To get you started we provide Python scripts that allows you to run the full star tracker simulation, creating your own scenes with your own parameters and test your star identification implementation.
If you port the library to other languages we encourage you to share the code. Get in touch with us to add the code to this page.
The following files are contained in the zip that can be found on the data page:
scripts/simulator.py: the python source code of the simulator that generates scenes.
scripts/example.ipynb: a sample ipython/jupyter notebook to generate scenes.
scripts/verify.ipynb: a sample ipython/jupyter notebook that evaluates a solution.
example/input_sample.csv: a sample input file with the same format as input.csv.
example/result_sample.csv: the resulting output file to the input_sample.csv.
Note: The hip_main.dat star catalog from the data directory is necessary to run some of the python scripts.
The Python library (simulator.py) contains a few classes for cameras, the star catalog, the star detector (handling the magnitude noise computations) and a scene class, as well as some utility functions. The example notebook shows which parameters can be set. It is the same script that generated the example data. The exact parameters used to create the competition's input data are, of course, a well-kept secret.
To run the notebook you need jupyter and the python libraries numpy, pandas, matplotlib and seaborn. Then open the example.ipynb notebook in jupyter and simply run it to generate random scenes and display the last one. The files in the example folder have been generated with this script and can be used to test your star identification algorithm.
The verify notebook only requires numpy and is very simple, comparing the contents of two csv files with the scoring function described on the evaluation page. Make sure to change the file names accordingly and it should run without problems.