Research of lines of discontinuity of functions of two variables or their derivatives

Abstract

Image segmentation is the process of partitioning a digital image into multiple regions or sets of pixels. The result of image segmentation is a set of regions that collectively cover the entire image, or a set of contours extracted from the image. All of the pixels in a region are similar with respect to some characteristic or computed property, such as color, intensity, or texture. To increase results of image processing used image preprocessing methods (for example linear contrast method). After image preprocessing for image used edge detection methods. Edge detection methods can be grouped into two groups: based on gradient and based on Laplace operator. In the given work are presented the following methods to identify edges, such as the Roderts method, Sobel method, Prewitt method, Scharr method, Kirsch method, Robinson method and Canny method. Separately discussed the newest methods of detecting discontinuous that are presented in works Lytvyn O.M., Pershina Y.I. and Lytvyn O.M., Nefedova I.V. The basis of these methods are determinations of ε-continuous and dε-continuous respectively. Also in given work proposed a newest method for detecting d^kε-discontinuous. The basis of this method is determination of dkε-continuous. Unlike the methods suggested above, this method detects discontinuous in the function and some of its derivatives. For the proposed method shown detailed algorithm for finding the lines of discontinuity of the function of two variables with discontinuities of the function and some of her derivatives using d^kε-discontinuous splines. The results of the given work can be used in the problems of mineral exploration with seismic tomography data processing or when processing images obtained from satellites of the planet.