Land Cover Classification using Sentinel-1 Radar Mission Interferometry
Loading...
URL
Journal Title
Journal ISSN
Volume Title
Sähkötekniikan korkeakoulu |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
2017-10-23
Department
Major/Subject
Space Science and Technology
Mcode
ELEC3039
Degree programme
NanoRad - Master’s Programme in Nano and Radio Sciences (TS2013)
Language
en
Pages
69+11
Series
Abstract
Synthetic Aperture Radar (SAR) has been widely used for many years in the field of remote sensing. SAR has valuable contribution due to its ability to provide complementary information to optical systems, penetration of radar waves through volumetric targets and high-resolution. SAR has the ability to operate during day and night. It provides operational services under all weather conditions. SAR imagery has many applications including land cover changes, environmental monitoring, climate change and military surveillance. This work focuses on land cover classification with SAR interferometry (InSAR) technique using Sentinel-1 space radar image pair. Sentinel-1 data were collected over the southern part of Estonia. Two SLC SAR images were acquired from both Sentinel-1A and Sentinel-1B with six days temporal difference. In this study, interferometric coherence and backscattering intensity processing chains have been set up and applied to Sentinel-1 SAR image pair. The Sentinel Application Platform (SNAP) has been used for processing of single pair for Sentinel-1 mission. The SNAP is an European Space Agency (ESA) software. The Sentinel-1 image pair processing has been done using Sentinel-1 Toolbox (S1TBX) which is a part of SNAP. Corine Land Cover (CLC) 2012 database has been used as a reference data with 20 m resolution. The CLC2012 contains land use/cover information for most of the European countries. A single optical image from Sentinel-2A was additionally used for feature extraction. An overall accuracy of 68% to 73% was achieved when performing classification into five classes (Urban, Field, Forest, Peat-land, Water) using supervised classification with k-nearest neighbour (kNN) algorithm. The accuracy assessment was done by using confusion matrices.Description
Supervisor
Praks, JaanThesis advisor
Antropov, OlegKeywords
synthetic aperture radar, SAR, land cover classification, InSAR, interferometric coherence, backscattering intensity