Authors:
Yutaro Yamanaka
1
;
Eichi Takaya
2
and
Satoshi Kurihara
2
Affiliations:
1
Faculty of Science and Technology, Keio University, Yokohama, Japan
;
2
Graduate School of Science and Technology, Keio University, Yokohama, Japan
Keyword(s):
Industrial Applications of AI, Tactile Tile, Ground Detection, Visual Impairment.
Abstract:
Tactile paving is a system used to help visually impaired individuals walk safely. However, it is difficult to recognize the surrounding tactile tiles on a first visit to an area. In this study, we propose a method for detecting tactile tiles integrated with ground detection using an RGB-Depth sensor. For the ground detection, we use the RANSAC algorithm and expand the region by using the breadth-first search. When detecting the tactile tiles, we perform thresholding and construct a model to identify candidate areas. Experimental results showed that the proposed method obtained a precision of about 83% in detecting tactile tiles on a paved asphalt road. It was possible to correctly distinguish tactile tiles from other objects by combining ground detection in many cases. On the other hand, there were many false detections of tactile tiles in challenging environments, and the processing speed should be improved for real-time navigation.