Accurate real-time stereo correspondence using intra- and inter-scanline optimization
Institute of Information and Communication Engineering, Zhejiang University, Hangzhou 310027, China
Journal of Zhejiang University SCIENCE C, 2012
@article{yao2012accurate,
title={Accurate real-time stereo correspondence using intra-and inter-scanline optimization},
author={YAO, L. and LI, D. and ZHANG, J. and WANG, L. and ZHANG, M.},
year={2012}
}
This paper deals with a novel stereo algorithm that can generate accurate dense disparity maps at real-time rate. The algorithm employs an effective cross-based variable support aggregation strategy within a scanline optimization framework. Rather than matching intensities directly, the use of adaptive support aggregation allows for precisely handling the weak textured regions as well as depth discontinuities. In order to improve the disparity results with global reasoning, we reformulate the energy function on a tree structure over whole 2D image area, as opposed to dynamic programming of individual scanlines. By applying both intra-scanline and inter-scanline optimization, the algorithm reduces the typical "streaking" artifact while maintaining high computational efficiency. The experimental results are evaluated on the Middlebury stereo dataset, which shows that our approach is among the best for all real-time approaches. We implement the algorithm on a commodity graphics card with CUDA architecture, running at about 35 fames per second for a typical stereo pair with resolution of 384×288 and 16 disparity levels.
February 15, 2012 by hgpu
