This is an implementation of iMAP SLAM in PyTorch, which is a cleaned up version of ueda0319's iMAP implementation. This version is edited for a beginner to understand, with descriptive variable names, lots of comments, and minor tweaks. Hopefully this can help others to learn NERF and NERF based SLAM algorithms. Thanks to ueda0319 for the original implementation!
- PyTorch 1.8
- opencv-python
- numpy
Download RGBD dataset from TUM(https://vision.in.tum.de/data/datasets/rgbd-dataset/download)
run with following commands
python run_slam_tum.py path/to/rgbd_dataset_freiburg1_teddy/