Mapping paddy field in East Japan using sentinel-1 time series assisted by sentinel-2

Shimpei Inoue, Chinatsu Yonezawa, Akihiko Ito

Research output: Contribution to conferencePaperpeer-review

Abstract

The distribution of paddy fields affects greatly regional budget of methane (CH4), which is the second influential greenhouse gas. Therefore, broad scale paddy field map is fundamentally important for methane budget estimation. However, only a few paddy field maps have been developed, because there remain serious technical issues making it difficult. For example, rice cultivation period overlaps roughly with rainy and cloudy seasons and differs among regions due to complicated factors such as rice cultivar and water management. To solve these problems, several methods using Sentinel-1 time series have been proposed, because it is robust against cloud-cover conditions. However, these methods that use only Sentinel-1 time series have an issue in accuracy of extracting paddy field, because the SAR sensor of Sentinel-1 is less sensitive to vegetation and flooded condition than optical sensors with multi-bands. In this study, to compensate for the weakness of Sentinel-1 time series in paddy field mapping, we propose a new method assisted by Sentinel-2, which is more reliable and robust for the extraction of flooded condition over various land surfaces. Our new method consists of two major steps. In Step 1, paddy fields were provisionally extracted by the thresholds based on the seasonal variation and the flooded condition using the time series of VH-polarized backscatter of Sentinel-1A. In Step 2, among the pixels specified to be provisional paddy field in Step 1, those which were classified as 'non-flooded' by the LSWI, NDVI and EVI calculated using the Sentinel-2 data during a flooded period, were excluded. Finally, we developed a paddy field map for East Japan in 2018 with 30m spatial resolution. The results show that our proposed method assisted by Sentinel-2 was effective in reducing mis-classifications especially for areas cultivated by other crops.

Original languageEnglish
Publication statusPublished - 2020 Jan 1
Event40th Asian Conference on Remote Sensing: Progress of Remote Sensing Technology for Smart Future, ACRS 2019 - Daejeon, Korea, Republic of
Duration: 2019 Oct 142019 Oct 18

Conference

Conference40th Asian Conference on Remote Sensing: Progress of Remote Sensing Technology for Smart Future, ACRS 2019
Country/TerritoryKorea, Republic of
CityDaejeon
Period19/10/1419/10/18

Keywords

  • Paddy field mapping
  • Remote sensing
  • Sentinel-1
  • Sentinel-2

ASJC Scopus subject areas

  • Information Systems

Fingerprint

Dive into the research topics of 'Mapping paddy field in East Japan using sentinel-1 time series assisted by sentinel-2'. Together they form a unique fingerprint.

Cite this