DEVELOPMENT OF DETAILED BUILDING DISTRIBUTION MAP TO SUPPORT SMART CITY PROMOTION -AN APPROACH USING SATELLITE IMAGE AND DEEP LEARNING–

• DOI: 10.5194/isprs-annals-x-4-w3-2022-189-2022
• Publication Date: 2022-10-14T09:52:31Z
• Authors: K. Okada; N. Nishiyama; Y. Akiyama; H. Miyazaki; S. Miyazawa
• Keywords: Promotion (chess); Demographics

Abstract. Detailed demographics play an important role in the development of smart cities. However, especially developing countries, maintenance and management this data is incomplete, which hinders promotion The objective study to develop a method create detailed building distribution maps from satellite images, will serve as basis for demographic support cities around world. target area several areas Tokyo where validation available. We first developed extracting buildings images then estimating use determine residents are distributed. Both methods deep learning. As result, it was possible extract with extraction rate (the number automatically extracted divided by verification) up 60.3% entire area. In addition, estimation use, our able classify detached non-detached average accuracy 78.7%