| 題名 | Aerial Image Discovery System with Accelerated Candidate Selections in Multi-Scale City Image Database |
| 著者 | *Kaoru Uchida, Shuo Hu, Bingjie Ma, Rheza Harliman (Hosei University) |
| ページ | pp. 220 - 227 |
| キーワード | aerial imagery, image matching, frequency analysis, rotation/scale invariance |
| アブストラクト | Aerial image discovery is one of the key technologies in various services using GIS, in which a match is to be identified in a large image database with a given query image with unknown location, orientation, and scale. The task is more challenging in city areas because the views have higher degree of homogeneity compared to non-urban areas. We propose a highly efficient image discovery system to find a match in the multi-layered image database of city aerial images with multiple magnitudes. The system first pre-selects matching candidates based on comparison results of image profiles such as frequency spectrum, then search in the database using scale-invariant matcher based on local feature descriptors. The matching focuses on the appropriate scale layer to accelerate search in the large database. This paper outlines our proposed approach and discusses its feasibility and performance based on the preliminary experimental results from our on-going research work. |