Submitted by Heiko Hirschmueller.

Submission data

Full nameSemi-Global Matching (GPU) (RVC 2020 baseline submission)
DescriptionReimplementation of SGM on the GPU. Census was used as matching cost. Additionally to disparity, a confidence and error is computed for each pixel. This is included in the runtime.

This is the implementation used on-board the Roboception rc_visard stereo sensors.
segmentation filter=80 pixels
median filter=5x5
Publication titleStereo processing by semiglobal matching and mutual information
Publication authorsHeiko Hirschmueller
Publication venueTPAMI 2008, Volume 30(2), pp. 328-341
Publication URL
Programming language(s)C++ with CUDA
HardwareGeForce GTX 980
Submission creation date10 Mar, 2018
Last edited10 Mar, 2018

High-res multi-view results

multi-view co.statueterrac.
No results yet.

Low-res many-view results

indooroutdoorlakesidesand boxstorage roomstorage room 2tunnel
No results yet.

Low-res two-view results

Infoalllakes. 1llakes. 1ssand box 1lsand box 1sstora. room 1lstora. room 1sstora. room 2lstora. room 2sstora. room 2 1lstora. room 2 1sstora. room 2 2lstora. room 2 2sstora. room 3lstora. room 3stunnel 1ltunnel 1stunnel 2ltunnel 2stunnel 3ltunnel 3s
binarytwo views10.080.603.422.300.3219.416.3318.9514.6425.1424.3233.3418.7919.8612.550.

SLAM results

allboxesboxes darkbuddhacables 4cables 5desk 1desk 2desk changing 2desk dark 1desk dark 2desk global light changesdesk ir lightdinodroneforeground occlusionhelmetkidnap 2lamplarge loop 2large loop 3large non loopmotion 2motion 3motion 4planar 1reflective 2scale changetable 1table 2table 5table 6table global light changestable local light changestable scenetrashbin
No results yet.