Vol.29, No.1, 2021, pp.247-258, doi:10.32604/iasc.2021.015274
Implementation of Multi-Object Recognition System for the Blind
  • Huijin Park, Soobin Ou, Jongwoo Lee*
Department of IT Engineering, Sookmyung Women’s University, Seoul, 04310, Korea
* Corresponding Author: Jongwoo Lee. Email:
(This article belongs to this Special Issue: Machine Learning and Deep Learning for Transportation)
Received 13 November 2020; Accepted 23 March 2021; Issue published 12 May 2021
Blind people are highly exposed to numerous dangers when they walk alone outside as they cannot obtain sufficient information about their surroundings. While proceeding along a crosswalk, acoustic signals are played, though such signals are often faulty or difficult to hear. The bollards can also be dangerous if they are not made with flexible materials or are located improperly. Therefore, since the blind cannot detect proper information about these obstacles while walking, their environment can prove to be dangerous. In this paper, we propose an object recognition system that allows the blind to walk safely outdoors. The proposed system can recognize obstacles and other objects through a real-time video stream and a sensor system, and provides the recognition results to the blind via a voice output. The system is able to figure out the current state of a pedestrian signal, the position of a bollard, and the direction of a tactile paving all at the same time using an object recognition module. In addition, its sensor determines whether there is an obstacle near the blind at a specific distance. We built a prototype of the object recognition system using a Raspberry Pi module, and then evaluated it with an experiment created for testing purposes, in which the system drives a programmable remote-control car. The experiment results showed that our object recognition system succeeds in detecting the obstacles and taking a safer route in order to avoid them.
blind; obstacles; object detection; sensor; Raspberry Pi
Cite This Article
H. Park, S. Ou and J. Lee, "Implementation of multi-object recognition system for the blind," Intelligent Automation & Soft Computing, vol. 29, no.1, pp. 247–258, 2021.
This work is licensed under a Creative Commons Attribution 4.0 International License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.