Paddy Rice Mapping in Thailand Using Time-Series Sentinel-1 Data and Deep Learning Model

Abstract

The elimination of hunger is the top concern for developing countries and is the key to maintain national stability and security. Paddy rice occupies an essential status in food supply, whose accurate monitoring is of great importance for human sustainable development. As one of the most important paddy rice production countries in the world, Thailand has a favorable hot and humid climate for paddy rice growing, but the growth patterns of paddy rice are too complicated to construct promising growth models for paddy rice discrimination. To solve this problem, this study proposes a large-scale paddy rice mapping scheme, which uses time-series Sentinel-1 data to generate a convincing annual paddy rice map of Thailand. The proposed method extracts temporal statistical features of the time-series SAR images to overcome the intra-class variability due to different management practices and modifies the U-Net model with the fully connected Conditional Random Field (CRF) to maintain the edge of the fields. In this study, 758 Sentinel-1 images that covered the whole country from the end of 2018 to 2019 were acquired to generate the annual paddy rice map. The accuracy, precision, and recall of the resultant paddy rice map reached 91%, 87%, and 95%, respectively. Compared to SVM classifier and the U-Net model based on feature selection strategy (FS-U-Net), the proposed scheme achieved the best overall performance, which demonstrated the capability of overcoming the complex cultivation conditions and accurately identifying the fragmented paddy rice fields in Thailand. This study provides a promising tool for large-scale paddy rice monitoring in tropical production regions and has great potential in the global sustainable development of food and environment management.

Su Kaynakları Yönetiminde Makine Öğrenmesi: Çeltik Sulaması Uygulama Örneği

“…However, this data is based on the farmer's decleration as stated above. It is known that mostly declared statements do not represent real production areas (Sitokonstantinou et al, 2021;Xu et al, 2021). As can be seen here, 27.69 km 2 more paddy rice cultivation was calculated in Ipsala district in 2021 than declared.…”

Mapping Paddy Cropland in Guntur District using Machine Learning and Google Earth Engine utilizing Images from Sentinel-1 and Sentinel-2

“…Rice is ingested by more than 50% of the Earth's inhabitants, and rice paddies cover more than 10% of the total cropland worldwide. On a global scale, rice is the main contributor among plant-based sources to the discharge of greenhouse gases [1]. Precise and current data on rice production are of paramount importance in achieving food and environmental objectives and safeguarding access to water.…”

“…In particular, the two-satellite constellation of Sentinel-2 enhances its temporal resolution to an impressive interval of 5 days. Numerous research efforts have presented the effective utilization of S1 data for the mapping of rice fields in different nations, such as Vietnam [9][10], China [11][12][13][14], the USA and Spain [15], India [16][17], the Mediterranean region [18], Iran [19], Bangladesh [20], South Korea [21], Indonesia and Malaysia [22], Philippines [23], Thailand [1,24], and Myanmar [25]. Simultaneously, S2 has also proven its efficiency in accurately mapping rice fields in the context of Egypt [26] and China [27][28][29].…”

Rice Mapping in a Subtropical Hilly Region Based on Sentinel-1 Time Series Feature Analysis and the Dual Branch BiLSTM Model

“…Kussul et al used 19 scenes of Sentinel-1A and Landsat-8 data and a shallow CNN for crop classification in Ukraine, which showed that the CNN architecture outperformed the multilayer perceptron architecture, resulting in accuracies above 85% for major crops [35]. Xu et al extracted temporal statistical features based on 758 scenes of Sentinel-1 images from late 2018 to 2019, covering Thailand, and fed the features into a U-Net model with a fully connected conditional random field to generate an annual rice map [36]. Experimental results showed that the method achieved the best overall performance compared to the SVM classifier and the feature selection strategy-based U-Net model, and the overall accuracy reached 91%.…”