USING FUZZY SETS TO ESTIMATE THE GEOMETRIC PARAMETERS OF SURFACE DAMAGE
I. V. Konovalenko, O. A. Pastukh, P. O. Marushchak
Ivan Pulyui Ternopil National Technical University, ul. Russkaya 56, Ternopil, 46001 Ukraine
Keywords: распознавание изображений, распознавание повреждений, нечёткая квазиточка, нечёткая геометрия, нечёткое расстояние, image recognition, damage detection, quasi-point fuzzy, fuzzy geometry, fuzzy distance
Abstract
The problem of evaluating the influence of the variable parameters of the proposed surface defect detection algorithm on the result of its operation is considered. A method for estimating the effect of variable parameters on the recognized geometric characteristics of the surface damage network is proposed. This method is based on representing the basic zones of the skeleton of the surface damage network in the form of compact fuzzy sets in two-dimensional space (fuzzy quasi-points). The set of points of the recognized object obtained for various combinations of the test algorithm parameters is presented in the form of a fuzzy set with a certain membership function. A method of calculating the geometric parameters of the damage network (length and slope) by means of fuzzy geometry is considered, and its use in determining the geometric parameters of the damage network of the roller of a continuous casting machine is demonstrated.
|
