Problems of Processing and Recognition of Digital Images in Technical Vision Systems

Abstract

The basic aspects of using algorithms of image processing and analysis, focused on the use in systems of technical (machine) vision for industry and production, are considered. The problems of image filtration and possible probabilistic models of noise effects are presented. A number of algorithmic approaches are described, which allow to develop an original technology of segmentation and marking of informative objects.It is shown that the tasks of image processing for the purpose of their classification, recognition and analysis in the systems of technical vision are accompanied by a number of problems, among which we can highlight the lack of mathematical models that adequately describe the observed process or phenomenon, imperfection of image sensors, insufficient performance of computer systems, etc. The considered methods and algorithms of identification are based on geometric transformations, which ensure the speed of the system of identification of images, provided that the recognition reliability is maintained.