VAMR 2025 Mini Project
You can check this playlist for example runs of the pipeline on our custom dataset, KITTI, Malaga, and Parking.
Disclaimer: The pipeline currently has VIO (MSCKF) integration, which is not yet fully implemented. However, this tentative version can be run on the msckf branch with its own README.
Before installing the package:
Make sure that the following scripts are run on the corresponding branches:
Malaga -> main
Parking -> main
KITTI -> main
custom -> video-out
Make sure that you download and unzip the data packages from the course website into a directory named data, create data directory at the root of the package (on the same level as the src, tests, etc directories) that has the following structure:
data/capture_20251229_142428
data/kitti05/kitti
data/malaga-urban-dataset-extract-07/malaga-urban-dataset-extract-07
data/parking/parking
AFTER installation (instructions below), To run the testing script once the package is installed, follow the instructions below:
Provided Dataset:
Navigate to tests directory
Open up the test_tracking.py file, you can use this to configure what dataset you run the tests on (modify the name on line 19 to your preferred dataset).
In the case of any unexpected pathing errors, the file path taken by the package for a given dataset are constructed in the dataset_handler.py found in the top level of the src directory
Simply run the test_tracking.py file once you have chosen your preferred directory using python or python3 followed by the name of the file in your terminal.
Custom Dataset:
Navigate to src directory and make sure the custom data flag is uncommented on line 223. The pipeline can be run with the following command:
python vo_pipeline.py
P.s: press any key to start the pipeline
How to correctly install this package:
Before beginning install create the python environent with:
cd ThreeBlindMice
python3 -m venv .venv
First, inside of VAMRProject/TreeBlindMice activate the venv environment using:
source .venv/bin/activate
Make sure you are inside the venv when running any testing scripts
Next run the install script:
pip install -e .
If you dont have setuptools up to date run:
pip install --upgrade pip setuptools wheel
but this shouldnt be an issue
Finally, select the venv as your python interpreter: on vscode with the python extension by microsoft installed: ctrl + shift + p, Python:Select interpreter, Python 3.10.12 (.venv) ./.venv/bin/python
System Specifications
Maximum number of used threads: 1 CPU Frequency: 4 GHz RAM: 16 GB