Tim Hutchison (@Tim Hutchison) and I both have an interest in processing data from the James Webb Space Telescope. We decided to test out our similar, but different techniques on the recently posted Pillars of Creation data from the JWST NIRCam using the five filters. A description of my approach is below. Tim's version is described in his posting (Tim's JWST NIRCam Pillars of Creation). Be sure to read his as well.

We will be posting a second image of the Pillars of Creation from data from the JWST MIRI. It's a beautiful but different representation of this target.

JWST – Creating the Pillars of Creation Work Flow
There are several ways to process the Pillars of Creation image from the JWST image database. This is one of them. I am not claiming that it is the best. It is, however, a simple step by step procedure that yielded the image you see above (This image has been down-sampled for faster downloading). It is also a process that can be used on any of the recent images that have been posted from the JWST. Feel free to both use and improve this process as you see fit.

Acquire the data. Data from almost all of the professional/space telescopes can be downloaded free from Mikulski database. Here is a link to that portal (https://mast.stsci.edu/portal/Mashup/Clients/Mast/Portal.html ). Now enter the target of interest. In this case it will be M16. This brings up data from all the telescopes. Search the page for JWST and click on that box. Now you have selected only those images and data from the JWST. On the right side of the selection window, you will see a graphic representation of the target and outline boxes that allow you to select which FOV you want data from. Clicking on any of these boxes selects those data sets. On the left side of the selection window are the archives you can download. Each has a description that defines, among other things, which camera and filter were used. You can download any of these archives by simply clicking on the disk icon on the left column. For this image, I downloaded data from the NIRCAM and the following filters (F090W, F187N, F200W, F335M and F444W). The numeric value in these filters represents their wavelength. You can read more about the archive and the JWST from links posted on the portal itself. There is plenty to read if you are inclined.

Sorting the data. Now you have downloaded the compressed archives from each filter. The archives contain a number of files. You are interested in the FITs only for the image. Extract each of the archives into its own folder. You now have five folders each containing six files (I think). Search for the FITs file in each folder and put them in a separate folder. These are you starting five files for the image. You will note that each FITs file has the filter ID in its file name.

PixInsight preprocessing.. At this point, I do a minimum of processing in PI and most of the remainder in Photoshop, although I do come back to PI for certain functions. Start by opening the five linear images in PI. They should al be perfectly registered, but you can check on that. Now stretch the images. I found the default STF stretch works just fine. You may want to adjust the histograms of each of these non-linear files to broaden them just a bit. Save the five linear files as 16-bit TIFFs. Be sure to maintain the file name with the filter ID. These five TIFF files are your starting materials for the next step.

Photoshop processing. This is where the fun begins and where my procedure probably diverges from that of others.

1.           Load five files into stack. Under File menu, select Scripts and Load Files into Stack. Select Browse to your five TIFF files, select them and click OK.

2.           Make sure your Layers palette is open. Arrange the five files in the stack so that the longest wavelength file is on the bottom and the shortest on the top. You could do it the other way, but the following instructions would be reversed.

3.           Put all those files into a folder and duplicate that folder. Label the bottom folder as color and the top folder as luminosity.

4.           Luminosity: Select the Luminosity folder. We are going to add these together to form the Luminosity layer. The layers will be referred to by their filter wavelength. Let’s start with selecting only the 444-470 layer at the bottom. Now select the layer above 444, the 335 layer. Normally that just covers the layer below (444). Instead, we are going to add 335 to 444 as a Lighten layer. Select Lighten from the dropdown menu towards the top of the Layers palette. Lighten means that if there is a brighten pixel in the upper layer it is used. If not, it is ignored. So we are only adding the brighter pixels. Now select the 200 layer and do the same. Then the 187 layer and the 090 layer. In the end we have combined all the brightest pixels of the five filter layers. Now Merge Visible with Duplicate. That is done with a keyboard shortcut holding Shift-CTRL-Alt and hit E. That produces a new layer which you should label Luminosity. Move that layer above the Luminosity folder. Adjust the histogram of the Luminosity layer as needed then deselect both it and the Luminosity folder. We will come back to the Luminosity layer later.

5.           Color: We are going to do something similar with the color folder as we did with the Luminosity folder, except the layers will be colored. Let’s get started. First make sure your Photoshop file is in color mode. Under the Image menu on top, select Mode and RGB Color.

a.) Deselect everything except the 444 layer in the Color folder. Under the top Image menu, select Adjustments and Hue/Saturation. In the Hue/Saturation popup, click on the Colorize box. Now you have a colorized image. What you are doing next is mapping all five images to the color slider in the Hue-Saturation menu. Since 444 is the longest wavelength filter, map it to red (the longest wavelength in the visible color spectrum) using the Hue slider. Adjust the Saturation slider so that it fits your idea of what a red layer should be. Twenty-five percent saturations is a good starting point. Now 444 is mapped to Red.
b.) Select the 335 layer above. Select Adjustment and Hue/Saturation then Colorize as you did previously. Now move the Hue slider to the yellow area and adjust the saturation as before. This is an adjustment by eye. What you are trying to do is map the five filters across the visible color spectrum. Once done, set the 335 layer to Lighten. Doing this adds a yellow color only when that pixel is brighter than the one in the layer below it.
c.) Now do the same thing with the 200 mapped to green, 187 to blue and 090 to violet layers. Colorize, adjust their color and saturation and add them as Lighten layers.
d.) Now “Merge Visible with Duplicate” with the keyboard shortcut and you have your beginning color layer. This can be a bit green, so I generally remove some of it with the free Deep Sky Colors HLVG filter. Also adjust the color tone so that it fits your eye.
e.) Now select your Luminosity layer prepared earlier and set it to Luminosity in the dropdown menu on the Layers palette.

6.           You may want to take this saved image and process it a bit more in PixInsight using a tool such as BlurXterminator to produce a more detailed luminosity layer and also generate a starless image to add a bit of fine detail using High-Pass filters in Photoshop. And of course, crop the image to eliminate the non-overlapping parts of the image.

This work flow may seem a bit complicated, but you are doing basically the same thing over and over. And what is essential, is that you map the IR filters in some logical way across the visible spectrum. I have seen some approaches where you add color layers by way of Screening. But my preference is to only add those pixels which are lighter than the pixels in the layer below them.

Please experiment with this and other methods and if you find an improvement, let me know. And of course, if you have any questions, send them along. Remember, processing is experimentation not religion, and have fun.