/EdgeDirectVO

Edge-Direct Visual Odometry

Primary LanguageC++

Edge-Direct Visual Odometry

If you find this useful, please cite the related paper:

@article{DBLP:journals/corr/abs-1906-04838,
  author    = {Kevin Christensen and
               Martial Hebert},
  title     = {Edge-Direct Visual Odometry},
  journal   = {CoRR},
  volume    = {abs/1906.04838},
  year      = {2019},
  url       = {http://arxiv.org/abs/1906.04838},
  archivePrefix = {arXiv},
  eprint    = {1906.04838},
}

Setup

This repository assumes the following directory structure, and is setup for the TUM-RGBD Dataset:

EdgeDirectVO
|-- tum_rgbd_dataset/
    |-- depth/
    |-- rgb/
    |-- assoc.txt
|-- tum_rgbd_dataset_ground_truth.txt

Be sure to run assoc.py to associate timestamps with corresponding frames.

Build and Run

cd build && cmake .. && make -j && ./EdgeDirectVO

Evaluation

After evaluating on a dataset, the corresponding evaluation commands will be printed to terminal. Simpy copy and run them in terminal in project root directory. Examples are shown below:

python evaluate_rpe.py groundtruth_fr1xyz.txt rgbd_dataset_freiburg1_xyz_results.txt --fixed_delta --delta 1 --verbose
python evaluate_ate.py groundtruth_fr1xyz.txt rgbd_dataset_freiburg1_xyz_results.txt --verbose