Implementation and Evaluation of Obstacle Recognition System for the Blind

Abstract

The visually impaired are exposed to many dangers while walking because of the lack of information on the various obstacles. In this paper, we proposed an obstacle recognition system that provides voice guidance in real-time for the visually impaired. The system informs on obstacles and finds locations of the obstacles such as traffic lights, bollards, and braille blocks. We developed an obstacle recognition system using Raspberry Pi and evaluated its performance by applying the system to a programmable remote control car. The performance evaluation proves that the system successfully avoids the obstacles and guide safely.