ROS2 Stereo Visual Odometry with Local Bundle Adjustment Using g2o

An educational ROS2 stereo visual odometry system utilizing local bundle adjustment with the g2o library. This project is designed as a learning tool, offering insights into visual odometry and optimization techniques without aiming for high-performance enhancements. Based on concepts discussed in "SLAMBook2 Chapter 13".

Features

• Stereo Visual Odometry: Utilizes stereo camera data to estimate the robot's motion.
• Local Bundle Adjustment: Employs the g2o library to optimize the positions of keypoints across multiple frames.
• Educational Focus: Designed to provide hands-on experience with visual odometry techniques without the complexity of performance-optimized code.

Result

Dependencies

• ROS2 (Humble or newer)
• g2o
• OpenCV
• Eigen
• Sophus

Usage

Launch the stereo visual odometry node using the following ROS2 command:

ros2 launch stereo_visual_slam slam.launch.py