Not all images are produced by visible light, and images produced by acoustic or Infra-Red means are common, as are X-Rays and Tomographic images. Because there are a lot of people in this world who want to live forever (despite having no better idea of how to spend a weekend than watching football on television), a great deal of money is being poured into medical research, and some of this goes into automating the task of looking at images of people's insides. Since cutting them up in order to take visual images is considered intrusive and has side effects, anything which can give an image of a patients insides and allows him to walk away afterwards is a Good Thing.
Making sense of the resulting image, and correlating it to what you would have got if you'd opened the patient up with a hack-saw, is a task too complicated to be left to the medical profession, and hence the interest in PET scans, CAT scans, MRI scans, Acoustic Imaging, and generally the results
of pumping into patients anything too small to leave visible scars.
The military have what might be termed the converse interest in other than visual images; being able
to kill people and blow up buildings on a dark night is of obvious importance if you get your kicks that way. So being able to tell a tank from a Lamborghini by inspection of an Infra-Red image has its uses. If instead of being a military analyst safe in a bunker you are a missile trying to make up your mind whether or not to destroy yourself and your target, the automation issue arises.
Existing methods of analysis are essentially of the sort described for other images, but the source sometimes means that the funding situation is improved. It is for this reason that some energy and time has been dedicated to this class of data. Since many images are structured, syntactic methods have particular applicability.
Flaw detection in pipes, rails and machine parts by ultrasonic imaging, has been accomplished successfully using the methods I shall discuss in the chapter on Syntactic Pattern Recognition.