2A2-J02 全方位画像を用いたロボット姿勢検出のための特徴抽出手法の開発(スワームロボティクス)

抄録

In this paper, the authors present the development of a method for detecting robust local features for the change of environments and the solution of an occlusion problem based on the feature quantity extraction technique called features from accelerated segment test (FAST) for an omnidirectional image taken indoor and outdoor using panoramic annular lens. This method can solve the above problems by detecting a set of feature points on edges of a ground surface captured by a visual sensor. In addition, the authors have verified how accurately the edges can be found by the proposed method. As a result, the proposed method is effective to extract an arc as the feature point, but it was not effective against edges of the ground surface detection.