mirror of
https://github.com/introlab/rtabmap.git
synced 2026-09-02 01:20:25 +08:00
69ac21f811
* Adding GFTT and SIFT Cuda support * Working CudaSift * Disable cudasift option when not available * Added check to avoid re-allocating gpu memory everytime parseParameters is called. Added workaround of to detect/ignore invalid descriptors * Added SIFT/PreciseUpscale and SIFT/Upscale parameters. Adjusted max octave to behave more like opencv * Refactored how maximum features are thresholded, to be more similar to OpenCV version * Updated loop closure benchmark scripts * Cuda optical flow tmp commit * Added Stereo/Gpu Vis/CorFlowGpu parameters (optical flow gpu integration for F2F odom and stereo correspondences) * Fixed build without opencv cuda * Fixed build with Opencv 4.10 * ZED: updated parameters to match zed sdk 4 * MRPT requires C++17 * updated max octave limit CudaSift
To reproduce results (based on parameters of this paper):
rtabmap-console \
--Rtabmap/StatisticLogged true\
--Rtabmap/StatisticLoggedHeaders false\
--Kp/DetectorStrategy 0\
--Rtabmap/MemoryThr 300\
--Rtabmap/LoopRatio 0.9\
--SURF/HessianThreshold 150\
--Mem/STMSize 30\
--Vis/MaxFeatures 400\
--Kp/TfIdfLikelihoodUsed false\
--Kp/MaxFeatures 400\
--Kp/BadSignRatio 0.25\
--Mem/BadSignaturesIgnored true\
--Mem/RehearsalSimilarity 0.20\
--Mem/RecentWmRatio 0.2\
-gt "~/Downloads/UdeS_1Hz.png"\
~/Downloads/UdeS_1Hz
Adding the ground truth file here is optional to show recall at 100% precision at the end of the process directly without using the octave/MATLAB script below. For NewCollege and CityCentre datasets, rtabmap-imagesJoiner can be used to assemble the left and right images together.
To analyze with Octave/MATLAB, drop LogF.txt and LogI.txt generated files from command above in ShowLogs directly, then execute showLogs.m.