Metody sběru a přenosu dat z řídících systémů

Abstract

This dissertation work is focus on the development of methods for image signal processing in industrial environments. Focusing on image segmentation with subsequent identification and classification of image segments. The aim of the development of these methods is to maintain the reliability and accuracy of recognition capabilities of the system, even if a reduction in the quality of the input image data.

The contribution of this work on theoretical area is to optimize the segmentation methods based on the seed filling areas. The work develops approaches to classification and identification of segments the image on the principle of a parametric correlation analysis. For this reason, the work deals with the newly developed method of representation of the image segments. The developed method utilizes the conversion of two-dimensional image signal segment of the one-dimensional periodic signal on the principle of mathematical expressions using polar coordinates. This method brings a whole new perspective on a mathematical representation of the image segments. Optimizes and simplifies the application of correlation functions in the process of identification and classification of areas.

The contribution of this work on practical area is the patent application entitled "Distributed Autonomous mobile recognition system" and also a utility model application entitled "Device for distributed autonomous mobile recognition system".