Closed qetuo105487900 closed 6 months ago
matlabbe
commented
9 months ago Yes, you can run without docker, but you will need to install cuda + pytorch (from source), then build rtabmap library with libtorch (for superpoint) and python (for superglue) support (cmake -DWITH_TORCH=ON -DWITH_PYTHON=ON ..).
qetuo105487900
commented
9 months ago i change CMakeList.txt in rtabmap such as this :
IF(WITH_TORCH)
set(Torch_DIR /home/lun/libtorch/share/cmake/Torch)
FIND_PACKAGE(Torch REQUIRED)
IF(TORCH_FOUND)
MESSAGE(STATUS "Found Torch: ${TORCH_INCLUDE_DIRS}")
ENDIF(TORCH_FOUND)
ENDIF(WITH_TORCH)and it find !!!
but when i run
roslaunch rtabmap_launch rtabmap.launch args:="-d
--SuperPoint/ModelPath /home/lun/rtabmap/archive/2022-IlluminationInvariant/scripts/superpoint_v1.pt
--SuperGlue/Path /home/lun/rtabmap/archive/2022-IlluminationInvariant/scripts/SuperGluePretrainedNetwork/demo_superglue.py
--Reg/RepeatOnce false
--Vis/CorGuessWinSize 0
--Kp/DetectorStrategy 11
--Vis/FeatureType 11
--Mem/UseOdomFeatures false
--Vis/CorNNType 6"
approx_sync:=falseit still display
Features2d.cpp:594::create() SupertPoint Torch feature cannot be used as RTAB-Map is not built with the option enabled. GFTT/ORB is used instead.WTF?!!
qetuo105487900
commented
9 months ago cmake -DWITH_TORCH=ON -DWITH_PYTHON=ON ..
-- PCL definitions don't contain "-march=native", make sure all libraries using Eigen are also compiled without that flag to avoid some segmentation faults (with gdb referring to some Eigen functions).
-- Found Sqlite3: /usr/include /usr/lib/x86_64-linux-gnu/libsqlite3.so
-- Found OpenMP: /usr/lib/gcc/x86_64-linux-gnu/9/libgomp.so;/usr/lib/x86_64-linux-gnu/libpthread.so
-- Add PCL_OMP to definitions
-- Found OpenCV: /usr/include/opencv4
-- Found PCL: /usr/include/pcl-1.10;/usr/include/eigen3;/usr/include;/usr/include/ni;/usr/include/vtk-7.1;/usr/include/freetype2;/usr/include/x86_64-linux-gnu
-- Found ZLIB: /usr/include
-- VTK_RENDERING_BACKEND=OpenGL2
-- Caffe2: CUDA detected: 12.2
-- Caffe2: CUDA nvcc is: /usr/local/cuda/bin/nvcc
-- Caffe2: CUDA toolkit directory: /usr/local/cuda
-- Caffe2: Header version is: 12.2
-- /usr/local/cuda/lib64/libnvrtc.so shorthash is 79a3de5a
-- USE_CUDNN is set to 0. Compiling without cuDNN support
-- USE_CUSPARSELT is set to 0. Compiling without cuSPARSELt support
-- Autodetected CUDA architecture(s): 8.6
-- Added CUDA NVCC flags for: -gencode;arch=compute_86,code=sm_86
-- Found Torch: /home/lun/libtorch/include;/home/lun/libtorch/include/torch/csrc/api/include
-- Found Python3
-- Found pybind11: /opt/ros/noetic/include (found version "2.10.3")
-- Found OpenNI2: /usr/include/openni2
-- Found DC1394: /usr/include/dc1394
-- Found g2o (targets)
-- g2o include dir: /opt/ros/noetic/include
-- Old g2o factory version detected without shared ptr (factory file: /opt/ros/noetic/include/g2o/core/factory.h).
-- GTSAM include directory: /opt/ros/noetic/lib/cmake/GTSAM/../../../include
-- FlyCapture2_INCLUDE_DIR=FlyCapture2_INCLUDE_DIR-NOTFOUND
-- FlyCapture2_LIBRARY=FlyCapture2_LIBRARY-NOTFOUND
-- Triclops_INCLUDE_DIR=Triclops_INCLUDE_DIR-NOTFOUND
-- Triclops_LIBRARY=Triclops_LIBRARY-NOTFOUND
-- FlyCaptureBridge_LIBRARY=FlyCaptureBridge_LIBRARY-NOTFOUND
-- Found Boost: /usr/include (found version "1.71.0") found components: thread filesystem system program_options date_time chrono atomic
-- Found libpointmatcher: /opt/ros/noetic/include;/usr/include/eigen3
-- Found Boost: /usr/include (found version "1.71.0") found components: thread filesystem system program_options date_time chrono timer serialization atomic
-- Found RealSense2:
-- Found octomap 1.9.8: /opt/ros/noetic/include
-- Found grid_map_core 1.6.4: /opt/ros/noetic/include;/usr/include/eigen3
-- Architecture: x86_64
-- Found Pthreads
-- --------------------------------------------
-- Info :
-- RTAB-Map Version = 0.21.4
-- CMAKE_VERSION = 3.28.3
-- CMAKE_INSTALL_PREFIX = /usr/local
-- CMAKE_BUILD_TYPE = Release
-- CMAKE_INSTALL_LIBDIR = lib
-- BUILD_APP = ON
-- BUILD_TOOLS = ON
-- BUILD_EXAMPLES = ON
-- BUILD_SHARED_LIBS = ON
-- CMAKE_CXX_FLAGS = -fmessage-length=0 -fopenmp -std=c++17
-- FLANN_KDTREE_MEM_OPT = OFF
-- PCL_DEFINITIONS = -DDISABLE_OPENNI2;-DDISABLE_PCAP;-DDISABLE_PNG;-DDISABLE_LIBUSB_1_0
-- PCL_VERSION = 1.10.0
--
-- Optional dependencies ('*' affects some default parameters) :
-- *With OpenCV 4.2.0 xfeatures2d = NO, nonfree = NO (License: BSD)
-- With Qt 5.12.8 = YES (License: Open Source or Commercial)
-- With VTK 7.1 = YES (License: BSD)
-- With external SQLite3 = YES (License: Public Domain)
-- With ORB OcTree = YES (License: GPLv3)
-- With SupertPoint = YES (License: GPLv3) libtorch=2.2.0
-- With Python3.8 = YES (License: PSF)
-- With Madgwick = YES (License: GPL)
-- With FastCV = NO (FastCV not found)
-- With PDAL = NO (PDAL not found)
--
-- Solvers:
-- With TORO = YES (License: Creative Commons [Attribution-NonCommercial-ShareAlike])
-- *With g2o 1.0.0 = YES (License: BSD)
-- *With GTSAM 4.2.0 = YES (License: BSD)
-- *With Ceres = NO (WITH_CERES=OFF)
-- With MRPT = NO (MRPT not found)
-- With VERTIGO = YES (License: GPLv3)
-- With cvsba = NO (WITH_CVSBA=OFF)
-- *With libpointmatcher 1.3.1 = YES (License: BSD)
-- With CCCoreLib = NO (WITH_CCCORELIB=OFF)
-- With Open3D = NO (WITH_OPEN3D=OFF)
-- With OpenGV = NO (OpenGV not found)
--
-- Reconstruction Approaches:
-- With OctoMap 1.9.8 = YES (License: BSD)
-- With GridMap 1.6.4 = YES (License: BSD)
-- With CPUTSDF = NO (WITH_CPUTSDF=OFF)
-- With OpenChisel = NO (WITH_OPENCHISEL=OFF)
-- With AliceVision = NO (WITH_ALICE_VISION=OFF)
--
-- Camera Drivers:
-- With Freenect = NO (libfreenect not found)
-- With OpenNI2 = YES (License: Apache v2)
-- With Freenect2 = NO (libfreenect2 not found)
-- With Kinect for Windows 2 = NO (Kinect for Windows 2 SDK not found)
-- With Kinect for Azure = NO (Kinect for Azure SDK not found)
-- With dc1394 = YES (License: LGPL)
-- With FlyCapture2/Triclops = NO (Point Grey SDK not found)
-- With ZED = NO (ZED sdk and/or cuda not found)
-- With ZEDOC = NO (ZED Open Capture not found)
-- With RealSense = NO (librealsense not found)
-- With RealSense2 2.50.0 = YES (License: Apache-2)
-- With MyntEyeS = NO (mynteye s sdk not found)
-- With DepthAI = NO (WITH_DEPTHAI=OFF)
--
-- Odometry Approaches:
-- With loam_velodyne = NO (WITH_LOAM=OFF)
-- With floam = NO (WITH_FLOAM=OFF)
-- With libfovis = NO (WITH_FOVIS=OFF)
-- With libviso2 = NO (WITH_VISO2=OFF)
-- With dvo_core = NO (WITH_DVO=OFF)
-- With okvis = NO (WITH_OKVIS=OFF)
-- With msckf_vio = NO (WITH_MSCKF_VIO=OFF)
-- With VINS-Fusion = NO (WITH_VINS=OFF)
-- With OpenVINS = NO (WITH_OPENVINS=OFF)
-- With ORB_SLAM = NO (WITH_ORB_SLAM=OFF)
-- Show all options with: cmake -LA | grep WITH_
-- --------------------------------------------
-- Configuring done (2.7s)
-- Generating done (0.4s)
-- Build files have been written to: /home/lun/rtabmap/buildrtabmap --version
without everything that i had do
RTAB-Map: 0.21.3
PCL: 1.10.0
With VTK: 7.1.1
OpenCV: 4.2.0
With OpenCV xfeatures2d: false
With OpenCV nonfree: false
With ORB OcTree: true
With SuperPoint Torch: false
With Python3: false
With FastCV: false
With OpenGV: false
With Madgwick: true
With PDAL: false
With TORO: true
With g2o: true
With GTSAM: true
With Vertigo: true
With CVSBA: false
With Ceres: false
With OpenNI2: true
With Freenect: false
With Freenect2: false
With K4W2: false
With K4A: false
With DC1394: true
With FlyCapture2: false
With ZED: false
With ZED Open Capture: false
With RealSense: false
With RealSense SLAM: false
With RealSense2: false
With MYNT EYE S: false
With DepthAI: false
With libpointmatcher: true
With CCCoreLib: false
With OctoMap: true
With GridMap: true
With cpu-tsdf: false
With open chisel: false
With Alice Vision: false
With LOAM: false
With FLOAM: false
With FOVIS: false
With Viso2: false
With DVO: false
With ORB_SLAM: false
With OKVIS: false
With MSCKF_VIO: false
With VINS-Fusion: false
With OpenVINS: falseYou can also do:
cmake -DWITH_TORCH=ON -DWITH_PYTHON=ON -DTorch_DIR=/home/lun/libtorch/share/cmake/Torch ..to avoid hard-coding it in the CMakeLists.txt.
The cmake status shows that torch is detected, so if rtabmap builds succesfully, it should have it.
After build is complted, you must install sudo make install. It seems rtabmap --version is not your version you just built, but the one from the binaries. I suggest also to uninstall the binaries, as you will also need to rebuild rtabmap_ros from source:
# uninstall binaries
sudo apt remove ros-$ROS_DISTRO-rtabmap*
# install rtabmap with torch support
cd rtabmap/build
cmake -DWITH_TORCH=ON -DWITH_PYTHON=ON -DTorch_DIR=/home/lun/libtorch/share/cmake/Torch ..
make -j6
sudo make install
# install rtabmap_ros with local rtabmap version:
cd ~/catkin_ws
git clone https://github.com/introlab/rtabmap_ros.git src/rtabmap_ros
catkin_make -j4
sorry to bother everyone, Superglue can run rtabslam without docker ? as follow step is correct?
roslaunch realsense2_camera rs_camera.launch align_depth:=true
roslaunch rtabmap_launch rtabmap.launch args:="-d \ --SuperPoint/ModelPath /home/lun/rtabmap/archive/2022-IlluminationInvariant/scripts/superpoint_v1.pt\ --SuperGlue/Path /home/lun/rtabmap/archive/2022-IlluminationInvariant/scripts/SuperGluePretrainedNetwork/demo_superglue.py \ --Reg/RepeatOnce false \ --Vis/CorGuessWinSize 0 \ --Kp/DetectorStrategy 11 \ --Vis/FeatureType 11 \ --Mem/UseOdomFeatures false \ --Vis/CorNNType 6" \ approx_sync:=false
i get this : QAQ
Features2d.cpp:594::create() SupertPoint Torch feature cannot be used as RTAB-Map is not built with the option enabled. GFTT/ORB is used instead.