色情報とエッジ情報を利用した航空写真からの大局的な地震被害の把握(A.被害予測と緊急対応(2),口頭発表)

概要

論文の詳細を見る
To grasp the range and seriousness of the damage is indispensable for starting up quickly life supporting activities after a wide area disaster, such as a big earthquake In this study utilization of aerial photograph images for this purpose is considered We investigated two cases: (1)only and aerial photograph after the earthquake is available, and (2) aerial photographs of the same area before and after the earthquake can be used In the first case, color and edge information is used for detecting the damaged areas We detect brown areas in the hue image, and also areas where the number of edges is large and edge orientations are somewhat uniform in the lightness image Then we make their intersection In the second case, we first match two images by specifying several corresponding points by hand and by applying an affine transformation Checking colors of each corresponding pixel pair in the matched two images, we detect areas having color differences to some extent We apply different criteria to chromatic colors and nonchromatic colors for compensation of the difference of light and view conditions of the two images taken Both methods yield fairly good detection of areas damaged by an earthquake