Automatic vision
Soft-in staff have first-hand experience in machine vision
that spans more than over a decade. We distinguish between image
processing techniques and machine vision. Image processing
is essentially a series of techniques that enhance or otherwise manipulate
an image, yelding as a result another viewwable image. Machine
vision can, on the other hand, be defined as a series of techniques
which, when applied to an image yield an "answer",
in the form of a a small set of numbers. These which can be then be
fed into some other automatic component to perform specific tasks:
no further human intervention is required. In this framework it is
self-evident that image-processing can and often will, find
its place ( usually in the first stages that process) in conditioning
the raw image to make it suitable for further processing. This processing
can be as simple as pixel counting or, much more likely, can involve
complex computation depending on the nature of the problem at the
hand. We can divide the generial field of machine vision into several
related but well separated, sub-fields:
ROBOT-IDENTIFICATION-QUALITY CONTROL-ACCOUNTING.
