Rice-planted area mapping using SAR and optical satellite data with machine learning in Southeast Asia

Rice is a staple cereal crop in Asia, and the continent accounts for about 90% of global rice production and consumption. Rice-planted area map is important parameter to estimate rice production for food security or economy, and also to quantify the carbon, water cycle or methane emission via paddy fields. Rice is mainly cultivated in the rainy season, Synthetic Aperture Radar (SAR) is therefore a robust tool because it penetrates cloud cover. Recently, machine learning has been widely used in many land cover related researches and distinct results were reported, however, limitation is that it needs a large amount of training data, it is normally time and cost consuming task. In this research, we utilized the combination of SAR and optical satellite data to efficiently produce the training data and identify rice-planted area. Training data were generated from the classification results derived from optical satellite data for the sampled regions, then a random forest classifier was applied to ALOS-2 PALSAR-2 ScanSAR data to identify rice-planted area of both rainy and dry season in Southeast Asian countries. It is also difficult to identify rice-planted area in this region since there are high variations in rice phenology. In order to compensate for the variations, we calculated metrics such as minimum, maximum, average of each polarization (HH and HV), and differences of HH and HV data from time-series ALOS-2 data. Classification models were fine-tuned for each country. Independent verification through visual interpretation using very high resolution images (VHRs) showed a high consistency with the classification results. The developed rice-planted area maps showed high accuracy in most countries and regions, however, verification using in-situ data or national statistics, as well as application to other regions, and years, is necessary to confirm the effectiveness of proposed methodology.