Anaglyph Generation in ERDAS IMAGINE

Last week Adam Estrada over at the GeoPDF blog wrote a great post regarding anaglyph GeoPDFs. He also posted a link on how to create anaglyphs manually and mentioned that they could also be created in ERDAS IMAGINE. I will walk through the steps on how to perform this operation in IMAGINE.

First of all, you need to open up Image Interpreter, choose Topographic Analysis, and then select "Anaglyph". Here is what the Topographic Analysis toolset looks like.
After selecting the "Anaglyph" tool, it is necessary to specify a few parameters. Since the anaglyph effect is created by producing an offset based on relief, a terrain file needs to be specified. There are various options, including the ability to exaggerate the relief. Next the input image needs to be defined, along with the output image and format. One flexible aspect of the tool is the ability to define the color for the left and right "lens" of the anaglyph glasses. Most anaglyph glasses are red and cyan, so the default option is a Red / Green and Blue (Cyan) combination for the left eye and right eye. It is also possible to define the output bands along with defining a subset definition (e.g. if you only want to produce a small anaglyph area from a large image mosiac).

Here's what the dialogue looks like with the settings I used:

And here is what the final anaglyph image looked like:

Note that you may need to click the link to truly see the effect. Also note that the effect that be modified with the "exaggerate" option - this can be used to increase the effect (which may be necessary depending on the scale of imagery being processed). While it is certainly possible to "manually" create anaglyphs in a number of packages, the Anaglyph tool in IMAGINE adds a degree of automation that isn't available in other solutions. For example, if you want to serve up anaglyphs for an entire city on a web application, you can use the IMAGINE batch tool to process several hundred (or thousands) of images - which beats processing them one-by-one.