@techreport{oai:ipsj.ixsq.nii.ac.jp:00052196, author = {小野, 晋太郎 and 池内, 克史 and Shintaro, Ono and Katsushi, Ikeuchi}, issue = {51(2006-CVIM-154)}, month = {May}, note = {実世界空間、特に都市や遺跡などの広域空間の仮想モデルを計算機内に構築・提示する研究が高い関心を集めている。大規模・複雑な空間を効率よく計測するためには、複数センサデータの統合や移動体による計測が必要である。本論文では、獲得した画像データ・幾何データの時空間解析を通じて移動体に積載したカメラ画像の自動統合や自己位置推定を行うとともに、それらの応用として運転シミュレーションや遺跡の狭隘部を計測するシステムを開発する。, It has been highly interested in the field of computer vision and graphics etc., to construct and represent virtual models of the real world in a computer. Especially, modeling of large-scale scene such as urban environment and cultural heritages are recently expected to be applied to various purposes. In order to acquire informations of such large-scale and complex scenes with efficiency, integration of multiple sensors and scanning by moving sensors are necessary. This paper proposes several methods for integrating video images and localizing sensors which are set on movable bodies through spatio-temporal analysis of image and range data acquired. Additionally, applicational systems to simulate driving road scene and scanning narrow areas in cultural heritages are developed.}, title = {移動体センサから得られる画像・幾何データの時空間解析とその応用}, year = {2006} }