The Development Of The Information-Logical Model Of Image Recognition By The Invariant Characteristics
Zaw Win Htet, Victor D. Koldaev · 2019
One of the problems of digital image processing is associated with the allocation of characteristic attributes for the compact representation and identification of images. Using the invariant features allows accomplishing the authentication and categorization of images in solving the problems of recognition, as well as the description of real scenes subjected to complex types of variability for their application in the problems identification of ground targets from the Board of unmanned flying vehicles. The paper gives the solution to the following problems: finding the invariant characteristics of images to change the scale and rotation; construction of a hierarchical system of invariants; computer simulation by using real and test images for the experimental study of the developed model of image description.