ug_stereomatcher

NEW Our matcher now works with CUDA 8.0 and ROS kinetic!

Authors (see package.xml for contact info): Paul Cockshott, Tian Xu, Gerardo Aragon-Camarasa, Susanne Oehler, Mozhgan Chimeh and J. Paul Siebert

SEE LICENSE BEFORE USING THIS SOFTWARE

A tutorial can be found at https://github.com/gerac83/ug_stereomatcher/wiki

The ug_stereomatcher ROS package comprises the University of Glasgow GPU stereo matcher and nodes to compute point clouds from disparity maps for each operational mode as described below.

The matcher has two operational modes:

1. Compute disparity maps over a image pyramid at full resolution in 10 seconds on 16MP RGB images
2. Compute foveated disparity maps for each level on the pyramid in 3 seconds on 16MP RGB images with a fixed fovea size of 615 by 407

NOTE: You need a CUDA capable graphics card in order to compile and run the software. For 16MP images, you need at least 2GB of graphics card RAM. This code has been optimised for the NVIDIA Geforce GTX 750i, 970 and 1080.

Publications

If you use this ROS package please cite the following paper(s):

• Cockshott, P., Oehler, S., Xu, T., Siebert, J. P., and Aragon-Camarasa, G., 2012 “Parallel stereo vision algorithm” in: Many-Core Applications Research Community Symposium 2012, 29-30 Nov 2012, Aachen, Germany.

The ROS package has been used and demonstrated in the following papers:

• Sun, L., Aragon-Camarasa, G. and Siebert, J. P., 2014. “Improving Cloth Manipulation by Modelling Wrinkles” in: 2014 IEEE International Conference on Robotics and Automation (ICRA 2014), Advances in robot manipulation of clothes and flexible objects Workshop. June 1, 2014. Hong Kong, China.
• Sun, L., Aragon-Camarasa, G., Siebert, J.P. and Rogers, S., 2013 “A Heuristic-Based Approach for Flattening Wrinkled Clothes,” in ‘Towards Autonomous Robotic Systems, TAROS 2013’, University of Oxford, 20-30 August 2013.
• Sun, L., Aragon-Camarasa, G. Rogers, S., Siebert, J.P., 2015 "Accurate Garment Surface Analysis using an Active Stereo Robot Head with Application to Dual-Arm Flattening," in ICRA 2015 (Accepted)

Database of stereo images

For benchmarking, we have produced a database of 80 stereo RGB images; it can be downloaded for free at:

https://sites.google.com/site/ugstereodatabase/

We have produced a new database comprising 50 clothing items from 5 categories: t-shits, shirts, sweaters, jeans and towels. Each item of clothing is captured over 21 different random configurations, resulting in a total of 1050 configurations. This database can be downloaded for free at:

http://researchdata.gla.ac.uk/270/

Companion technical report:

http://arxiv.org/abs/1311.7295