Openstartracker roadmap

[Andrew Tennenbaum]

By 05/24/2018:
Detailed derivation of openstartracker via probability theory
Probability theoretic derivation of probability of match
Camera misalignment correction
In space camera misalignment correction procedure.

By 08/01/2018:
Easy in space recalibration
Add "just distractors" unit test
Rewrite openstartracker algorithm as classes with fewer side effects
Faster algorithm based on constellation probabilities rather than individual star probabilities

By 01/01/2019:
Distinguish non-stars from stars via eigenvector medthods and frame to frame relative position (H0: motion due to centroiding noise; H1: motion due to not being a star)
Proper data collection for non-stars
Distortion correction
Save and load databases

By 01/01/2019 (depending on various factors):
Documentation on image mask, collision resolution, weighted triad (may change)
Use Tetra for calibration

Future work:
Setup continuous integration
For initial guess, generalize constellations to allow a user selectable number of stars rather than just 2
GPU/multicpu parallelization
Add brightness
Neural net centroiding/ non-blind motion blur compensation
Reconstruct large field of view using multiple images taken over time
Multi camera startracking
Earth tracker
Horizon sensor
Moon tracker
Sun tracker
Delay compensation