Dip based indoor guidance system for blind Person

Abstract

DIP based Indoor guidance system for Blind Person This Projects aims at the Development of a portable device that will help the blind users to navigate in a building independently by the provision of appropriate guidance and instructions. Pictures taken by the user through the device, will be matched with an already stored database of images, the best match found will dictate the corresponding audio instructions about the user’s current location in the building. The device is now capable of finding the best matched image from the database and guide the blind person by generating the appropriate sound according to the location of blind person. Picture taken from the camera which is interfaced with raspberry pi 3 is matched with the images of database and guide the blind person. This is possible by using SURF algorithm and FLANN index. The key points are extracted by the SURF algorithm of the test image and FLANN index make the index of descriptor vectors and match the key points of test image with the images already stored in the database one by one and generating the audio commands for every possible match that could occur during the process. These Audio Commands will contain the instructions that will update the user about his/her current precise location in the building.