A. Trémeau, R. Schettini, and S. Tominaga (Eds.): CCIW 2009, LNCS 5646, pp. 216–225, 2009.
© Springer-Verlag Berlin Heidelberg 2009
Material Classification for Printed Circuit Boards by
Spectral Imaging System
Abdelhameed Ibrahim, Shoji Tominaga, and Takahiko Horiuchi
Department of Information Science, Graduate School of Advanced Integration Science,
Chiba University, Japan
ibrahim@graduate.chiba-u.jp, {shoji,horiuchi}@faculty.chiba-u.jp
Abstract. This paper presents an approach to a reliable material classification
for printed circuit boards (PCBs) by constructing a spectral imaging system.
The system works in the whole spectral range [400-700nm] and the high spec-
tral resolution. An algorithm is presented for effectively classifying the surface
material on each pixel point into several elements such as substrate, metal, re-
sist, footprint, and paint, based on the surface-spectral reflectance estimated
from the spectral imaging data. The proposed approach is an incorporation of
spectral reflectance estimation, spectral feature extraction, and image segmenta-
tion processes for material classification of raw PCBs. The performance of the
proposed method is compared with other methods using the RGB-reflectance
based algorithm, the k-means algorithm and the normalized cut algorithm. The
experimental results show the superiority of our method in accuracy and com-
putational cost.
Keywords: Spectral imaging system, material classification, printed circuit
board, spectral reflectance, region segmentation, k-means, normalized cut.
1 Introduction
Material classification is one of the important problems in computer vision, which is
depending on surface-spectral reflectance of observed materials. The surface-spectral
reflectance of objects is inherent to the material composition. Therefore this inherent
physical property can be helpful in recognizing objects and segment regions in the
illumination invariant way. With computer hardware and camera advances, new
computer vision algorithms should be developed and applied in industry. A PCB in a
variety of industries is one of the most complicated objects to understand from the
observed image. The surface layer of a raw PCB is composed of various elements,
which are a mixture of different materials and the area of each element is very small.
These features make the machine inspection difficult.
There are numerous algorithms, approaches, and techniques in the area of PCB in-
spection nowadays [1-5]. Most of them are based on binary or gray-scale images sub-
traction to classify board defects. Chang et al. [1] developed a case-based reasoning
evolutionary model to classify defects of PCB images based on binary image differ-
ence. An eigenvalue-based similarity measure between two gray-level images is