街路画像を用いたセマンティックセグメンテーションによる歴史的町並みの研究

−兵庫県の西国街道宿場町を対象として−

抄録

In this study, we aimed to contribute to a better understanding of the current state of the historic streetscapes in the post towns along SAIGOKU-KAIDO in Hyogo, which are currently inadequately preserved. In the analysis, we used semantic segmentation, a type of deep learning, to look for the spots with a high proportion of historical components, the ones that were supposed to be eye-catchers, etc. The analysis produces maps with the characteristics of the post towns along SAIGOKU-KAIDO, which enables you to grasp the current situations and organize where to re-landscape.