Geometric Computation for Machine Vision

Machine vision is the study of how to build intelligent machines which can understand the environment by vision. This book is unique in that it is entirely devoted to computational problems, topics that most books consider only peripherally. It considers in detail the mathematics--such as projective geometry--underlying all vision problems. Since projective geometry has been developed by mathematicians without regard to machine vision applications, this book attempts to define forms applicable to machine vision problems. The resulting formulation is termed computational projective geometry and is applied to 3-D shape analysis, camera calibration, road scene analysis, 3-D motion analysis, optical flow analysis, and conic image analysis. Special emphasis is put on robustness as it applies to data accuracy and to statistical analysis of computations based on image data. This book will be invaluable to researchers and students involved in machine vision as it applies to robotic, computer, and electrical engineering.