碩士 / 國立中央大學 / 資訊工程研究所 / 100 / Visually impaired people occupy 0.24% of the total population in Taiwan, they are the minority people who need to be looked after. The major challenge encountered by them is that of way finding, i.e., the ability of a person to find his way to a destination. There are a lot of literatures about way finding for visually impaired people, they all have pros and cons comparing with each other. In this thesis, a computer vision based algorithm is proposed, using the color information and the marker encoding method used in Augmented Reality, we design these color markers which can be placed in the environment. In practice, users used a smartphone to find the marker, and the phone will send a feedback to the users. A route can be built by several markers and saved in the database, when a user input his destination to the phone, it will conduct matching between the markers from the environment and the markers from the database so as to guide the user to their destination.
The color markers are the modification of the binary markers which are commonly used in Augmented Reality. To get the color pattern from the image stably, we use the information from YCbCr and HSV color spaces to get marker’s color region in the image. Then, we try to find contours in the binary image. For each contour, we check if it fit the requirement for the marker. If it does, we decode it to generate the encoded message. Finally, the encoded message is translated to way finding information and feedbacks are sent to the users.
Experimental results show that our color marker can be used in indoor and outdoor environments and can guide the users successfully to their destination. It reveals the validity and feasibility of the system in way finding for visually impaired people.
| Identifer | oai:union.ndltd.org:TW/100NCU05392060 |
| Date | January 2012 |
| Creators | Yi-Hsuan Chen, 陳逸軒 |
| Contributors | Kuo-Chin Fan, none, 范國清, 莊啟宏 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 83 |
Page generated in 0.2719 seconds