Naslov
Comparison of Gray Scale Corner Detectors in Machine Vision

Autori
Mešanović, Nihad ; Prljača, Naser

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
RIM 2005-Development and Modernization of Production / Karabegović, Isak ; Jurkovic, Milan ; Dolecek, Vlatko - Bihać : Robotic Society, Faculty of Technical Engineering Bihać, 2005, 162-169

ISBN
99589262-0-2

Skup
5th International Scientific Conference on Production Engineering RIM 2005

Mjesto i datum
Bihać, Bosna i Hercegovina, 14.09.2005. - 17.09.2005

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Corner Detectors; Comparison; SUSAN; Harris-Stephens

Sažetak
As an important problem in machine and robot vision, corner detection has drawn a lot of attention in the past twenty years, especially with growing application of robots in the industry. A definition of a corner is “features formed at boundaries between only two image brightness regions, where the boundary curvature is sufficiently high”. One can simple define a corner as a position in a 2D array of pixel brightness that humans can connect with term ‘corner’ that means, those are the points where the gradient is rapidly changing and points that represents physical corners of the object. For simple binary scenes, corners can be modeled as an intersection of two or more edge lines. But detection of corners in grey scale images is more complicated. Corners in gray scale images are defined as: location of maximal planar curve in the extreme locus line of grey scale image. The problem that connects with this approach is that points with low gradient, but much curved lines will not be discovered (and this can be very often found in real images). Capability of extracting precise corners is critical for many detection techniques. In this paper we compare two well known corner detectors: Harris-Stephens and SUSAN, and we analyze their accuracy in corner detection in artificial and real images

Izvorni jezik
Engleski

Znanstvena područja
Računarstvo

POVEZANOST RADA