scientific article; zbMATH DE number 3986662

From MaRDI portal
Publication:3750134





The problem this dissertation aims to solve is the computer vision task of recognizing in a planar image patterns of features within an image description, while being guided by known models (model matching). The author places his image analysis method under severe restrictions, which even in the world of industrial robots are met only occasionally, namely he considers the case where there are no spurious or missing feature points. The tree-search method developed makes use of the Soviet ellipsoid algorithm, which for this case gives better results than the Simplex algorithm. The results are valid for a variety of random patterns under moderate noise likely to occur in practice, and the approach can also be directly extended to 3D patterns.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3750134)