GNC-Payload

This repo strives to house everything associated with the payload of the Argus satellite. This includes:

• Scripts for downloading Landsat and Sentinel data (image_downloader folder)
• Scripts for training and running inference of the region classifier and landmark detector (vision_inference folder)
• Orbit determination algorithms (orbit_determination folder)
• Code for mocking IMU measurements (sensors folder)
• Code for simulating images given a position and orientation of the satellite in ECEF (image_simulation folder)

Code from the following repos was used as a starting point:

• Earth Engine Downloader
• Vision Pipeline Inference
• Image Simulation
• Vision Training Pipelines

Setting Up

Conda Environment Setup

An environment_training.yml has been provided. The easiest way to set up a v-env is to run:

conda env create -f environment_training.yml

Back-up Conda Environment Setup

Depending on your OS the above command may not work (tested on Ubuntu 22 (works) and Windows (doesn't seem to work)). In this case set up manually via:

conda create --name MY_ENV python=3.12

Activate your environment MY_ENV and run the requirements file:

conda activate MY_ENV
pip install -r ./requirements.txt

This should set up the environment and packages needed but this can be flaky if pip uses cached packages it finds on your machine.

Installing this repo as a package

After activating your conda environment, do the following to install the necessary modules as editable packages:

pip install -e .

User-Specific Configuration

Copy user_config.example.yaml to user_config.yaml and fill in the necessary fields.

cp user_config.example.yaml user_config.yaml

Downloading Vision Models to run Vision Inference

The trained models and CSV files are stored on the Argus CubeSat Google Drive (here). Download the desired version and unzip it to the desired location. Update the models_directory field in your user_config.yaml file accordingly.

Pre-Commit Hook

A pre-commit hook can be set up to run pylint and black formatter.

The hook can be set up by running the following command in the root directory of the repository:

pre-commit install

Then every time you run git commit ... the pre-commit hook will run the linter and formatter and check that everything is in order. Only commit to your branch if ALL tests pass. Otherwise, address and correct the shown errors. These tools can also be run separately or on specific folders or files by specifying for example:

pylint ./sensors/bias.py
black ./orbit_determination/nonlinear_least_squares_od.py
pylint ./vision_inference