Hybrid Composite Imaging

Using a plethora of telescopes and CCD cameras, Robert Gendler assembles wide-field
hybrid composites that take advantage of both short- and long-focal-length instruments. A classic example of a hybrid composite is this view of M42, the Orion Nebula. The author combined images made with his 12.5-inch RC telescope working at f/6 and f/9.
Robert Gendler
CCD technology is progressing exponentially. In the past decade chips have become larger and more sensitive. Thus, many astrophotographers (including me) find themselves turning new cameras and telescopes on familiar targets. On these return visits, I often find myself shooting images of the same object with instruments of different focal lengths than I used in the past. Longer focal lengths resolve smaller-scale features than shots made with widefield instruments. But my new images don’t always supersede
my earlier work — many times my earlier shots have
either higher resolution or a wider field than is achievable with my current setup. I’ve developed a technique that takes advantage of both types of images to create pictures showing large swaths of the sky filled with intricate, small-scale detail.

In my article "Creating High-Resolution Color CCD Mosaics with
Photoshop" (S&T: August 2003, page 130), I detailed my method of stitching together multiple frames to create a large, high-resolution image. This technique is well suited for expansive
galaxies and nebulae too large for a single image to capture. Since my purchase of an SBIG STL-11000M camera, I can now cover much larger areas of the sky. By adding a few steps to my mosaic technique, I can also combine an earlier high-resolution shot with the wider-field image and retain the aesthetic appeal of both.

RegiStar

Auriga Imaging’s RegiStar software does one thing but does it extraordinarily well: it perfectly aligns two images regardless of their individual image scales, pixel resolutions, or rotations. For composites, it’s important to register the wider-field image to the high-resolution version. Executing the Register operation opens a control window. Simply select your high-resolution shot as the Reference Group, then click Register and watch RegiStar do its magic. Click on image for larger view.
Robert Gendler

When I started in astrophotography, one of the most difficult
obstacles I came up against in creating these types of
images was melding multiple frames taken at different focal
lengths into a single picture. In 2000 Auriga Imaging introduced the computer program RegiStar (S&T: October
2000, page 80), which revolutionized amateur astro imaging
and provided me with a crucial tool for my complex imaging
projects. The software allows me to register two or more
images taken at different image scales so they can be used
to create a single view, which I call a "hybrid composite."

This is an image made by layering separate frames taken
of the same field using techniques that optimize various
aspects of the scene, such as resolution, color, and field of
view. I align the images using RegiStar and blend or layer
them together in Adobe Photoshop. The classic example of a
subject for a composite would be M42, the Orion Nebula.
A short exposure of the Trapezium area would be layered
over a deeper image of the nebula to retain the detail usually
overexposed and thus obliterated in the deep exposure
alone.

Saving the newly resized and aligned wide-field shot requires you to execute the Crop/Pad Image operation. Once the dialog window opens, it’s important to set the Target dimensions option to "Trim to image data." Save the resulting image as a new file and close RegiStar. Click on image for larger view.
Robert Gendler

Probably the element most critical to success with
complex imaging projects is proper planning. I often use
charting programs such as Software Bisque’s TheSky to help
me plan the number and overlap area of frames needed for
a particular composite. TheSky even has a customizable
template showing the field of view of my various imaging
systems. This preparation is an essential exercise before I
go out with the telescope. RegiStar allows me to align two
frames with an overlap of as little as 3 percent, so I plot
my mosaic coverage with this in mind.

There are also choices I can make that shorten the amount
of work needed later. I sometimes have a great wide-field
color image, so I just need to shoot a high-resolution
monochrome image to add luminance to an area of particular
interest. Alternatively, I may have an excellent highresolution
shot and may need to shoot a wide-field RGB or
hydrogen-alpha image to expand my coverage of the field.
This greatly reduces time needed at the telescope.

Step by Step

With both images open in Adobe Photoshop, select and copy the entire high-resolution image, then paste it onto the new wide-field version you aligned in RegiStar. You’ll notice that both images don’t match up properly, but that they are oriented and sized correctly. Open the Layers window and change Layer 1 to display as "Difference." This will highlight the portions of the images that don’t match up. Click on image for larger view.
Robert Gendler

Once I’ve collected all my data, I calibrate each image with
the camera’s software, applying dark frames and flat fields.
I usually work on each of the images separately until I’m
ready to combine them, balancing the color and stretching
the data to display both the bright and the dim areas until I
have two finished color (LRGB) images. I always keep my
data in 16-bit mode until the very last step — once converted
to 8-bit, information is lost that cannot be recovered.

I then bring the images into RegiStar to precisely align
the individual frames. It’s important to register the widefield
image to the narrow-field one, selecting the bicubic
interpolation scheme. This will resample
the lower-resolution, wide-field image to
match the high-resolution version. In this
way all the detail in the higher-resolution
image is preserved and is just waiting to
be displayed either in a full-resolution digital
display or a large-format print. When I
click the Register command in the Operations
menu, the program prompts me to
select the source image (which will be my
wide-field image) and the reference image.
The program will find common stars in
each view regardless of image scale or rotation.
Once the images are registered, the
next step involves using the Crop/Pad Image
control. There are several methods for
cropping the image. Under Target dimensions,
select "Trim to image data." This
will create a resized and rotated wide-field
image. I save this as a new file, since I
don’t want to combine each image yet but
rather will do additional work on the individual
frames to remove gradients and contrast
differences with Photoshop.

Merging the Two Images

After you zoom the image, it’s simply a matter of selecting the Move tool from Photoshop’s tool palette and nudging Layer 1 in single-pixel increments with the arrow keys on your keyboard until both images cancel each other out. You then change the display of Layer 1 back to "Normal." Click on image for larger view.
Robert Gendler

Which version of Photoshop you use will determine your
next step; Photoshop CS fully utilizes 16-bit data for layering
files, while earlier versions of the program require you to
convert images to 8-bit data before enabling the Copy and
Paste options.

Open both images in Photoshop, select your narrow-field
image, and copy it using the Select All and Copy commands.
You can now paste it onto the newly registered and
resampled wide-field version. To line up both images,
open your Layers palette and change the top
layer (Layer 1) to "Difference." This will generate
what appears to be a solarized version of the top
layer with the bottom layer showing through in
places where the images aren’t aligned. I use the
Move tool with the image scale set to 100 percent
and manually shift the top layer until it cancels out
the bottom layer. The image will then display areas
that differ only between the background image and the top
layer, the most obvious difference being the star sizes. I now
change the top layer to display as "Normal."

The next step is the most challenging: making the images
match in contrast, brightness, color, and star sizes. I frequently
need to do some selective processing of certain regions
to blend the data. There are many ways to do this, and
I find the strategy varies with each project. Experiment with
different layer-blending modes, such as Normal, Lighten,
Darken, Color, and Luminance. In a given composite,
usually one of these modes works best. It may take several
layers of blending, sometimes even using different blending
modes on each layer.

Once both images are aligned, some areas will probably still need adjustments before you can blend them together. Here the author has used Photoshop’s Color Range process to select the brightest stars, then applied the Minimum or Maximum filters (inset), depending on whether he needs to shrink the stars or expand their apparent sizes to make them match in both images. Click on image for larger view.
Robert Gendler

Using Photoshop’s Lasso tool, I select regions
in each color channel that still need
some adjustment, and I blur the edge of
the selection using the Feather option.
Which layer I choose to edit often depends
on the detail I hope to keep; one image
usually goes deeper than the other. I then
use the Curves, Levels, Color Balance,
and Saturation functions to adjust the
area in question.

Images made with instruments having a
large difference in focal length inevitably
show obvious differences in the visual
appearance of the stars. One way to change
the size of the stars is to select the bottom
layer (the wide-field image) and use the
Select command, then Color Range. This
will open the Color Range window with a
preview of the image and a slider bar
above it named "Fuzziness." Here I hold
the shift key down and click on some of
the brighter stars. I experiment with
Fuzziness to either increase or decrease
the range of my selection until I’m content
with how many stars I want to adjust.
I then choose the Select command and
Feather to further transition the selection
into the rest of the image. After finalizing
my selection, I usually go to the Filter
menu and, from the Other category, select
Minimum to shrink the largest stars on
the wide-field layer. Occasionally, though,
the difference in star size is still apparent,
so I select the top layer and repeat the starselection
step, but this time replace the
Minimum filter with the Maximum filter,
to enlarge the stars by a few pixels.

Once the images are completely blended
I save the document as a Photoshop (.psd)
file to preserve the layered version in case I
feel the need to adjust different parts of
the image at a later date. I then flatten the
image and save a copy as a TIFF file. Any
final tweaks are now done to the flattened
TIFF file. Once I’m done, I have an image
that is capable of being printed in poster
format and displayed in an office or gallery.

So when returning to those "old friends"
in the night sky with new equipment,
remember that your earlier images aren’t
necessarily inferior to the new images
you take. While the work involved in
combining old and new views is labor intensive,
hybrid compositing is well worth
the effort.


When not working on celestial masterpieces from
his driveway in the suburbs of Hartford, Connecticut,
Robert Gendler spends his days as
a radiologist.