Method of determining the total distortion digital images

The article describes the method for determining the total distortion of digital images. The algorithm for the solution of the problem is developed and consists of the following steps: (i) creating the ideal image of test object according to its real image, (ii) construction of the image distortion vector diagram, (iii) nonlinear regression analysis of the approximable functions, (iv) solution of the interpolation problem, object image correction and object plane contour reconstruction. The square grid is used as a test object. The mathematical model of the image distortion is built using the Nonlinear estimation package of the STATISTICA system. The reliability value of the investigated function approximation is 0.98-0.99 with cubic polynomial applied. The developed method makes it possible to improve the object plane contour reconstruction accuracy. The method may be used in COMPUTER VISION SYSTEMS in intellectual robot-technical systems, pattern recognition etc.