Road Extraction using Result of Divided K-Means Algorithm from Aerial Images

概要

論文の詳細を見る
In application of geographical information system, it is important to extract artificial objects on the ground automatically from aerial images. We aim at road which has parts of connection between areas. But analysis of aerial images needs much time to process because the images are very large. In this paper, we present a method to reduce the time to analysis by dividing images. The size of image to be divided depends on edge density. If a region has high edge density in image, the region is divided into smaller size. Also we present a method using color feature on the assumption that road has similar color. Using these methods, we propose road extraction method using divided K-Means algorithm, in which we use color (CIE L*u*v*) and location (x, y) information. Using the result of road extraction, road map is described according to a logical model of road. We show the result of road extraction. Finally, we evaluate the computational cost and the quality of the result of road extraction.
社団法人日本写真測量学会の論文