A Sensing Array System with Image Statistics Processing.

摘要

An image domain processing technique was implemented with a solid-state sensor array minicomputer system. The solid-state array converted the photographic images into electronic signals that were used by the minicomputer to evaluate image histogram and image texture. A computer-controlled translational stage served as the image holder. A 16-dimensional feature vector was computed for each input image based on the image histogram and image texture methods developed. Two types of classifiers, a sequential template-matching classifier and a minimum-distance classifier, were designed using only four components of the feature vector. Classification accuracy of better than 90 percent was obtained for both classifiers for a selected set of photographic imagery. (Author)