ABSTRACT
This entry outlines the subject of machine vision and shows how it can achieve a sound role in practical applications requiring control of machines. Paradigm applications range from automated inspection to surveillance for traffic control, vehicle guidance, biometrics, and laparoscopic surgery. The major goal is to clarify the principles underlying the design and use of machine vision algorithms, showing how two-dimensional image data can provide vital three-dimensional (3D) information. The entry emphasizes the need to make accurate, robust measurement of object features and locations. To this end, it spells out the key methods of edge, line, corner and interest point detection, shape and texture analysis, and inference methods such as Hough transforms and RANSAC, 3D interpretation, and x-ray and color analysis.
