Automatic Detection and Counting Passersby Using Motion-Based Tracking

カテゴリ: 部門大会

論文No: SS3-7

グループ名: 【C】平成29年電気学会電子・情報・システム部門大会講演論文集

発行日: 2017/09/06

タイトル(英語): Automatic Detection and Counting Passersby Using Motion-Based Tracking

著者名: Harras Abdelrahman(徳島大学),辻 明典(徳島大学),寺田 賢治(徳島大学)

著者名(英語): Abdelrahman Harras|Akinori Tsuji|Kenji Terada

キーワード: 自動計数|モーショントラッキング|歩行者歩行者|Automatic Counting|Motion Tracking|Passersby

要約(日本語): Automatic detection and counting moving objects are essential technique in the field of computer vision used for many applications such as counting large number of objects in buildings, roads, and railway stations, etc. The purpose of this study is to detect and track moving passersby by using motion-based tracking. The background subtraction algorithm based on a Gaussian model is applied to detect the passersby. In addition, morphological operations are applied to eliminate noise and the Kalman filter is used to determine a motion of each tracked passerby. Our developed system used a USB camera fixed at the entrance of hall in a laboratory environment to obtain the image sequence. From the experimental results, our system achieved the detection ratio is 94% and the tracking is 88 % correct in case of one passerby. The detection ratio is 69% and the tracking is 61% correct in case of two passersby.

PDFファイルサイズ: 374 Kバイト