Implementation of U-Net for Paddy Field Mapping Using Very High-Resolution Satellite Imagery

Abstract

Mapping rice fields using remote sensing is one method that can be used to determine the number of rice fields, especially in Indonesia. Using this method can increase effectiveness in agricultural resource management. This research uses Pleiades optical satellite image data with very high resolution which is capable of displaying data information on a larger scale. The rice field classification model in this study uses U-net to classifier between rice fields and non-rice fields. The performance of applying this model for the classification of paddy fields and non-rice fields is 96%. These results show that the U-net model is capable of classifying small rice fields with high accuracy