Keywords: FAST Feature Detection, LK Feature Tracking, Five-Point Motion Estimation

port from https://github.com/yueying/LearningVO

Results

Lucas Kanade

SIFT

ROMA

Dependencies

python3 -m venv .venv
source .venv/bin/activate
poetry install

Usage

Modify the path in test.py to your image sequences and ground truth trajectories, then run

python test.py --method roma --path ~/datasets/KITTI/dataset_sample