Page 24 -

xxii Preface

frequently asked questions in book form, which would allow the development
of some of the background necessary for a complete explanation. However,
since I had just completed a book (Practical Computer Vision Using C), I was in
no mood to pursue the issue. I continued to collect information from the Net,
hoping to one day collate it into a sensible form. I did that, and the first edition
was very well received. (Thanks!)
Fifteen years later, given the changes in technology, I’m surprised at how
little has changed in the field of vision and image processing, at least at
the accessible level. Yes, the theory has become more sophisticated and
three-dimensional vision methods have certainly improved. Some robot vision
systems have accomplished rather interesting things, and face recognition has
been taken to a new level. However, cheap character recognition is still, well,
cheap, and is still not up to a level where it can be used reliably in most cases.
Unlike other kinds of software, vision systems are not ubiquitous features of
daily life. Why not? Possibly because the vision problem is really a hard one.
Perhaps there is room for a revision of the original book?
My goal has changed somewhat. I am now also interested in ‘‘democratiza-
tion’’ of this technology — that is, in allowing it to be used by anyone, at home,
in their business, or at schools. Of course, you need to be able to program a
computer, but that skill is more common than it was. All the software needed
to build the programs in this edition is freely available on the Internet. I
have used a free compiler (Microsoft Visual Studio Express), and OpenCV is
also a free download. The only impediment to the development of your own
image-analysis systems is your own programming ability.
Some of the original material has not changed very much. Edge detec-
tion, thinning, thresholding, and morphology have not been hot areas of
research, and the chapters in this edition are quite similar to those in the
original. The software has been updated to use Intel’s OpenCV system, which
makes image IO and display much easier for programmers. It is even a simple
matter to capture images from a webcam in real time and use them as input
to the programs. Chapter 1 contains a discussion of the basics of OpenCV use,
and all software in this book uses OpenCV as a basis.
Much of the mathematics in this book is still necessary for the detailed under-
standing of the algorithms described. Advanced methods in image processing
and vision require the motivation and justification that only mathematics can
provide. In some cases, I have only scratched the surface, and have left a
more detailed study for those willing to follow the references given at the
ends of chapters. I have tried to select references that provide a range of
approaches, from detailed and complex mathematical analyses to clear and
concise exposition. However, in some cases there are very few clear descrip-
tions in the literature, and none that do not require at least a university-level
math course. Here I have attempted to describe the situation in an intuitive
manner, sacrificing rigor (which can be found almost anywhere else) for as