2A1-D21 Pedestrian Detection and Following based on Matching for Mobile Robot
スポンサーリンク
概要
- 論文の詳細を見る
This paper presents a novel method for real-time pedestrian detection and following by a mobile robot. In this method motion areas are segmented from real-time image frames captured by a monocular camera. From the segmented and filtered motion areas, contour is detected using a proposed scan conversion algorithm. The normalized contour is then compared with predefined pedestrian's shape contour templates. When satisfactory match is found the pedestrian is localized and bounded using a rectangular box. After localized the pedestrian, if the user instruct the robots to follow the pedestrian the mobile robot will follow. The effectiveness of the system is evaluated using a mobile robot-PeopleBot, where the robot follows the pedestrian movements.
- 一般社団法人日本機械学会の論文
著者
-
Hasanuzzaman Md.
National Institute Of Informatics (nii)
-
Inamura T.
National Institute Of Informatics (nii)