Stacking different resolution images in WBPP Pleiades Astrophoto PixInsight · Astrogerdt · ... · 6 · 716 · 2

Hi guys, 

I am taking images using a dual setup with DSLRs that produce different resolution images. The preprocessing is done in WBPP. Here are some screenshots showing my configuration: 

Post-Calibation tab: 


WBPP creates separate integrations for each camera. As far as I could find out, there is no way to tell WBPP directly to create only one integration of all images combined. 
As of now, my solution to this is the following: let WBPP only calibrate all the frames and then perform LocalNormalization, ImageIntegration and DrizzleIntegration on all frames manually using a manually created LN reference image. While this works and offers the benefit of integrating everything as one image set, this also has some problems: 
First, it involves some manual work that is spread out over a long period of time. That is inconvenient and error-prone, since it is easy to miss some small details when you do other stuff while PI is working on the last task. 
Secondly, this leads to a lot of pixel rejection around the stars since both setups have slightly different PSFs (due to the used lenses and rescaling of one image group). 

However, instead of doing this, I thought of another idea how to handle this situation that I would like to discuss here. Maybe this is slightly better and could also benefit others that are in the same situation. 

My new idea is to let WBPP run as shown in the settings above and perform every step except for DrizzleIntegration (i.e. calibration, StarAlignment, LocalNormalization and ImageIntegration). This already takes care of most of the manual work. 
After this, I would perform a DrizzleIntegration of all images combined in one step. 

This method has the following positive aspects: 
First, it is less error-prone. The only step that I would have to do manually is the final DrizzleIntegration. 
Secondly, since the pixel rejection is done for each set of images independently during ImageIntegration, this would reject less pixels around stars due to differing PSFs. 

However, there is one Problem I see here: LocalNormalization normalizes both images against different reference images, that (in most cases) will have different brightness. So in practice, this would integrate images with different brightness levels, which I assume to be a major drawback. 


From a purely theoretical standpoint, what are your thoughts on both workflows? Do you think the second workflow would offer some benefit? 

CS Gerrit

Try thisTo combine images from a dual imaging setup with different resolutions in PixInsight, you need to first align and rescale the images to match the same resolution and dimensions. Here’s a step-by-step workflow for combining these files after preprocessing: Step 1: Choose a Reference Image • Select one of your stacked images (e.g., from the higher resolution camera) as the reference for alignment. This will be the resolution that the other image will be rescaled to. Step 2: Use StarAlignment for Registration 1. Open the StarAlignment tool in PixInsight.2. Set the Target Image to your chosen reference image (higher resolution stack).3. Load the second image (lower resolution stack) as the Source Image.4. In the Star Detection and Matching settings, you may need to adjust the Star Detection Sensitivity if the images differ significantly in field of view or scale.5. Set the Output Image directory and ensure Rescale is enabled in the Registration settings. This step resizes the lower resolution image to match the higher resolution reference.6. Execute the process. Step 3: Crop if Necessary • If the images have slightly different fields of view, use the DynamicCrop tool to crop them to the same dimensions. Step 4: Combine the Images 1. Use the PixelMath tool to combine the images. This step can vary depending on whether you want to blend them, average them, or perform a different operation.For example, if you want a simple average:• Open PixelMath and enter the following expression: (Image1 + Image2) / 2  Replace Image1 and Image2 with the names of your aligned and rescaled images. 2. If you prefer other blend modes (e.g., weighting one image more than the other), adjust the expression accordingly:• For a 70/30 blend: 0.7*Image1 + 0.3*Image2  3. Apply the PixelMath expression to a new image to generate the combined output. Step 5: Final Adjustments • Perform any necessary post-processing, like noise reduction, color calibration, and stretching, to fine-tune the result. By following these steps, you can effectively combine the separate stacks from your dual DSLR setup, even with different image resolutions. Let me know if you need more help with specific settings!

Oskari Nikkinen:
Not what you were looking for, but Siril will happily integrate images of any pixel scale to one stack with no complaints.

Same with PI, but that isn't necessarily the best way to go about it.

Hi Gerrit,

I think using WBPP at least through the registration steps is perfectly fine but honestly may as well go all the way to create the 2 master files and follow the steps as
@Jeff Fishman states above. I would look through the WBPP log for the reference image selected for the hi res image stack and use that as your reference image when aligning the 2 masters using Star Alignment. After that you can use image integration and drizzle integration as you stated you like to do. I would not worry too much about not running local normalization as although I am a fan using this to simplify the gradients I find the new Gradient correction tools can easily accommodate this.

Kevin