Paddy rice mapping in Red River Delta, Vietnam, using Sentinel 1/2 data and machine learning algorithms

This study focuses on building a paddy rice map for the Red River Delta region (Vietnam) during the spring crop of 2019. Experiments were conducted with traditional machine learning models (XGBoost, LightGBM) and deep learning models (U-Net, Linknet, DeeplabV3+) based on satellite data from Sentinel...

Full description

Saved in:
Bibliographic Details
Published inJournal of spatial science Vol. 69; no. 1; pp. 103 - 119
Main Authors Ngo, Truong Xuan, Bui, Nam Ba, Phan, Hieu Dang Trung, Ha, Hoang Minh, Nguyen, Thanh Thi Nhat
Format Journal Article
LanguageEnglish
Published Taylor & Francis 02.01.2024
Subjects
deep learning
machine learning
Paddy rice mapping
Red River Delta
Sentinel 1
Sentinel 2
Online AccessGet full text

Cover

More Information
Summary:This study focuses on building a paddy rice map for the Red River Delta region (Vietnam) during the spring crop of 2019. Experiments were conducted with traditional machine learning models (XGBoost, LightGBM) and deep learning models (U-Net, Linknet, DeeplabV3+) based on satellite data from Sentinel 1, Sentinel 2, and topographic maps. The experimental models all gave good evaluation results. The rice maps have good agreements with statistics from the government. The results highlighted that the combination of synthetic aperture radar and optical data with machine learning, deep learning models is an effective approach for short-term high-resolution paddy rice mapping.
ISSN:1449-8596
1836-5655
DOI:10.1080/14498596.2023.2174196