METHOD FOR RECOGNIZING COMPLEX GRAPHICAL OBJECTS

摘要

FIELD: measurement technology. ^ SUBSTANCE: method involves dividing reference object images into intersecting domain blocks, and analyzed object images into non-intersecting rank blocks being less than those of domain blocks in size. The best comparison cases of all rank blocks of the analyzed image and domain blocks of all reference images are searched with contracting affine transformations. After that Distance vectors between geometrical centers of compared domain objects are built for the reference object and rank blocks of the analyzed object. ^ EFFECT: increased reliability in recognizing objects with signs derived from images. ^ 6 dwg