satellite sensors and low spatial resolution. The main sources of receiving salinity data are remote sensing data from ESA SMOS, NASA Aquarius and SMAP satellites. Using different machine learning algorithms, we can get models or equations, representing dependency between studied environmental variable and different spectral channels of remote monitoring data. After receiving and collecting remote sensing data in database this system uses machine learning algorithms to find dependency between collected field data and different spectral bands of the remote sensing data. Our goal was to form an analytical system based on remote sensors and machine learning algorithm to analyse, predict and evaluate water ecosystems for fisheries and environmental protection.