aswathselvam/Mono-V-SLAM

Mono-V-SLAM

Authors

• Aswath Muthuselvam(118286204)

• Akhilrajan Vethirajan(117431773)

• Vishal Kanna Sivakumar(117764314)

• Shailesh Pranav Rajendran(118261997)

  • Date: 04/29/2022
  • Course: ENPM673 - Perception for Autonomous Robots

---

Output

Trajectory

Pointcloud

Simulation Video

Youtube video - link

Folder structure

$ tree -L 2
.
├── code
│   ├── main.py
│   └── utils
│        ├── load_dataset.py
│        ├── compute_pose.py
|        ├── bundle_adjust.py
|        ├── loop_closure.py
|        ├── recover_scale.py
│        ├── plotting.py
|        └── bag_of_words.py
├── dataset
└── README.md

Setup code

Setup dataset

1. Download odometry data set (grayscale, 22 GB), odometry ground truth poses (4 MB), odometry data set (calibration files, 1 MB) dataset from here.
2. Organize your <datatset-folder> folder in this structure:

<dataset-folder>
    ├── poses
    └── sequences

Setup environment

• Add the dataset folder location

# ~/.bashrc
export SLAM_DATA_FOLDER = <datatset-folder>

Install dependencies

• Install python requirements:

  pip install requirements.txt

• Install Pangolin for visualization from: https://github.com/stevenlovegrove/Pangolin

Run the code

python3 code/main.py