I LOVE finding out about different ways to appreciate the Sun and light in general. Use this forum to post your info or questions about various outside the mainstream ways to appreciate our life giving star!
While it is fairly easy to spot the H-alpha line in the solar spectrum, the other Balmer series lines are harder to identify. To find H-beta, I typically have used an H-beta filter to help with the process. But this isn't possible with the H-gamma and H-delta lines. So I decided to use a combination of my Hydrogen spectral tube, foamboard SHG and direct sunlight to create some "maps". I found I could use these maps to find to correctly identify the lines (I had the spectral tube nearby so I could check if I was correct).
hydrogen solar spectrum.JPG (47.46 KiB) Viewed 1081 times
I took a series of screenshots of the solar spectrum as it appears using FireCapture with a ZWO 183MM camera. I used Photoshop to crop the images and adjust the levels a bit. If you click on the images, they appear with better resolution.
I had also problems to identify lines and to focus the telescope in full sunlight using SharpCap. Sometimes it took only a lot of time, but in the worst case I got the wrong line. So I decided to write a python program to help with adjustment. At the moment, it is more or less a sketch of gui to see what is really needed. Getting the video from the ASI178 camera works, but motors and controller are still waiting on my desk. Target is support during adjustment and line identification, data rate is not sufficient for acquisition.
For line identification a tabular reference spectrum taken from BASS2000 is re-binned to the approximate spectral resolution and plotted as curve and as virtual spectrum next to spectrum and curve taken from SHG. Markers and annotations at lines from a table help to identify. The same gui tab holds also the buttons to control the grid adjustment motor.
Wavelength.jpg (164.83 KiB) Viewed 1053 times
Focus of SHG was the least problem but I want to be quick when changing the range especially in UV. Therefore three regions along the spectrum are zoomed and also displayed as curves.
Spectrum.jpg (107 KiB) Viewed 1053 times
Focus of telescope is supported by zooming on the edge of the spectrum as well as three zoom regions along the spectrum. Numerical measures of sharpness will be added later.
I think the most critical area of the spectrum is violet and near ultraviolet, from 4000 to 3880A, containing Ca H, Ca K and CN band. Here the chromatic aberration changes very fast with wavelength, producing the "fishtail" in normal spectroscopy. Therefore we can not use any features too far from the target wavelength.
Most of sharpness measurement methods provided in libraries like openCV do not fit to our problem, as we have different sources of blurring in the two principal directions. I will use the terminus "vertical" for the cross slit or spectral direction and "horizontal" for the spatial direction along the slit, as we see the picture this way in SharpCap or Firecapture. If possible, I would like to use only simple means of evaluation, not looking at tilt or smiley of the spectrum.
We need to focus the spectrum first with the motor on the camera, using vertical details close to the line we want to observe. This focus is independent of the telescope focus, but due to flexture, temperature expansion and optical consideration it needs to be done shortly before acquisition. Maybe the region to analyze needs to be a little larger than the region we finally take in the video. So we have some 50 to 100 pixels on both sides of the target line to evaluate focus. Autocorrelation of signal or simple differentiation in vertical direction to get only the high frequency content may give good numbers.
For telescope focuser, things are more complicated:
First we have the width of the edges (e.g. from 20 to 50% of maximum signal) to determine sharpness. But in this case we need to exclude the spectral line itself, otherwise we will see a protuberance as lack of sharpness. This exclusion can be based on intensity or on area within our picture.
Second we have surface details along the line. Differentiation in horizontal direction will help, but we need to limit our analyze area to the line features. Turbulence and vibration will strongly change our measures, so I expect a lot of difficulties to arise here.
