brickbots
commented
1 month ago Just tested this under the stars and it works FANTASTIC! The estimate roll seems to match very, very closely and it provides a really nice experience when using the chart.
Thank you @TakKanekoGit for this great contribution 💯
Compensates for the roll offset between the roll seen by the camera and the expected roll at the (RA, Dec). This should also fix issue https://github.com/brickbots/PiFinder/issues/117.
The roll at the camera center is different from the roll at the target pixel. To address this, the (RA, Dec, Roll) at the camera center get stored as solution["RA_camera"], solution["Dec_camera"] and solution["Roll_camera"]. The coordinates at the target pixels are still the same: solution["RA"], etc.
The roll offset is calculated at the camera center in integrator.py and stored as solution["Roll_offset"] every plate solve. solution["Roll"] at the target is calculated from the expected roll at the target (RA, Dec) and compensated with the Roll_offset.
This should display the constellation with the correct roll as seen by an observer. This PR hasn't been system-tested on a PiFinder or field-tested. It's possible that the sign for the roll_offset is the wrong way round. If so, this can be changed in integrator.estimate_roll_offset().