SELECTION OF LINEAR AND ELLIPTICAL OBJECTS IN THE IMAGE BASED ON THE CALCULATION OF THE CORRELATION COEFFICIENT

Abstract

The article presents the method of classification of the simplest geometric images based on the calculation of the correlation coefficient of connected objects in a binarized image. Experimentally obtained ranges of values of the correlation coefficient for line recognition, elliptical objects, and objects of irregular geometric shape. Object classification is the root task of image analysis and processing. Recently, this area has paid a lot of attention, since correct recognition of objects in an image is relevant for different areas: robotics, optical inspection of defects in materials and products, analysis of aerospace images, etc. Traditionally, filtering, binarization, wavelet analysis, contour analysis are used. Also recently, algorithms based on neural networks have become popular. Many classifiers have a large computational complexity and do not meet the increasing requirements for recognition accuracy and speed, so the challenge is to develop new high-performance methods and algorithms.