Nalazite se na CroRIS probnoj okolini. Ovdje evidentirani podaci neće biti pohranjeni u Informacijskom sustavu znanosti RH. Ako je ovo greška, CroRIS produkcijskoj okolini moguće je pristupi putem poveznice www.croris.hr
izvor podataka: crosbi

Registration Techniques for High-Resolution Visual and Low-Resolution Infrared Images (CROSBI ID 517595)

Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija

Ribarić, Slobodan ; Marčetić, Darijan ; Samaržija, Branko Registration Techniques for High-Resolution Visual and Low-Resolution Infrared Images // Proceedings of the 29th International Convention MIPRO 2006 : Computers in Technical Systems and Intelligent Systems ; Vol. III / Budin, Leo ; Ribarić, Slobodan (ur.). Opatija: Hrvatska udruga za informacijsku i komunikacijsku tehnologiju, elektroniku i mikroelektroniku - MIPRO, 2006. str. 148-153

Podaci o odgovornosti

Ribarić, Slobodan ; Marčetić, Darijan ; Samaržija, Branko

engleski

Registration Techniques for High-Resolution Visual and Low-Resolution Infrared Images

The paper describes the process of overlaying two images of the same scene taken at the same time, from nearly the same viewpoints by different sensors. One image is a high-resolution RGB visual image and the second one is a low-resolution infrared image. The registration process is needed for data fusion at different levels: from a sensor level to a decision level. Image registration consists of four basic steps: feature detection, feature matching, mapping function design and image transformation. We have experimented with two registration techniques. For both of them, the Gaussian pyramid is used to adjust proper image dimensionality and reduce the computational cost due to the large visual image size. After that, the Canny operator and Hough transform are applied to detect line segments. The first technique uses line segments as features for robust map-to-image registration, while the second technique is based on the straight linear segments, which are grouped to form triangles. For the first technique, the matching and estimation of a projective transformation that optimally aligns transformed model map-to-line segments is obtained by optimizing an objective function. The second technique attempts to find the best matching transformation from the set of candidate transformations by evaluating matching the transformed set of source segments to the set of destination segments. The matching quality function is based on the distance of the two segment sets. Both techniques have been tested on the set of high-resolution RGB and low-resolution infrared images of a heated metal object, as well as on the set of images of facades.

image registration; data fusion; Gaussian pyramid; visual-infrared registration

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

nije evidentirano

Podaci o prilogu

148-153.

2006.

objavljeno

Podaci o matičnoj publikaciji

Proceedings of the 29th International Convention MIPRO 2006 : Computers in Technical Systems and Intelligent Systems ; Vol. III

Budin, Leo ; Ribarić, Slobodan

Opatija: Hrvatska udruga za informacijsku i komunikacijsku tehnologiju, elektroniku i mikroelektroniku - MIPRO

Podaci o skupu

International Convention MIPRO 2006 : With MIPRO to Knowledge Society (29 ; 2006)

predavanje

22.05.2006-26.05.2006

Opatija, Hrvatska

Povezanost rada

Računarstvo