柏受信アンテナで得られた画像データに写る意味を解釈するアルゴリズムの開発

衛星が撮影した画像に写る雲の分類:

Satellite images of shallow clouds show a dazzling variety of organization

patterns. These patterns can crucially impact the role that shallow clouds

play in a changing climate. Naturally, researchers would really like to

understand these complex structures better. In a large number of satellite

pictures, researchers have found that the organization of clouds can be

roughly divided into four types：sugar, flower, fish and gravel. However, it

is not efficient to find the types of cloud in many satellite images

manually. This research verified that U-Net model which is one of the

semantic segmentation models can help researchers to classify the cloud

structure more efficiently.

 

 

超解像を用いた衛星画像の物体検出に関する研究（研究紹介）: