Self-Localization via Server Communication and Development of Keypoint Map

• Project/Area Number: 26330305
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Intelligent robotics
• Research Institution: National Institute of Advanced Industrial Science and Technology
• Principal Investigator: Banno Atsuhiko 国立研究開発法人産業技術総合研究所, ロボットイノベーション研究センター, 研究チーム長 (70356187)
• Research Collaborator: YOKOZUKA Masashi
• Project Period (FY): 2014-04-01 – 2017-03-31
• Project Status: Completed (Fiscal Year 2016)
• Budget Amount: ¥4,810,000 (Direct Cost: ¥3,700,000、Indirect Cost: ¥1,110,000); Fiscal Year 2016: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000); Fiscal Year 2015: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2014: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
• Keywords: 3次元環境構築 / 自己位置推定 / 画像特徴点 / 全方位画像

Outline of Final Research Achievements

We developed a method to acquire three-dimensional map only from images taken by several types of cameras using the technique of Structure from Motion. We set up a route of Tsukuba Mobility Robot Experimental Zone, and acquired a lot of movies with high definition omnidirectional cameras. Two types of methods for 3D reconstruction were produced: sequential and global methods. In order to cope with an increase of the number of images, we introduced an approximate solution method was introduced. As a result, computation time was reduced less than 3 % of the conventional method while suppressing accuracy degradation. Furthermore, instead of treating the omnidirectional image as a two-dimensional conventional image, three-dimensional pixel grid was set as a spherical image, and feature points and feature point descriptors could be directly extracted from the spherical surface. Using these, we developed a system that calculates the self-position on the map when a query image was taken.

Research Products

• [Int'l Joint Research] スイス連邦工科大学チューリッヒ校(スイス)
• [Journal Article] 信頼できるインフラストラクチャーの更新に向けて (2016)
  • Author(s): 阪野貴彦
  • Journal Title: 検査技術 Volume: 21 Pages: 1-4
• [Presentation] Accurate Depth-Map Refinement by Per-Pixel Plane Fitting for Stereo Vision (2016)
  • Author(s): M. Yokozuka, K. Tomita, O. Matsumoto, A. Banno
  • Organizer: International Conference on Pattern Recognition
  • Place of Presentation: メキシコ
  • Year and Date: 2016-12-07
  • Int'l Joint Research
• [Presentation] Keypoint extraction and matching for generalized image grid (2015)
  • Author(s): Atsuhiko Banno
  • Organizer: Joint research for robot beyond human capacity for emergency handling
  • Place of Presentation: ドイツ・カールスルーエ
  • Year and Date: 2015-03-20
• [Presentation] SFMによるつくばモビリティ実験特区の環境復元 (2014)
  • Author(s): 日本機械学会ロボティクス・メカトロニクス講演会
  • Organizer: 横塚 将志、阪野 貴彦、松本 治
  • Place of Presentation: 富山
  • Year and Date: 2014-05-26
• [Book] 機械工業年鑑 (2016)
  • Author(s): 「18．ロボット・メカトロニクス」
  • Total Pages: 1
  • Publisher: 日本機械学会