Developing localization and navigation assistance tools for visually impaired people havebeen received many intention in the autonomous robotics community [6]. Most of the worksfocus on finding efficient localization solutions based on positioning data from differentsensors such as GPS, laser, Radio Frequency Identification (RFID), vision or the fusion ofseveral of them. Loomis et al. [17] surveyed efficiency of GPS-based navigation systemssupporting visually impaired people. The GPS-based systems share similar problems: lowaccuracy in urban-environments (localization accuracy is limited to approximately 20 m),signal loss due to multi-path effect or line-of-sight restrictions due to the presence of buildingsor even foliage. Kulyukin et al. [15] proposed a system based on Radio Frequency Identification (RFID) for aiding the navigation of visually impaired people in indoor environments.The system requires the design of a dense network of location identifiers. Helal et al. [13]proposed a wireless pedestrian navigation system. They integrated several signals such asvoiced, wireless networks, Geographic Information System (GIS) and GPS to provide thevisually impaired people with an optimized route. Recent advanced techniques in computervision offer substantial improvements with respect to localization and navigation services inknown or unknown environments. The vision-based approaches offer not only safe navigation,nhưng cũng cung cấp một mô tả rất phong phú và có giá trị của môi trường. Ví dụ, [2]phát triển một ứng dụng được đặt theo tên LocateIt, giúp người mù, xác định vị trí các đối tượng trong hồmôi trường.
đang được dịch, vui lòng đợi..
