Home // INFOCOMP 2013, The Third International Conference on Advanced Communications and Computation // View article

A Near Real-Time Color Stereo Matching Method for GPU

Authors:
Naiyu Zhang
Hongjian Wang
Jean-Charles Créput
Julien Moreau
Yassine Ruichek

Keywords: GPU; Real-Time Stereovision; SAD-ALD; Adaptive Window; CUDA

Abstract:
This paper presents a near real-time stereo matching method with acceptable matching results. This method consists of three important steps: SAD-ALD cost measure, cost aggregation in adaptive window in cross-based support regions and a refinement step. These three steps are well organized to be adopted by the GPU's parallel architecture. The parallelism brought by GPU and CUDA implementations provides significant acceleration in running time. This method is tested on six pairs of images from Middlebury dataset, each possibly declined within different sizes. For each pair of images it can generate acceptable matching results in roughly less than 100 milliseconds. The method is also compared with three GPU-based methods and one CPU-based method on increasing size image pairs.

Pages: 27 to 32

Copyright: Copyright (c) IARIA, 2013

Publication date: November 17, 2013

Published in: conference

ISSN: 2308-3484

ISBN: 978-1-61208-310-0

Location: Lisbon, Portugal

Dates: from November 17, 2013 to November 21, 2013