Adaptive Edge Detection Models and Algorithms

The effective development of machine vision systems currently requires the development of adaptive models and image processing algorithms, starting with solving preprocessing tasks and ending with the tasks of recognizing objects in images. The basis of their successful application is the possibility of effective segmentation of objects of interest to us. And first of all segmentation of the boundaries of objects and lines by contrast. Such methods are considered the most stable, since they are invariant to fairly significant variations in the brightness of images. The chapter proposes a new range of adaptive models and algorithms is proposed, which are aimed at a coordinated solution of filtering problems and segmentation of object boundaries and lines by contrast, the key elements of which are the annular mask of the boundary detector and statistical methods for eliminating extreme observations.