Low-cost stereo vision system for supporting the visually impaired's walk

Authors:
Thanathip Limna;Pichaya Tandayya;Nikom Suvanvorn
Affiliations:
Prince of Songkla University, Hat Yai, Songkhla, Thailand;Prince of Songkla University, Hat Yai, Songkhla, Thailand;Prince of Songkla University, Hat Yai, Songkhla, Thailand
Venue:
Proceedings of the 3rd International Convention on Rehabilitation Engineering & Assistive Technology
Year:
2009

Citing 5
Cited 0

A Pixel Dissimilarity Measure That Is Insensitive to Image Sampling

IEEE Transactions on Pattern Analysis and Machine Intelligence
Parallel Low-Level Image Processing on a Distributed-Memory System

IPDPS '00 Proceedings of the 15 IPDPS 2000 Workshops on Parallel and Distributed Processing
Drishti: An Integrated Navigation System for Visually Impaired and Disabled

ISWC '01 Proceedings of the 5th IEEE International Symposium on Wearable Computers
Depth Discontinuities by Pixel-to-Pixel Stereo

ICCV '98 Proceedings of the Sixth International Conference on Computer Vision
A Vision System for Providing 3D Perception of the Environment via Transcutaneous Electro-Neural Stimulation

IV '04 Proceedings of the Information Visualisation, Eighth International Conference

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an obstacle detection system for the visually impaired to use with a walking stick. The system in this work can find obstacles applying the depth discontinuities Pixel-to-Pixel (P2P) stereo algorithm which is one of the Intensity-based Stereo Matching (ISM) techniques in stereo vision that can find objects with featureless surface such as whiteboard, door, etc. However, the ISM technique is time consuming and not fast enough for real-time usage. Applying parallel computing using Message Passing Interface (MPI) helps reduce the computing time and enables realtime usage. In this work, we estimate the distance between objects and the visually impaired by applying the V-disparity. Our system can accurately detect objects within the range of 5 meters using 12-centimeter based-line low-cost webcams.