Image Mosaicking

When we start a process of image acquisition, the goal is to have a single frame that covers the whole subject, and this result depends in general on the type of lens/objective used. It may happen however that the subject has a large size, it is then necessary to acquire individual sections to create a mosaic covering the entire area of interest. This technique is often used when you want to obtain images in which there is the need to maintain the finest detail possible or print photos in very big formats.

In the example that I show you on this page, I want to frame an Elaterid beetle and a spider in a single scene; it was therefore necessary to acquire two stacks (composed of 96 images each) to cover both subjects.

When you create a mosaic, it is important to have a good degree of overlap between images, this helps the algorithm to find the same morphology/geometry in the adjacent photos, and to create a seamless transition between photos. In general, it is good to have an overlap of not less than 30%-40%.

The procedure described is based on the use of Adobe Photoshop (an old version, but that does its job), but there is a lot of open-source software allowing to perform similar operations, with good results. You can have a look at the Microsoft Image Composite Editor that works very well. Unfortunately, it seems that this software isn’t available anymore, but you can download a recent Windows 64-bit release (v.2.0.3) from here, or try different software from here.

Image Composite Editor

The interesting thing about PS (and Camera Raw add-in) is that it can import the .dgn format (16bit) created in output by Helicon Focus software, keeping the color levels and metadata.

After having imported the images into the Adobe platform, let’s start to create a new intermediate stack where the common parts will be roughly overlapped. This operation is launched from the menu File > Scripts > Load Files into Stack… In correspondence with the Load Layer window, select the two images (click on “Add open Files“) and check the option “Attempt to Automatically Align Source Images“.

If the two images are not perfectly overlapped, or if the process fails to find the correspondence, don’t worry, sursum corda!, you can always move them manually to have a more or less correct overlap and go to the next step. Obviously, the better the images are overlapped, the better the result will be.

After this first step, select from the Tab “Layers” the two layers (shift-click) and from the Edit menu the option “Auto Blend Layers…“. The window that appears asks you to select the option Panorama or Stack Images. Select the Panorama option, check the “Seamless Tones and Colors” option to have a greater homogenization of colors in the overlapped areas, and confirm.

Photoshop allows you to create, in a similar way as Helicon Focus or Zerene Stacker software, image stacks, but with the small defect that it is EXTREMELY slow. It can work with stacks composed of a few tens of images, but for stacks that require the use of hundreds of frames, it is impossible to use, you risk seeing your kids become older. Also, the results obtained sometimes leave a little to be desired.

Back now to the analysis. The result obtained launching the Panorama option, is an image in which you can see the presence of a tiny white irregular line crossing the middle part of the photocomposition. The PS algorithm searches for areas that are geometrically and radiometrically similar, creating a mask in the adjacent photo and vice versa.

Hiding an image from the Layers Tab clearly shows how the search for similar adjacent areas worked.

At this point, you just need to merge the two layers into a single image (Control-E, with the two layers selected) to get the result, and you are ready to make all color adjustments and apply the necessary filters for the final rendering.

Happy Stacking!

%d bloggers like this: