English Español Français

Leo Triplet with DSLR

by Alejandro Tombolini


Post processing of an integrated DSLR image.

Image of Marc Hoeppner publicated in PixInsight Forum - Post processing until the final image of an integrated image taken with Carbon Newton 200/800, Canon 1000Da, EQ6, Mgen autoguider. Only 25*180 seconds light, 10*180 seconds darks, 20 bias and 20 flats. - Main notes: CanonBandingReduction and DBE to control background, MultiscaleMedianTransform for noise reduction and contrast in the galaxies and Morphological Transformation and MultiscaleLinearTransform to process stars.


The firs step is to apply CanonBandingReduction script to reduce some banding.  Check  "Protect from Highlights" to avoid some dark banding where are located the galaxies. 

Apply DynamicBackgroundExtraction to extract some background gradient. The pattern is a little odd and I will reduce the smothing factor to remove it, so in this case is important to carefully place the samples where there is not stars.

Select a preview that represent the background of the image and apply BackgroundNeutralization tool. The values of upper and lower limit can be extracted from Statistics tool. 

For calibrating the color take the same preview used to neutralize the background as reference of the background in ColorCalibration tool with the new corresponding values updated, and make an Aggregated image that include the three galaxies to be use as reference image of the white reference. 

Use CloneStamp tool to delete a green hot pixel near one of the galaxies.

Generate a PSF with DynamicPSF tool to be used as External PSF on Deconvolution tool.  

Apply Deconvolution to increase details on galaxies and to improve stars profile controling the ringing with a convenient Deringing value determinated by trial and error. 

Stretch of the image using MaskedStretch tool . Set the value of Target background a little lower than default and select the same preview previously usde as background reference. 

Generate a mask to protect the background and saturate the images. Apply CurvesTranformation while protecting with the mask.

Normaly the saturation increase the chrominance noise. To reduce the noise apply ACDNR only to chrominance (uncheck lightness) without using mask. See before and after noise reduction: 

Protecting the background with the previously created rangemask, apply MultiscaleLinearTransform to increase details on the galaxies.

The next two steps are two consecutive applications of MultiscaleMedianTransform to noise reduction in the galaxies using two different rangemasks to protect the background. Created as detailed: 

Note that besides nois reduction I have also finetuned the Bias in some Layers to increase contrast. 

Now generate a structure_map mask with StarAlignment tool and adjust the numbers of layer to avoid that detect structures in the galaxies. Using that mask to protect the background apply MorphologicalTransformation tool to reduce stars. 

Generate a new mask with StarMask tool that include the biggest stars of the image with the objective of repair its saturated centers. 

Apply MultiscaleLinearTransform deleting the first three layer to have a better profile of those stars.

Finally extract the Lightness of the image to use as mask and protect bright region while adjust the briht of the background.

Apply GammaStretch to reduce the bright of the background protecting the galaxies. You can choose here how much to darken the background.

Final image: 

Last Update: 28/06/2016 16:43 UT

English Español Français