HRI-JP Honda Research Institute Japan (HRI-JP) – Research and development of advanced technologies

Publications > 熱画像と距離カメラのマルチモーダル情報統合に基づく動的頭部検出システムの検討

Research

publication_search

Advanced Search

September 2010

熱画像と距離カメラのマルチモーダル情報統合に基づく動的頭部検出システムの検討

  • article   (437.23KB)
    Copyright (C) 2010 RSJ 本著作物の著作権は(社)日本ロボット学会に帰属します。本著作物は著作権者であ る日本ロボット学会の許可のもとに掲載するものです。 ※ページ番号=1H3-4
  • Nakamura_2010_681   (1.47KB)
  • 中村圭佑, 中臺一博, インジュギョカン, 中島弘史,
  • in 第28回日本ロボット学会学術講演会予稿集 (RSJ 2010),
  • 日本ロボット学会,
  • 2010,
  • Conference paper

This paper addresses a newly developed human-head detection system utilizing two different modalities, that is, an infrared thermal sensor and a Time-of-Flight (ToF) distance camera. A lot of previous work on human detection has been forcusing mainly on visual sensors, thus it has difficulties in dealing with environmental changes such as illumination. Our multimodal fusion approach is robust for such environmental change. Therefore, it suits for environmental scene analysis. The human detection algorithm is simple enough working in real time with maintaining the performance. It also has a function of human tracking by introducing a particle filter with a low computational cost. The experimental results with the multimodal approach successfully verified a novel performance improvement compared to single modality approach.

Search by Other Conditions

Keywords
Entry type
Years
to
Authors
Language
Refereed