@article{oai:ipsj.ixsq.nii.ac.jp:00234952, author = {永田, 吉輝 and 浦野, 健太 and 米澤, 拓郎 and 河口, 信夫 and Yoshiteru, Nagata and Kenta, Urano and Takuro, Yonezawa and Nobuo, Kawaguchi}, issue = {6}, journal = {情報処理学会論文誌}, month = {Jun}, note = {人流データは,運輸や施設運用,災害対策等多くの分野で活用が広がっているが,その利活用にはプライバシやコスト等様々な課題が存在する.一方,公共施設の人流分析では,数分の誤差・数%の人数誤差を含む人の移動傾向をとらえられれば十分な場合が多い.本稿では,大規模施設内の各エリアを結ぶゲートで,プライバシフリーな通過情報を収集し,広域な人の移動経路を推定する手法を提案する.本稿におけるプライバシフリーな通過情報とは,人の通過時刻と通過方向に加えて,匿名性の高い人の特徴を通過ごとに記録した情報である.本稿の手法は,このプライバシフリーな通過情報を利用して,データの収集段階からプライバシを考慮した移動経路推定を実現する.実証実験では,中部国際空港において,移動経路推定用の通過情報と正解用の人追跡データを収集した.その結果,許容通過時間誤差が3分以内という条件で,エリア内の人の入退場推定における精度は81%であり,シミュレーション結果とおよそ一致した.また,より高精度な通過センサや複数の人の特徴を利用した場合のシミュレーション結果では,より高い精度を実現した.よって本手法は,センサの設置間隔の調整や収集する通過情報の粒度を調整すれば,公共空間における移動経路推定に利用可能であると期待される., People flow data is widely used in many fields, such as transportation, facility operation, and disaster prevention, but there are various privacy and cost problems associated with its use. On the other hand, in the analysis of people flow in public facilities, it is often sufficient to capture the movement trends of people, including errors of a few minutes and errors of several percent in the number of people. In this paper, we propose a method for estimating wide-area travel paths of a person by collecting passing data at gates connecting various areas of a large-scale facility. Since the passing data does not contain information that completely identifies individuals, it is possible to estimate privacy-free travel routes. In a demonstration experiment, we collected passing data for estimating travel routes and human tracking data for correct answers at the Central Japan International Airport. The results showed that with an allowable transit time error of 3 minutes or less, the accuracy in estimating the entry/exit of a person in an area was 81%, which is approximately the same as the simulation results. The simulation results also suggest that higher accuracy can be achieved by changing the accuracy of the sensor and the human characteristic information used. Therefore, it is expected that this method can be used for travel route estimation in public spaces by adjusting the sensor spacing and the granularity of passing data to be collected.}, pages = {1018--1032}, title = {プライバシフリーな通過情報を用いた大規模施設における移動経路推定手法}, volume = {65}, year = {2024} }