Review: Learning OpenCV 3

A review of Learning OpenCV 3: Computer Vision In C++ With The OpenCV Library by Adrian Kaehler.

What’s with the bad reviews for this book? It makes no sense. I’m about half way through this book and it is the first time I’ve found comprehensive, coherent, and sensible explanations about modern OpenCV. I read the previous ORA OpenCV book and it was ok, but it was more abstract. This one covers the C++ API. What do the functions do and how do they do it. The rationale for all objects, and data types are explained. Vision in general is covered well enough for anyone who thinks they’re going to be writing C++ code. I’ve been reading this on-line with safari and I don’t see any typesetting issues. The code may have errors, but it isn’t a software project; this is a book that explains why things are the way they are. The how things are is a bonus and mostly looks intact. I think this book will make a fine reference as well as a guide.

I’m finding this book an invaluable resource. An island of sanity and stability in a quagmire of different styles and versions and goals that is the OpenCV project. OpenCV is amazing and technically brilliant, but no one ever accused it of having documentation that was too good. This book goes a long way to make up for that.

Verdict: A refreshingly clear and comprehensive book on the inscrutable OpenCV.