New coma corrector!

I’ve been battling some odd colour fringes around stars with my photos, and I figure that the only part of the optical train that might be causing this is the Bintel coma corrector. I noticed that the lens has a slight rattle to it and that the locking nut was ever so slightly loose.

I decided to pick up a Baadar MPCC for several reasons – it seems quite popular for astrophotography, it is much lighter than the Bintel CC, and it is decently priced.

MPCC

Quite a significant drop in weight! It also sits flush against the focus tube, meaning I don’t have to worry about getting the right distance as much as I did with the Bintel CC. And since it sits closer, the weight of the camera has less of an effect on the overall balance of the telescope.

Some test shots show that coma is greatly reduced around the edges of the image (better than with the Bintel CC), and although it’s a bit early to tell, the colour fringing appears to have been eliminated.

NGC 2467 – Skull and Crossbones nebula

NGC 2467

Thirty-Three 5-minute light frames, calibrated, registered, stacked, and post-processed using PixInsight. This may well be the most exposures I have taken of one object in one night, and the resulting reduction in noise is fantastic!

Date taken: 05/01/2016
Location: Adelaide, South Australia
Camera: Canon 450d w/ IR filter removed, GSO coma corrector
ISO800
Mount: HEQ5PRO
Scope: GSO 8″ f/5 Newtonian
Autoguider: Orion Starshoot AG
Imaging: BackyardEOS w/ PHD dithering
Guiding: PHD2
50 bias frames, 20 dark frames, 50 flat frames.
Total integration time: 165 minutes

PRE-PROCESSING DETAILS:

  • BatchPreProcessing, calibration and debayer
  • Blink to find and remove any really bad frames
  • Star Alignment
  • Blink to make sure everything aligned well
  • Image integration (using Winsorized Sigma Clipping)
  • Automatic Background extraction

PROCESSING THE IMAGE:

  • Background Neutralisation
  • Colour Calibration
  • Histrogram Transformation
  • TGVDenoise (masked, L channel extracted post-histogram extraction)
  • UnsharpMask (masked, with a range mask so only the interesting nebulae were affected)
  • Histrogram Transformation x2 to further stretch the image after noise reduction
  • Curves Transformation x2 to bring out the colours
  • scnr to remove the green tinge
  • DynamicCrop, removing the less interesting edges of the images

Far less effort was required to reduce the noise in this image – TGVDenoise was able to eliminate almost all of the small scale noise in a single pass without ruining the details of the nebula.

Easy drift alignment

One of the problems facing beginners is how do they align their mount with the celestial pole. You can use the polar scope that is built into your mount (if it has one!), but if like me you’re in the southern hemisphere you might find it to be pretty difficult to use.

The method used here is called Drift Alignment by Robert Vice, or D.A.R.V for short.

Things you’ll need

  • String
  • A bit of wood with a 1M dowel screwed into it, 90° to the flat surface
  • Your equatorial mount
  • An inclinometer
  • A laptop
  • A DSLR/CCD
  • Your telescope
  • GPS (A phone with GPS will do fine)

What to do

The first thing you need to know is which way is South/North. You can use a compass here to get the rough direction of the pole nearest you. A better way to work out the north-south line would be during the day, get a stick and mount it 90° to the ground. Look up the Solar Noon for your location and free up that time. You’ll want something at this point to mark the north-south line – I used string. When you reach the time you found out to be solar noon for your location, mark out using string/tape/whatever you have along the line of the shadow cast by your stick. This is the north-south line that you will need to roughly get your mount aligned to the celestial pole.

You can start setting yourself up before sunset, there’s no point in wasting good imaging time setting up the telescope and mount! The first thing you need to do is get the tripod set up where you want it, pointing roughly at the celestial pole. You should be able to eyeball it by looking down the tripod leg and comparing it to the line you now have indicating the north-south line. It doesn’t have to be absolutely perfect since we’re going to be doing further refinement later on using the mounts built in adjustments. That said, the closer you get it now the less adjustment you’ll have to do later. Next, set your inclinometer on top of the tripod. At first set it east-west and try to get it level by adjusting the tripod legs. Once you have it sitting absolutely level, turn it so it’s aligned with the north-south line and adjust the north/south tripod leg to level it out. You should now be seeing 0° no matter how you orientate the inclinometer. Make sure the legs are settled, if they can be pushed out further you will have to re-level the tripod! When you attach the accessory tray it will push out the legs a little. Unless you are on a very uneven surface or didn’t make sure the legs are fully pushed out, this shouldn’t affect the inclination of the mount.

Mount the equatorial head on top of the tripod, and screw in the primary locking shaft and accessory tray. When the locking knob is wound up completely it will push against the tripod legs, locking them into place. At this point you should work out your latitude from your phone or other GPS device. We’re going to need it to set the correct latitude on the equatorial head. Mount the inclinometer to the dovetail mount and switch it on. Use the latitude adjustment screws on the equatorial head until the inclinometer reads the same latitude as your GPS. We’ll refine this further later on.

Remove the inclinometer and set up the counter weights on your mount. Mount the telescope and camera, and balance everything. Hook up the power and any other cables (eg, USB to the camera) and fire up your laptop. Turn everything on and follow the prompts on your mounts computer to configure the right lat/lon, time, timezone, date, daylight savings etc. Skip the star alignment for now, and just go into settings and get the mount moving at sidereal rate. Slew the mount to point it towards a bright-ish star and get your focus roughly right. It doesn’t need to be perfect for this method to work but it will work better with a sharper image.

telescope-setup

Using your hand controller, slew the mount to point north if you’re in the southern hemisphere, around 0° Dec.  Fire up BackyardEOS on your laptop and switch on the camera if you haven’t already. Just use the frame&focus mode for now. Set the ISO level to something low – 200 usually works well but you can use 400 if there are no brighter stars visible. Set the exposure time to 130 seconds – the longer you expose the more accurate your alignment will be. Set the slew rate on your telescope to 1x, and start the capture. Wait until 10 seconds have passed and then press the Left/West slew button on your hand controller for 60  seconds. At 70 seconds of exposure time, press the Right/East button on your hand controller.

When the exposure finishes, you’ll see a number of trails with a bright spot at one end. If you’re extremely lucky you’ll just have single straight lines, which means no adjustment is required! It’s far more likely you’ll see something that looks like a < symbol. In this case, we need to make some azimuth adjustments to the mount. There’ll be two knobs on the celestial pole side of the mount (the front I guess?). Adjust those a few turns in one direction, lock them off, and try the exposure again. You’ll get an idea for which way to adjust the knobs for your setup. For me it’s clockwise if the star appears to move up, counter-clockwise if the star appears to move down.

Screenshot (8)Picking a reasonable bright star.Screenshot (9)Pointing north, drifting down. The corrected needed here is counter-clockwise. Screenshot (10)Same star as above after the initial adjustment, still a bit more counter-clockwise needed.Screenshot (11) Whoops! Too far! Slight adjustment clockwise.Screenshot (12)Perfect!

Once you’re satisfied with the azimuth alignment of your mount, slew it over to the east or west, again around 0° Dec. As before, slew speed 1x, 130 second exposure, and wait 10 seconds before pressing the Left/West button. Hold down the West button for 60 seconds and then at 70 seconds, press and hold down the East button. Since we set up the latitude using our inclinometer before, it’s far less likely to need any huge adjustments. If you’re always imaging from the same location, you can leave it set at the same latitude. As before, experiment with the latitude adjustment screws until you get a single straight line and not a <.

You can repeat the above steps as many times as you like, but you’ll see diminishing returns. It’s a good idea to go back and repeat the north/south check before you go any further. You should now perform a 2 or 3 star alignment if you want to use the go-to functionality of your mount.

NGC281 – A response to “Astrophotography Tutorials”

NGC281

Doug has a great channel called Astrophotography Tutorials that I highly recommend. He recently uploaded some raw data and asked his subscribers to have a go at processing it. This is my first attempt at using narrowband data, and I think I might need to look at upgrading from the DSLR!

Rather than type out the process I went through, I captured the screen as I processed the data:

NGC2438

Fourteen 5-minute light frames, calibrated, registered, stacked, and post-processed using PixInsight. I took this after the Horsehead Nebula was too low in the sky to get any useful images.

The gathering of stars just above the nebula are commonly known as M46, this region is within the constellation Puppis.

Date taken: 13/03/2015
Location: Adelaide, South Australia
Camera: Canon 450d w/ IR filter removed, GSO coma corrector
ISO800
Mount: HEQ5PRO
Scope: GSO 8″ f/5 Newtonian
Autoguider: Orion Starshoot AG
Imaging: BackyardEOS w/ PHD dithering
Guiding: PHD2
50 bias frames, 20 dark frames, 50 flat frames.
Total integration time: 70 minutes

Pre-Processing details:

  • BatchPreProcessing, calibration only
  • BatchDebayer
  • Star Alignment
  • Image integration (using Winsorized Sigma Clipping)
  • Automatic Background extraction
  • Dynamic Crop to remove some of the dodgy edges
  • Automatic Background extraction to remove any remaining unwanted gradients
  • Extracted luminance from the RGB for later on

Processing the RGB Image:

  • Background Neutralisation
  • Colour Calibration
  • scnr to remove the green tinge
  • atrouswavelettransform (masked, duplicate image, STF autostretch)
  • multiscalemediantransform (masked, duplicate image, STF autostretch)
  • Histrogram Transformation
  • atrouswavelettransform (masked, L channel extracted post-histogram extraction)
  • ACDNR (masked, L channel extracted post-histogram extraction)
    atrouswavelettransform (blurring the image)

Processing the luminance image:

  • atrouswavelettransform (masked, duplicate image, STF autostretch)
  • multiscalemediantransform (masked, duplicate image, STF autostretch)
  • Histrogram Transformation
  • atrouswavelettransform (masked, cloned image with auto clip shadows/highlights)
  • ACDNR (masked, cloned image with auto clip shadows/highlights)

Putting the images together:

  • LRGBCombination, saturation initial tweak.
  • Curves, increase contrast, bring out details

IC434 – The Horsehead Nebula

ic434-lrgb-final

Seventeen 5-minute light frames, calibrated, registered, stacked, and post-processed using PixInsight. I have tried and failed to capture the Horsehead nebula before, so I’m glad to finally tick this one off!

Date taken: 13/03/2015
Location: Adelaide, South Australia
Camera: Canon 450d w/ IR filter removed, GSO coma corrector
ISO800
Mount: HEQ5PRO
Scope: GSO 8″ f/5 Newtonian
Autoguider: Orion Starshoot AG
Imaging: BackyardEOS w/ PHD dithering
Guiding: PHD2
50 bias frames, 20 dark frames, 50 flat frames.
Total integration time: 85 minutes

Pre-Processing details:

  • BatchPreProcessing, calibration only
  • BatchDebayer
  • Star Alignment
  • Image integration (using Winsorized Sigma Clipping)
  • Automatic Background extraction
  • Dynamic Crop to remove some of the dodgy edges
  • Automatic Background extraction to remove any remaining unwanted gradients
  • Extracted luminance from the RGB for later on

Processing the RGB Image:

  • Background Neutralisation
  • Colour Calibration
  • scnr to remove the green tinge
  • atrouswavelettransform (masked, duplicate image, STF autostretch)
  • multiscalemediantransform (masked, duplicate image, STF autostretch)
  • Histrogram Transformation
  • atrouswavelettransform (masked, L channel extracted post-histogram extraction)
  • ACDNR (masked, L channel extracted post-histogram extraction)
    atrouswavelettransform (blurring the image)

Processing the luminance image:

  • atrouswavelettransform (masked, duplicate image, STF autostretch)
  • multiscalemediantransform (masked, duplicate image, STF autostretch)
  • Histrogram Transformation
  • atrouswavelettransform (masked, cloned image with auto clip shadows/highlights)
  • ACDNR (masked, cloned image with auto clip shadows/highlights)

Putting the images together:

  • LRGBCombination, saturation initial tweak.
  • Curves, increase contrast, bring out details
  • Rotated 90°

M42 – Orion Nebula – one year later

 Update! I’ve reprocessed the image:

m42_HDR_latest

The steps were mostly the same as before, but the HDR stage was tweaked a bit:

  • RGBWorkingSpace changed to 1 for each channel
  • DynamicPSF to generate PSF for Deconvolution
  • Luminance mask created from RGB image
  • Deconvolution used to bring out some more details
  • Multiple HistrogramTransformation runs to stretch the image carefully, and align the channels
  • HDRMultiscaleTransform with the brightest stars masked, once on the larger scale (~10) and once on a smaller scale (~7)
  • SCNR to reduce green
  • ACDNR to reduce noise
  • HistogramTransformation to stretch the image a little further
  • RangeMask+StarMask used to protect the background & bright stars
  • LocalHistrogramEqualisation used on the nebula to brighten things up
  • CurvesTransformation to increase contrast
  • UnsharpMask to sharpen some details slightly
  • Slight Convolution to reduce a bit of the harshness from UnsharpMask
  • ColourSaturation three times, one with a mask protecting everything but the brightest stars, one protecting the background and bright stars to increase saturation on the nebula, and once with the previous mask inverted to reveal some colour in the dusty background.
Original post:

m42-HDR

I’ve shot this target before!

I haven’t done a HDR shot before, most targets don’t really need it in my experience. There is one huge exception – M42! The core of M42 is super bright compared to the rest, and then there’s the dust which usually goes unseen until you take much longer exposures.

M42 is easy to capture, hard to master.

General Details:
  • Date taken: 14/12/2014
  • Location: Adelaide, South Australia
  • Camera: Canon 450d w/ IR filter removed, GSO coma corrector
  • ISO800
  • Mount: HEQ5PRO
  • Scope: GSO 8″ f/5 Newtonian
  • Autoguider: Orion Starshoot AG
  • Imaging: BackyardEOS w/ PHD dithering
  • Guiding: PHD2
  • 50 bias frames, 20 dark frames (for the 300 second lights only!), 50 flat frames
  • 15 x 10 second exposures
  • 15 x 60 second exposures
  • 12 x 300 second exposures
  • Total integration time: 4650 seconds
Pre-Processing details (performed on each exposure length individually):
  • BatchPreProcessing, calibration only
  • BatchDebayer
  • Star Alignment with drizzle (took a lot of tweaking to get the 10 second exposures to align correctly!)
  • Blink to check for and remove any bad frames
  • Image integration with drizzle (using Winsorized Sigma Clipping)
  • Automatic background extraction on the three integrated RGB images to remove the light pollution
Processing the resulting RGB Images:
  • BackgroundNeutralization
  • ColourCalibration
  • SCNR to remove the green tinge
Prep the images for HDR:
  • StarAlignment of the three different integrations so they’ll line up
  • Blink to make sure they are lined up OK
  • DynamicCrop to cut out any bad edge bits
HDR Time:
  • HDRComposition to combine the three different exposure length integrations
  • RGBWorkingSpace
  • ACDNR to reduce noise
  • HistogramTransformation x2 to bring out the background dust
  • HDRMultiscaleTransform to reveal the core of M42
  • ColorSaturation to enhance the colours a bit
  • CurvesTransformation to enhance contrast and luminance
  • LocalHistogramEqualisation to brighten the image a bit

Migrating to Android Studio 1.0.0 – missing method “runProguard”

When I updated Android Studio to the 1.0.0 release, I found the Gradle project sync was no longer working. A quick look at the messages revealed it wasn’t happy with the method “runProguard”.

Error:(16, 0) Gradle DSL method not found: 'runProguard()'
Possible causes:
The project '<AppName>' may be using a version of Gradle that does not contain the method.
Open Gradle wrapper file
The build file may be missing a Gradle plugin.
Apply Gradle plugin

According to the Android build system documentation migration guide, it has been replaced with minifyEnabled. You’ll need to open the build.gradle file for your app and correct it like this:

  buildTypes {
 release {
 -runProguard false
 +minifyEnabled false
 proguardFiles getDefaultProguardFile('proguard-android.txt'), 'proguard-rules.pro'
 }
 }

NGC1365 – Great Barred Spiral Galaxy

ngc1365_final

Twenty-five 5-minute light frames, calibrated, registered, stacked, and post-processed using PixInsight. I was having a bit of trouble before this shot, my stars had colour fringes caused by the coma corrector. I have started performing my collimation with the coma corrector in place, which seems to have reduced this problem significantly. I think it’s just because the coma corrector isn’t as snug a fit as it should be in the focuser tube.

  • Date taken: 27/11/2014
  • Location: Adelaide, South Australia
  • Camera: Canon 450d w/ IR filter removed, GSO coma corrector
  • ISO800
  • Mount: HEQ5PRO
  • Scope: GSO 8″ f/5 Newtonian
  • Autoguider: Orion Starshoot AG
  • Imaging: BackyardEOS w/ PHD dithering
  • Guiding: PHD2
  • 50 bias frames, 20 dark frames, 50 flat frames.
  • Total integration time: 125 minutes

Pre-Processing details:

  • BatchPreProcessing, calibration only
  • BatchDebayer
  • Star Alignment
  • Image integration (using Winsorized Sigma Clipping)
  • Dynamic Background extraction, twice, first time to remove the majority of the light pollution, then a second run to reduce some of the gradients from nearby light sources.
  • Dynamic Crop to remove some of the dodgy edges, also centering the image on NGC1365
  • Extracted luminance from the RGB for later on

Processing the RGB Image:

  • BackgroundNeutralization
  • ColourCalibration
  • ATrousWaveletTransform to reduce noise
  • HistogramTransformation to reveal the galaxy etc
  • ACDNR to further improve noise
  • ATrousWaveletTransform a few more times with various masks to reduce noise and even out the background
  • Blurred the RGB image with ATrousWaveletTransform by removing all the detail from the first 4 layers!

Processing the luminance image:

  • Much the same as before, ATrousWaveletTransform to reduce noise
  • HistogramTransformation to stretch the image
  • More ATrousWaveletTransform, this time using a mask generated from a cloned image, HistogramTransformation used to auto clip highlights & shadows
  • Same mask, ACDNR.
  • Bit more HistogramTransformation now that there is a lot less noise
  • Pixel math: star_mask+range_mask, masking everything but the stars and galaxy
  • LocalHistogramEqualization to bring out some more detail in the galaxy
  • Tiny bit of bias in ATrousWaveletTransform to sharpen up the galaxy

Putting the images together:

  • LRGBCombination!
  • Saturation increased a tad, 0.350 or so, but otherwise just applied L to RGB
  • CurvesTransformation to tweak the saturation and luminance a tad, and increase contrast

M8 – Lagoon Nebula – Take 3

m8-reprocessed

I have re-used the data from my previous attempt, this time the processing was done entirely in PixInsight. If I recall correctly, there was 10 light frames used for a total integration time of 50 minutes.

  • Date taken: 14/08/2014
  • Location: Adelaide, South Australia
  • Camera: Canon 450d w/ IR filter removed, GSO coma corrector
  • ISO800
  • Mount: HEQ5PRO
  • Scope: GSO 8″ f/5 Newtonian
  • Autoguider: Orion Starshoot AG
  • Imaging: BackyardEOS w/ PHD dithering
  • Guiding: PHD2

The colour calibration is much more accurate this time around. I have also managed to reduce the noise considerably during the integration phase by using Winsorized Sigma Clipping. This allowed me to be a lot less aggressive with the noise removal tools, resulting in a more defined image.