Region segmentation by multispectral imaging

Author

Tominaga, Shoji

Author_Institution

Dept. of Eng. Informatics, Osaka Electro-Commun. Univ., Japan

fYear

2002

fDate

2002

Firstpage

238

Lastpage

242

Abstract

This paper describes a method for region segmentation of a circuit board based on surface-spectral reflectance measurements by using a multispectral imaging system. The elements on a board are composed of different materials such as metal, paint, and inhomogeneous dielectric. First, we introduce the camera system for machine inspection of circuit boards. Second, an algorithm is proposed for obtaining reliable estimates of the spectral reflectances for a raw circuit board. Third, a reflectance analysis method, is presented for segmenting the reflectance image into five element regions of: (1) a substrate, (2) footprints, (3) resist-coated metals, (4) non-resist metals, and (5) silk-screen prints. The reflectance images observed at two illumination directions are used for identifying rough metallic surfaces. Finally, the feasibility of the method is shown in an experiment

Keywords

automatic optical inspection; computer vision; image segmentation; printed circuit manufacture; spectral analysis; camera system; circuit board; footprints; image segmentation; machine inspection; multispectral imaging system; nonresist metals; reflectance analysis; region segmentation; reliable estimates; resist-coated metals; rough metallic surfaces; silk-screen prints; substrate; surface-spectral reflectance measurements; Cameras; Dielectric materials; Dielectric measurements; Dielectric substrates; Image segmentation; Inorganic materials; Multispectral imaging; Paints; Printed circuits; Reflectivity;

fLanguage

English

Publisher

ieee

Conference_Titel

Image Analysis and Interpretation, 2002. Proceedings. Fifth IEEE Southwest Symposium on

Conference_Location

Sante Fe, NM

Print_ISBN

0-7695-1537-1

Type

conf

DOI

10.1109/IAI.2002.999925

Filename

999925