ORB-SLAM3 for Ubuntu 24.04 LTS

This repository contains ORB-SLAM3 with fixes and modifications for compatibility with Ubuntu 24.04 LTS (Noble Numbat).

System Requirements

• OS: Ubuntu 24.04 LTS
• Compiler: GCC 13.3.0, G++ 13.3.0
• CMake: 3.28.3 or higher
• Memory: At least 4GB RAM recommended
• Storage: ~2GB for dependencies and build files

Prerequisites

1. Install Build Tools

sudo apt update && sudo apt upgrade -y
sudo apt install -y build-essential cmake git pkg-config

2. Install Core Dependencies

# Eigen3 (Linear algebra library)
sudo apt install -y libeigen3-dev

# OpenCV 4.6.0 (Computer vision library)
sudo apt install -y libopencv-dev

# Pangolin dependencies
sudo apt install -y libglfw3-dev libgl1-mesa-dev libglu1-mesa-dev libepoxy-dev

3. Install Optional Dependencies

For RealSense Camera Support

If you plan to use Intel RealSense cameras, build librealsense from source:

# Remove apt version if installed
sudo apt remove librealsense2-dev librealsense2-utils

# Install build dependencies
sudo apt install -y libusb-1.0-0-dev libgtk-3-dev

# Clone and build librealsense
cd ~
git clone https://github.com/IntelRealSense/librealsense.git
cd librealsense
git checkout v2.55.1

mkdir build && cd build
cmake .. \
    -DCMAKE_BUILD_TYPE=Release \
    -DBUILD_WITH_DDS=false \
    -DBUILD_EXAMPLES=true \
    -DBUILD_GRAPHICAL_EXAMPLES=true

make -j$(nproc)
sudo make install
sudo ldconfig

# Setup udev rules
sudo cp ../config/99-realsense-libusb.rules /etc/udev/rules.d/
sudo udevadm control --reload-rules && sudo udevadm trigger

Checking Camera Device Mapping

To identify whether the connected camera is the internal webcam or the external USB camera, run:

ls /dev/video*

This shows all available video devices.

To identify which /dev/videoX belongs to which camera:

v4l2-ctl --list-devices

Example output:

icspring camera: icspring camer (usb-0000:00:14.0-3):
    /dev/video2
    /dev/video3
    /dev/media1

Integrated_Webcam_HD: Integrate (usb-0000:00:14.0-5):
    /dev/video0
    /dev/video1
    /dev/media0

From this output:

• /dev/video0 and /dev/video1 correspond to the internal laptop webcam
• /dev/video2 and /dev/video3 correspond to the external USB camera

The ORB-SLAM3 mono_webcam.cc file should use the correct device index:

cv::VideoCapture cap(2, cv::CAP_V4L2);

where 2 corresponds to /dev/video2.

Installation

1. Setup Directory Structure

# Create workspace
cd ~
mkdir orb-slam3-root && cd orb-slam3-root

# Set environment variable (add to ~/.bashrc for persistence)
export ORB_SLAM3_ROOT_PATH=~/orb-slam3-root
echo "export ORB_SLAM3_ROOT_PATH=~/orb-slam3-root" >> ~/.bashrc

2. Build Pangolin

cd $ORB_SLAM3_ROOT_PATH

# Clone Pangolin
git clone https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin
git checkout v0.9.2

# Build
mkdir build && cd build
cmake ..
make -j$(nproc)

Test Pangolin (optional):

$ORB_SLAM3_ROOT_PATH/Pangolin/build/examples/HelloPangolin/HelloPangolin

3. Clone This Repository

cd $ORB_SLAM3_ROOT_PATH
git clone  "https://github.com/irocu26/ORB-SLAM3.git"
cd ORB-SLAM3

4. Build ORB-SLAM3

chmod +x build.sh
./build.sh

The build process will:

• Compile DBoW2 (vocabulary management)
• Compile g2o (graph optimization)
• Compile Sophus (Lie algebra)
• Uncompress vocabulary file
• Build ORB-SLAM3 library and examples

5. Install Sophus Globally

This step is required if you plan to use ORB-SLAM3 with ROS2:

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3/Thirdparty/Sophus/build
sudo make install

To uninstall later:

sudo rm -rf /usr/local/include/sophus
sudo rm -rf /usr/local/share/sophus

Testing

Option 1: Test with Webcam

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3
./Examples/Monocular/mono_webcam \
Vocabulary/ORBvoc.txt \
Examples/Monocular/Webcam.yaml

Move your camera around. You should see feature tracking and map building in real-time.

Option 2: Test with EuRoC Dataset

Download the MH_01_easy sequence:

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3
mkdir datasets && cd datasets

# Download from ETH Research Collection
wget https://www.research-collection.ethz.ch/bitstream/handle/20.500.11850/263889/1/MH_01_easy.zip

# Extract
unzip MH_01_easy.zip
mkdir MH01
mv mav0 MH01/mav0
cd ..

# Run test
./Examples/Monocular/mono_euroc \
    Vocabulary/ORBvoc.txt \
    Examples/Monocular/EuRoC.yaml \
    datasets/MH01 \
    Examples/Monocular/EuRoC_TimeStamps/MH01.txt \
    dataset-MH01_mono

Option 3: Test with TUM RGB-D Dataset

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3/datasets
mkdir TUM && cd TUM

# Download a sequence
wget https://cvg.cit.tum.de/rgbd/dataset/freiburg1/rgbd_dataset_freiburg1_xyz.tgz
tar -xvf rgbd_dataset_freiburg1_xyz.tgz

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3

# Run RGB-D example
./Examples/RGB-D/rgbd_tum \
    Vocabulary/ORBvoc.txt \
    Examples/RGB-D/TUM1.yaml \
    datasets/TUM/rgbd_dataset_freiburg1_xyz \
    datasets/TUM/associations/fr1_xyz.txt

Available Examples

After successful build, you'll find these executables in Examples/:

Monocular

• mono_euroc - EuRoC MAV dataset
• mono_kitti - KITTI dataset
• mono_tum - TUM monocular
• mono_tum_vi - TUM VI dataset

Monocular-Inertial

• mono_inertial_euroc - EuRoC with IMU
• mono_inertial_tum_vi - TUM VI with IMU

Stereo

• stereo_euroc - EuRoC stereo
• stereo_kitti - KITTI stereo
• stereo_tum_vi - TUM VI stereo

Stereo-Inertial

• stereo_inertial_euroc - EuRoC stereo with IMU
• stereo_inertial_tum_vi - TUM VI stereo with IMU

RGB-D

• rgbd_tum - TUM RGB-D dataset

Key Modifications for Ubuntu 24.04

This repository includes the following fixes for Ubuntu 24.04 compatibility:

1. CMakeLists.txt (line 33):

   find_package(OpenCV 4.6)  # Matches Ubuntu 24.04 default

2. CMakeLists.txt (line 41):

   find_package(Eigen3 REQUIRED)  # Version number removed

3. ThirdParty/DBoW2/CMakeLists.txt (line 32):

   find_package(OpenCV 4.6 QUIET)

4. ThirdParty/Sophus/CMakeLists.txt (line 23):

   SET(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Werror -Wextra -std=c++11 -Wno-error=array-bounds -Wno-deprecated-declarations -ftemplate-backtrace-limit=0")

5. ThirdParty/Sophus/CMakeLists.txt (line 35):

   find_package(Eigen3 REQUIRED)  # Version number removed

6. Examples/Monocular/mono_euroc.cc (line 83):

   ORB_SLAM3::System SLAM(argv[1],argv[2],ORB_SLAM3::System::MONOCULAR, true);
   // Viewer enabled by default

Environment Variables

Add these to your ~/.bashrc for convenience:

# ORB-SLAM3 root path
export ORB_SLAM3_ROOT_PATH=~/orb-slam3-root

# Add library paths (if needed)
export LD_LIBRARY_PATH="$ORB_SLAM3_ROOT_PATH/ORB-SLAM3/lib:$LD_LIBRARY_PATH"
export LD_LIBRARY_PATH="$ORB_SLAM3_ROOT_PATH/Pangolin/build:$LD_LIBRARY_PATH"

Troubleshooting

Build Errors

Problem: error: 'array' was not declared in this scope Solution: Already fixed in this repository. Ensure you're using the modified Sophus CMakeLists.txt.

Problem: OpenCV version mismatch Solution:

# Check your OpenCV version
pkg-config --modversion opencv4

# Should be 4.6.0 on Ubuntu 24.04

Problem: RealSense warnings about fastcdr/fastrtps Solution: This is harmless if you don't use RealSense. To fix completely, rebuild librealsense with -DBUILD_WITH_DDS=false (see Prerequisites section).

Runtime Errors

Problem: Segmentation fault when loading vocabulary Solution: Ensure you've extracted the vocabulary:

cd $ORB_SLAM3_ROOT_PATH/ORB-SLAM3/Vocabulary
tar -xf ORBvoc.txt.tar.gz

Problem: Camera not detected Solution:

# Check camera device
ls -l /dev/video*

# Test with v4l2
v4l2-ctl --list-devices

Problem: Black screen/no tracking Solution:

• Ensure adequate lighting
• Move camera slowly at first
• Point at textured surfaces (not blank walls)
• Check that images are arriving (press 'v' in viewer to toggle views)

Performance Tips

1. CPU Usage: ORB-SLAM3 is CPU-intensive. For best performance:

   • Close unnecessary applications
   • Use make -j$(nproc) to utilize all cores during build

2. Memory: The vocabulary file loads into RAM (~80MB). Ensure sufficient memory.

3. Real-time Performance:

   • Monocular: 30 FPS on modern CPUs
   • Stereo: 20-30 FPS
   • RGB-D: 30 FPS

Citation

If you use ORB-SLAM3 in your research, please cite:

@article{ORBSLAM3_TRO,
  title={{ORB-SLAM3}: An Accurate Open-Source Library for Visual, Visual-Inertial 
         and Multi-Map {SLAM}},
  author={Campos, Carlos AND Elvira, Richard AND Rodriguez, Juan J. G\'omez AND 
          Montiel, Jos\'e M. M. AND Tard\'os, Juan D.},
  journal={IEEE Transactions on Robotics}, 
  volume={37},
  number={6},
  pages={1874-1890},
  year={2021}
}

License

ORB-SLAM3 is released under GPLv3 license.

Acknowledgments

• Original ORB-SLAM3 by Carlos Campos et al.
• Ubuntu 24.04 compatibility fixes
• Community contributions for stereo fixes

Support

For issues specific to Ubuntu 24.04 compatibility, please open an issue in this repository.

For general ORB-SLAM3 questions, refer to the original repository.

Related Resources

• ORB-SLAM3 Paper
• Pangolin Documentation
• EuRoC Dataset
• TUM RGB-D Dataset