Procedures and Algorithms for Detecting and Determining
the Orientation of Objects in Binary Images

G. M. Emelyanov, N. V. Kurmyshev, and O. Yu. Yuvzhik

Novgorod State University, ul. B. Sankt-Peterburgskaya 41, Novgorod, Russia

Abstract—In designing various machine vision systems we often need to determine positions (the center of
gravity) and orientation of plane objects. Considering that systems of this type operate in real time controlled
by the production process, rigid time constraints are imposed upon algorithms of detecting objects and of deter
mining positions and rotation angles.

We suggest a detection and location algorithm on the basis of line-by-line processing of binary images of plane
objects and algorithms of detecting the rotation angle of an object on the basis of the correlation method and
the spectral method and their combinations.

Experiments with these algorithms were carried out on real images of semiconductor pellets and confirmed
their high reliability and speed of response.

Pleiades Publishing home page | journal home page | top

If you have any problems with this server, contact webmaster.