オープンデータ時代の地球観測データ利用小特集 事例紹介
深層学習を用いた衛星画像の超解像手法
2018 年 38 巻 2 号 p. 131-136
詳細
2018 年 38 巻 2 号 p. 131-136
This paper presents a super resolution (SR) method for satellite imagery based on deep learning. From this preliminary study, we found that the performance of SR strongly depends on the variety of the training data set. A new strategy that combines SR and selection is proposed to improve the accuracy and flexibility of the method.
