NavList:

Bill,

+/- 2 pixels is to be expected. This translates to about +/- 1' which isn't much. I just pick the center of the Sun by eye after zooming in x4. Calibration is done from a known GPS position on shore and picking a day with no wind and a sharp horizon. Take images every 4 minutes of time from 35° to 1°. Plot on graph paper the corrections in minutes of arc to get a zero intercept for each image. Make the cheat sheet for every 100 pixels by picking off the completed graph. This will make for easy mental interpolating on follow-up images. A little trial and error may be needed to select the two digit multiplier that results in the smallest correction values. A rough starting multiplier can be had by using the Sun's exact diameter from Calsky and determining minutes of arc per pixel from a centered Sun image with the lens to be calibrated.

Greg Rudzinski

From: Bill Lionheart
Date: 2016 Jan 28, 13:42 -0800

Thanks Greg I enjoyed repeating your calculations. I couldn't get the same pixel distance though. I used Gimp to get the centre of mass of the sun, then used the measurement tool to measure to the horizon. I got 2403 pixels not 2401 which gives me 1.1' difference to you. I wonder if you used the raw rather than jpeg? Or maybe just how we segmented the sun.

I'd love to know more about your lens calibration. How did you make the table? I thought the error would be most at the top and bttom of the frame and zero in the middle.

Did you have a GPS fix to compare?

Thanks

Bill