Skip to main content

​
System Requirements

ORB-SLAM3 has been tested on Ubuntu 16.04 and 18.04, but should be easy to compile on other Linux platforms.
A powerful computer (e.g., Intel i7 processor) is recommended to ensure real-time performance and more stable, accurate results.

​
Prerequisites

Before building ORB-SLAM3, you need to install the following dependencies:

​
C++11 or C++0x Compiler

ORB-SLAM3 uses C++11 thread and chrono functionalities. Make sure you have a compatible compiler: 
g++ --version  # Should be 4.7 or higher

​
OpenCV

ORB-SLAM3 uses OpenCV for image manipulation and feature processing.
OpenCV 3.0 or higher is required. ORB-SLAM3 has been tested with OpenCV 3.2.0 and 4.4.0.
# Install OpenCV (Ubuntu/Debian)
sudo apt update
sudo apt install libopencv-dev

# Verify installation
pkg-config --modversion opencv
For manual installation, download from opencv.org.

​
Eigen3

Eigen3 is required by g2o for linear algebra operations.
Eigen 3.1.0 or higher is required.
# Install Eigen3 (Ubuntu/Debian)
sudo apt install libeigen3-dev

# Verify installation
ls /usr/include/eigen3
For manual installation, visit eigen.tuxfamily.org.

​
Pangolin

Pangolin is used for visualization and user interface. 
# Install dependencies
sudo apt install libgl1-mesa-dev libglew-dev

# Clone and build Pangolin
git clone https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin
mkdir build && cd build
cmake ..
make -j4
sudo make install
For detailed instructions, see the Pangolin repository.

​
Python and NumPy

Python with NumPy is required for trajectory evaluation and alignment with ground truth. 
# Ubuntu/Debian
sudo apt install libpython2.7-dev python-numpy

# Or for Python 3
sudo apt install python3-dev python3-numpy

​
ROS (Optional)

If you plan to use the ROS examples, install ROS Melodic (Ubuntu 18.04) or ROS Kinetic (Ubuntu 16.04). 
# For Ubuntu 18.04 (ROS Melodic)
sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'
sudo apt-key adv --keyserver 'hkp://keyserver.ubuntu.com:80' --recv-key C1CF6E31E6BADE8868B172B4F42ED6FBAB17C654
sudo apt update
sudo apt install ros-melodic-desktop-full
See the ROS installation guide for more details.

​
Installation Steps

1

Clone the Repository

Clone the ORB-SLAM3 repository from GitHub:
git clone https://github.com/UZ-SLAMLab/ORB_SLAM3.git ORB_SLAM3
cd ORB_SLAM3
2

Build Third-Party Libraries

ORB-SLAM3 includes modified versions of DBoW2, g2o, and Sophus in the Thirdparty folder. The build script will compile these automatically.
The build.sh script will:
  • Build DBoW2 (place recognition)
  • Build g2o (graph optimization)
  • Build Sophus (Lie algebra)
  • Extract the ORB vocabulary
  • Build the ORB-SLAM3 library and examples
Make the build script executable:
chmod +x build.sh
3

Run the Build Script

Execute the build script to compile everything:
./build.sh
This process may take several minutes depending on your system. The script will:
  1. Build DBoW2 in Thirdparty/DBoW2/build
  2. Build g2o in Thirdparty/g2o/build
  3. Build Sophus in Thirdparty/Sophus/build
  4. Extract the ORB vocabulary from Vocabulary/ORBvoc.txt.tar.gz
  5. Build the main ORB-SLAM3 library
  6. Compile all example executables
4

Verify Installation

After successful compilation, verify that the library and executables were created:
# Check that the library was built
ls lib/libORB_SLAM3.so

# Check that example executables exist
ls Examples/Monocular/mono_tum
ls Examples/Stereo/stereo_euroc
ls Examples/RGB-D/rgbd_tum
If these files exist, the installation was successful!

​
Build Outputs

The build process creates the following:
  • lib/libORB_SLAM3.so: The main ORB-SLAM3 shared library
  • Examples/*/: Executable binaries for each sensor configuration:
    • Monocular/: Monocular SLAM examples
    • Stereo/: Stereo SLAM examples
    • RGB-D/: RGB-D SLAM examples
    • Monocular-Inertial/: Monocular + IMU examples
    • Stereo-Inertial/: Stereo + IMU examples
    • RGB-D-Inertial/: RGB-D + IMU examples

​
Building ROS Examples (Optional)

If you want to use ORB-SLAM3 with ROS:
1

Add to ROS Package Path

Add the ORB-SLAM3 ROS examples to your ROS package path:
echo "export ROS_PACKAGE_PATH=\${ROS_PACKAGE_PATH}:$(pwd)/Examples/ROS" >> ~/.bashrc
source ~/.bashrc
2

Build ROS Nodes

Run the ROS build script:
chmod +x build_ros.sh
./build_ros.sh
This creates ROS nodes for Mono, Stereo, RGB-D, and their inertial variants.

​
Troubleshooting

​
OpenCV Version Mismatch

If you encounter OpenCV version errors: 
# Check your OpenCV version
pkg-config --modversion opencv

# If needed, install a specific version
sudo apt install libopencv-dev=4.4.0*

​
Compilation Errors with C++11

Ensure your compiler supports C++11: 
# Update GCC if necessary
sudo apt install gcc-7 g++-7
sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-7 60
sudo update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-7 60

​
Pangolin Not Found

If CMake cannot find Pangolin: 
# Make sure Pangolin is installed to a standard location
export CMAKE_PREFIX_PATH=/usr/local:$CMAKE_PREFIX_PATH

# Or specify the path explicitly
cmake -DPangolin_DIR=/path/to/Pangolin/build ..

​
Permission Issues

If you encounter permission errors during build: 
# Make sure all scripts are executable
chmod +x build.sh build_ros.sh

# If installing system-wide, you may need sudo
sudo make install

​
Next Steps

Now that ORB-SLAM3 is installed, you’re ready to run your first example!

Quick Start Guide

Learn how to run your first SLAM example

Build docs developers (and LLMs) love

Get started for freeTalk to us