The Pleiades – A complete Workflow with a DSLR and PixInsight

A cold winter night, a clear sky, the Pleiades are shining high in the sky. In this post I will tell you, step by step, how I took and processed this image.

The Subject

The open cluster Messier 45 in the constellation Taurus, better known as The Pleiades or The Seven Sisters, is a large object, spectacular when observed with a low-magnification telescope, thanks to the large number of hot and bright stars concentrated in a small space. In long-exposure photographs, the reflection nebula that surrounds the cluster is shown in all its splendour, revealing filaments and complex structures of rare beauty.

Here is the position of the subject in the sky and the field framed by my widefield camera system in some images obtained with the Stellarium software I used for planning. As you can see, the large field available allows me to easily frame the entire cluster and the surrounding nebulosity, with plenty of room for dithering and crop.

The Shooting

I took the shoot from my backyard at the edge of the urban area of Arezzo, the only place I could use at the moment because of Covid restrictions. The light pollution here is quite evident (I can’t give a correct estimate, but I think around Bortle 6-7), but the night was very clear and the Pleiades were very high in the evening sky.

Here is the specifications of my widefield rig, which I will describe in detail in another post. The large field of view allows me to comfortably include the whole cluster and the reflection nebula by centering the frame on the star Alcyone.

The use of the CLS filter, while negatively impacting on the colour rendering, allows me to reach guided exposures of 180s at ISO 1600. Given the limited focal length of the camera system and the large size of the pixels of the 6D (6.54 microns) the shot is undersampled: for this reason I will use a 2x drizzling integration.

0. Pre-Processing

For preprocessing I used PixInsight‘s new WeightedBatchPreprocessing script, leaving most of the options as default. I used an entire set of calibration frames including bias, flat and dark, as per the diagram below:

The image below compares a 50% crop of a single 180 seconds shot (to which only Debayer and AutomaticBackgroundExtractor were applied) with an integration of 55 calibrated 180 seconds shots stacked using 2x DrizzleIntegration (again,  AutomaticBackgroundExtractor was applied and the image was resampled to achieve the same scale despite the drizzling)

This is the result of the 55 x 180s integration using 2x drizzling.

1. Fitting / Colour Dominant Correction

The integration shows a strong green dominant, due to the use of the CLS filter. To eliminate it, I started by fitting the image: after separating the RGB channels, I applied the LinearFit process, leaving the default settings, to the red and blue channels, setting the green channel as reference. I then recomposed the RGB image using the ChannelCombination process.

After fitting, the image no longer shows the dominant but a light pollution gradient is clearly visible, which will be removed in the next steps. Autostretch has been applied to both images.

2. Crop

Thanks to the large field of view, the image offers many opportunities for composition: I decided on a square image, with some space around the main subject, which I will place in the center. I used the DynamicCrop process.

3. Gradient Correction

I corrected the light pollution gradient with DynamicBackgroundExtraction, manually placing a set of samples and raising the tolerance parameter a bit. Autostretch was applied again to the corrected image.

4. Colour Calibration

For the colour calibration phase I used the PhotometricColorCalibration process, by simply setting the shooting parameters.

Since the dominant caused by the CLS filter remained quite evident even after colour calibration, I made a pass with the SCNR process, setting it for green reduction. Now the colour is acceptable and I can proceed further.

5. Deconvolution

Despite the drizzling, the shot is significantly undersampled, so I did not find it worthwhile to spend too much time on deconvolution. However, I used the Deconvolution process, using a motion blur PSF, to correct a small elongation of the stars due to bad tracking, visible in this 100% detail: the effect is an increase in sharpness.

6. Noise Reduction

For the noise reduction step, which is very important given the limited integration time and the use of a DSLR, I used the two-step method described by Jon Rista.

The first step uses the TGVDenoise process to reduce high-frequency noise. A high-strength, low-contrast mask is applied to the image for this purpose, allowing very aggressive TGVDenoise settings to be used. However, finding the correct settings requires a lot of trial and error: I chose a rather drastic action, which greatly reduces noise, perhaps at the cost of a small loss of detail, as can be seen in this 100% crop.

In the second noise reduction step, I used the MultiscaleMedianTransform process to reduce low-frequency noise in both luminance and chrominance modes. Chrominance noise in particular is very noticeable and very difficult to reduce. Again, a mask is used to effectively protect the details.

At this point the linear phase processing is complete, and i can proceed with the stretching of the image!

7. Stretch

Applying an autostretch with  ScreenTransferFunction, the result seemed too contrasty and the stars too obtrusive. So I decided to use the MaskedStretch process conservatively and adjust the black and white points and contrast in the following steps.

8. Local Contrast

In order to enhance the structure in the reflection nebula surrounding the Pleiades and to make its fine filaments more obvious, I used the LocalHistogramEqualization process. The parameters are adjusted by trial and error, with the aim of achieving a pleasing but not excessive result.

9. Noise Reduction

Stretching and increasing the contrast had an impact on the noise and graininess of the image, increasing its perception. I then proceeded to a further noise reduction step, this time in non-linear phase, using the ACDNR process, again protecting the details with a mask. From this  100% detail you can see a slight drop in sharpness, which I hope will not have a negative impact on the final result.

10. Contrast and Curves Correction

We are now at the final stage of processing, using CurvesCorrection I adjusted the contrast and the black and white points. As for the colour, I increased the saturation and adjusted the RGB components according to the hue I wanted to achieve.

11. Final Corrections

Finding that there was still a residue of green dominant, I reduced it with a further pass of SCNR.

As a final step in Pixinsight, I made a pass with the DarkStructuresEnhance script to give greater emphasis and contrast to the structure of the nebulosity.

I used Photoshop for the last few small colour corrections, in particular to remove the annoying reddish colouration of the sky background, which has obviously not been correctly neutralised in the previous steps. Now all that’s left to do is resample, apply the correct sharpening and export the image in jpeg format!

Here is the final image at full resolution of 2400 x 2400 pixels.


Shooting and processing (several times) this image was a great exercise, from which I learned a lot. Although I am satisfied with the result, I am looking forward to photographing The Pleiades again next autumn, aiming to improve the shooting phase in particular and use a much longer integration time.

Leave a Reply

Your email address will not be published. Required fields are marked *