Open AccessParallel Implementations of the CSBP Stereo Vision Algorithm
Author(s) -
Lucas Veronese,
Lauro Jose Lyrio,
Jorcy de Oliveira Neto,
Avelino Forechi,
Claudine Badué,
Alberto F. De Souza
Publication year - 2011
Language(s) - English
Resource type - Conference proceedings
DOI - 10.5753/wscad.2011.17272
Subject(s) - cuda , computer science , implementation , stereopsis , pixel , computer vision , algorithm , parallel algorithm , artificial intelligence , parallel computing , programming language
We developed two parallel versions of the Constant Space Belief Propagation algorithm (CSBP - one of the best stereo algorithms currently known) [1]: one in OpenMP and one in C+CUDA. For images with 640x480 pixels, the sequential version has a performance of 1.16 frames per second (FPS), the OpenMP parallel version has a performance of 3.7 FPS, while the C+CUDA version has a performance of 17.3 FPS in high-performance desktop machines. These results are important because they enable the implementation of autonomous vehicles with sensors like camera, which is one of the objectives of a PRONEX project being currently developed in LCADDI/ UFES. One of the goals of this project is to implement an autonomous vehicle from a commercial automobile.