Python implementation of the Stereo PTAM (parallel tracking and mapping), See the paper: "[S-PTAM: Stereo Parallel Tracking and Mapping](docs/S-PTAM: Stereo Parallel Tracking and Mapping.pdf) Taihu Pire et al. RAS17"
Example of S-PTAM on KITTI dataset
- Python 3.6+
- cv2
- numpy
- g2o (python binding of C++ library g2o) for BA and pose graph optimization
- pangolin (python binding of C++ library Pangolin) for visualization
- Multithreads Tracking, Mapping, and Loop Closing;
- Covisibility Graph (representing the relation between keyframes, mappoints and measurements);
- Local Bundle Adjustment and Pose Graph Optimization;
- Motion Model (used for pose prediction, then for reliable feature matching);
- Point Clouds and Graph visualization;
python3 main.py