Spatial Distribution of Paddy Growth Stage Using Sentinel-1 based on CART Model

The utilization of high temporal Sentinel-1 satellite imagery with radar sensors for identification paddy growth stage monitoring had been used. The Sentinel-1 can provide data for one specific area with the same orbit in every 6 days and freely available to be downloaded. This advantage can be used to identify paddy growth stages that usually have approximately 120 days life cycles. In addition, radar satellite is suitable for tropical region like Indonesia which is cloud free and promises for operational level in monitoring. This paper describes a study of Classification and Regression Tree (CART) algorithm of paddy growth stage based on Sentinel-1 data. Sentinel-1 data for Indramayu region in West Java from January 2021 to April 2021 are downloaded and processed using SNAP software. It is found that the CART algorithm can estimate the spatial distribution of the next paddy growth stages based on previous data. It shows that the paddy growing stage between terrestrial estimates and CART-based SAR image data processing shows a similarity trend and pattern. Total area obtained by CART estimates (pixel based) always larger compares to area obtained by terrestrial estimates which is point based.