I imaged the Heart and Soul Nebula on 2018-01-14. It was quite cold at -18°C. I haven’t imaged anything in awhile and the cold temperatures usually makes things more difficult. I was concerned that some piece of the work flow wouldn’t work.
I did have a lot of trouble framing the field of view. The Heart and Soul nebula aren’t visible so it required a reasonably precise telescope pointing to get things started. Assuming the scope is pointing in the general area, then a series of pictures is taken that hopefully show this nebula and can be used to frame the image.
I had changed out the HD11 for the AT106 scope and piggy backed the camera with the Canon 70-200mm lens on a standard camera tripod head. I did a quite alignment and then did a precise goto to aim the scope. However, there aren’t any bright stars in the FoV so it took awhile to frame the object and get the guide camera/scope pointed at the guide star. With a straight through arrangement for the quide camera, it was physically challenging to get under the scope to see.
I almost gave up when the computer decided to stop working just after i got everything framed. At -18c my hands were frozen as was the rest of me. I rebooted the PC and went in side for a bit to warm up. Fortunately round 2 was a much easier setup and i refound the quide star and got things running pretty quickly.
I started with 180s exposures and retreated into the house where i can monitor things using a remote desktop. After a few exposures, i decided that 180s wasn’t long enough to capture the nebula, even though the stars were almost saturated. I did a few exposures at 360s. However, the platesolve2 app that FlexRx uses could not solve the image. This would be ok, but it leaves the meteguide shiftrate at an abnormally high value and i lost about a 1/2 hr of data due to bad tracking (high shift rate).
The remote desktop allows me to control BackyardEOS remotely. So after several unsuccessful attempts to get platesolve to work, i shut down flexRx and restarted the image run. I also decided to go with 240s exposures for the remainder of the capture.
After about 2hr, i woke up (yes, i sleep while the camera is working) to check things out. More bad luck! The clouds had rolled in which were not predicted by the CSC. So i shut down for the night and closed up the observatory.
I use PixInsight for image processing. This includes all calibration steps, alignment, stacking and post processing. I had a fair amount of difficulty aligning the subs. Red and Blue channels were fine. But there were many green channel subs that would not align. It took several passes using different settings and different reference images to get everything aligned. (I calibrate in raw, bayer format. Then debayer. Then split the channels into RGB and align all channels, all subs in one go. Then stack each channel separately. And finally combine the channel stacks into an RBG image for post processing.)
There are a lot of stars in the FoV of the Heart and Soul Nebula and the nebulae are faint in comparison to the star field. A standard processing would result in a dense star field obscuring the nebulae. I tried various techniques to actually remove (or reduce) the stars but none were successful. Generally this involves building a perfect star mask and then using a Morphological Transformation to reduce the stars. The MT attempts to shrink the stars and back fill with the surrounding colours. (Sort of like content aware fill in PhotoShop, but with shrinking star rather than a selection.)
I did minimize the stars along the way by choosing more moderate star masks and stretching the nebula while holding the stars more or less constant. The trick is in keeping a nice profile to the stars so they don’t look clipped or bright stars don’t look flat or washed out.
The purest view is the stars are there and an accurate representation of the FoV should include them. I think i managed a nice balance between stars and nebula.