thesisnumber: MT0051096 \\ lang: Japanese \\ tool: LaTeX \\ major: 工学 \\ title: 歩行者分布の実時間計測と地図生成 \\ ptitle: \\ etitle: Real-time Observation and the Map Creation of Information about Pedestrians Distribution \\ eptitle: \\ author: 前田 宏二 \\ pauthor: \\ eauthor: Koji Maeda \\ epauthor: \\ date: 2001.2.8 \\ division: 17 \\ supervisors: 千原 國宏 教授, 横矢 直和 教授, 眞鍋 佳嗣 助教授 \\ esupervisors: Kunihiro Chihara, Professor; Naokazu Yokoya, Professor; Manabe Yoshitsugu, Associate Professor \\ cmembers: 千原 國宏 教授, 横矢 直和 教授, 眞鍋 佳嗣 助教授 \\ ecmembers: Kunihiro Chihara, Professor; Kotaro Minato, Professor; Manabe Yoshitsugu, Associate Professor \\ keywords: 歩行者ITS,歩行者分布, 魚眼レンズ,画像処理,程度修飾 \\ pkeywords: \\ ekeywords: Pedestrians for ITS, Pedestrians Distribution, Fish-eye, Image Processing \\ epkeywords: \\ abstract: 近年、歩行者各自が街を歩く際の不自由さの軽減、危険の回避、街中情報の提供を目的とした、歩行者ITS(Intelligent Transport Systems)が注目されている[1]。このシステムの中で、現在や過去の歩行者分布情報(以下分布情報)を地図上に提示して利用できれば有益である。たとえ歩行者専用通路でも、混雑に不自由さや危険を感じることは、高齢者・身体障害者に限らず健常者にもあり、混雑具合の把握や予測をしたいときに障害物情報として利用できる。また、観光客にとって人の集まる場所は観光マップの付加情報にもなり、マーケティングや都市整備計画の際には貴重な検討材料にもなる。本稿では、歩行者分布情報の実時間提示・配信や、統計的情報を引き出すための蓄積を行うために、広範囲を対象とした歩行者分布観測システムを提案する。 \\ pabstract: \\ eabstract: Recently, ITS (Intelligent Transport Systems) for pedestrians is paid much attention and there are many related researches, out of which, Human-Navigation System is the most actively studied. However, pedestrian traffic information is not really utilized. When walking in town, for instance, we often want to avoid pedestrians' congestion, particularly for handicapped and higher age people. If the pedestrian traffic information is available as a real-time map for each pedestrian, it can provide variety of effective services for ITS. Storing the statistics data is also useful not only for individuals' scheduling but the marketing business. This paper proposes the system architecture that allows to monitor the pedestrian traffic by distributed camera terminals and to map the gathered information onto a display server via network. We also propose an effective method to detect traffic information using image processing with wide-angle camera and show some experimental results. \\ epabstract: \\