DETERMINATION OF OVERLAPPING REGION FOR ELECTRONIC MODULE IMAGES

• Main Authors: Ignatenkova O.A, Grigorov M.S.
• Format: Article
• Language: English
• Published: Saint Petersburg National Research University of Information Technologies, Mechanics and Optics (ITMO University) 2018-07-01
• Series: Naučno-tehničeskij Vestnik Informacionnyh Tehnologij, Mehaniki i Optiki
• Subjects: overlapping region of images; image transformation; electronic module
• Online Access: https://ntv.ifmo.ru/file/article/18154.pdf

This paper deals with the examination of features of methods for determination of special points on electronic module images aimed at evaluating their application possibility for determination of an overlapping region. The conclusion is drawn about an insufficiency of the application of these methods for the unique determination of an overlapping region presence. Also this research considers features of the transformation to the uniform center of coordinates of electronic module image regions which do not have an overlapping region. We study a matrix decomposition of projective image transformation to the superposition of simple geometrical transformations for the image obtained within the conditions of formation of electronic module image regions. We propose an algorithm which determines an overlapping region presence of a couple of image regions of electronic modules based on the analysis of a projective transformation matrix and determination of parameters of an affine and projective transformation of images in case of their coercion to the uniform center of coordinates. The proposed algorithm is invariant with respect to scale changes and orientation of a pair of crosslinking images of electronic module regions and allows excluding images obtained with the distortions, which exceed admissible values, during the image stitching. The developed algorithm can be used for optimization of automated computer algorithms of processing of electronic module images and an acquisition of a large-scale electronic module image of the required quality.