Application of random forest algorithm to Sentinel-1 for plantation detection: case study of Tesso Nilo ecosystem

Giusti Ghivarry; Adhera Sukmawijaya

doi:10.1117/12.2618285

22 December 2021 Application of random forest algorithm to Sentinel-1 for plantation detection: case study of Tesso Nilo ecosystem

Giusti Ghivarry, Adhera Sukmawijaya

Author Affiliations +

Proceedings Volume 12082, Seventh Geoinformation Science Symposium 2021; 1208208 (2021) https://doi.org/10.1117/12.2618285
Event: Seventh Geoinformation Science Symposium (GSS 2021), 2021, Yogyakarta, Indonesia

Abstract

Investigation of possible variations of the combination of VV and VH polarization available for SENTINEL-1 Synthetic Aperture Radar (SAR) imagery has been conducted for plantation detection. The dual-polarization (VV and VH) Interferometric Wide Swath (IW) along descending orbit over Tesso Nilo Ecosystem collected between 1st to 25th March 2021 stored inside Google Earth Engine dataset catalog is used for this objective. Thus initial preprocessing is excluded. The Random Forest method is implemented for two purposes, (1) to identify variable importance of variable used in the classification process, (1) the supervised pixel-based classification. Training samples used in the classification process were collected from visual interpretation of SENTINEL-1 composite image whereas the validation sample is obtained from the Google earth high-resolution imagery. The result shows that variable (VV/VH), (VV-VH), and RVI has the highest degree of importance for oil palm, pulpwood, and forest detection respectively. There is a pattern where the derivative variables of the VV and VH polarization have a high degree of importance. The same pattern appears in the classification results, where scenario 13 ((VV-VH), (VV/VH), ((VV+VH)/2), RVI) has the highest overall accuracy value of 91.74%. Scenario 13 produces the user accuracy of 94.22%, 93.89%, 81.82%, and 95.45% for oil palm, pulpwood, forest, and other land use respectively. The scenario also produces a high producer accuracy of 93.86%, 93.89%, 81.82%, and 97.67%. The combination of available polarization derivative variables with SAR data capabilities can be utilized to build wide-scale plantation monitoring and management systems.

Citation Download Citation

Giusti Ghivarry and Adhera Sukmawijaya "Application of random forest algorithm to Sentinel-1 for plantation detection: case study of Tesso Nilo ecosystem", Proc. SPIE 12082, Seventh Geoinformation Science Symposium 2021, 1208208 (22 December 2021); https://doi.org/10.1117/12.2618285

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available

Members: $17.00

Non-members: $21.00 ADD TO CART

PROCEEDINGS
7 PAGES

DOWNLOAD PAPER SAVE TO MY LIBRARY

GET CITATION

RIGHTS & PERMISSIONS

Get copyright permission Get copyright permission on Copyright Marketplace

KEYWORDS

Polarization

Synthetic aperture radar

Composites

Vegetation

Backscatter

Detection and tracking algorithms

Image classification

Show All Keywords

Keywords/Phrases

Search In:

Publication Years