Navigation system for visually impaired

Abstract

Object recognition and navigation are highly challengeable for visually impaired even though the environment is familiar for him. There are different approaches for addressing this problem domain using technologies such as Global Positioning System (GPS) technology, Radio Frequency Identification (RFID) sensor technology etc. Although existing systems which use these different technologies for addressing this approach there are still limitations and issues due to usability and cost effectiveness of them to be addressed. The proposed system enables the blind person to directly interact with a mobile application and request the needed object as a voice command. Then it will process through image recognition and optimal path finding modules and identify the correct object and will guide the blind person to navigate to it. The system use Scale Invariant Feature Transformation (SIFT) algorithm for track each and every object in the assumed indoor environment. And then using cell based map identify the object location, user location and provide navigate information to the user for navigating to that object optimally. This solution has been evaluated using several experimental scenarios. Although the results are stilI not enough to provide strong conclusions, they indicate that the system is suitable for visually impaired people as object recognition and navigation system.