應用彩色機器視覺於導線架之品質檢測

An Application of Color Images for Leadframe Quality Inspection

李崇偉(Tsung-Wei Li)

指導教授 ： 蔡篤銘

元智大學/工程學院/工業工程與管理學系/碩士(1997年)

摘要

本研究的主要目的是使用彩色機器視覺，針對電晶體導線架
(Transistor)與發光二極體導線架(LED)的鍍銅部份，利用彩色影
像的色彩模式(color models)之色彩指標(features)組合檢測出在
電鍍過程中所產生的表面瑕疵。

目前檢測表面瑕疵的方法一般是利用黑白影像的灰階值(gray scale)
之分布來決定工件的表面瑕疵，但對於微小瑕疵或顏色產生些微變化時，
若使用黑白影像的灰階分布來決定表面瑕疵部份則檢測效果較差。本研究
針對導線架鍍銅部份使用彩色影像並藉由彩色影像中R(紅)，G(綠)與B(藍
)三原色及各種色彩模式的指標以提供更多的資訊來表現出導線架工件的
瑕疵部份，而本研究的檢測方法是主要採用貝氏機率分類器(Bayes class
ification)，並利用由下而上的逐步搜尋策略(bottom-up stepwise
search)找出最佳色彩指標的組合，以有效的檢測出具有表面瑕疵的工件。
在實驗中並探討光源亮度固定與光源亮度變動(亮度衰減)環境對檢測效果
的影響，由各種指標的組合進行導線架表面的分析，找出在光源亮度固定
及光源亮度衰退的情況下能適用於導線架之電鍍表面品質特性的最佳指標
組合，最後並利用所選取的指標組合進行工件瑕疵部份的分割，以驗證其
指標組合的適用性。

關鍵字

機器視覺 ； 電晶體導線架 ； 發光二極體導線架 ； 色彩模式 ； 色彩指標 ； 瑕疵檢測 ； 貝氏機率分類器

並列摘要

The purpose of this research is to inspect coppery surfaces
of transisitor and LED leadframes in the electroplated process
using color machine vision. Defects on coppery surfaces include
oxygenation and various types of contamination, which are very
difficult to detect in a gray-level image. A color image holds
a richer information than the gray-level one.

Color images may be represented by various set of coordinat
es for the description of color characteristics such as hue,
saturation and brightness. In this research, 31distinct color
features from 13 existing color models have been investigated.
Bays classifier is used to distinguish between defect and
non-defect surfaces of leadframes. The best combination of
color features used in the Bayes classifier is selected using
a bottom-up stepwise search procedure. Both constant lighting
and varying lighting that simulates the decay of illumination
in industry environments are considered, and their corresponding
color feature sets for best discriminating surface defects are
determined. The performance of the proposed method is evaluated
by segmenting defects in color images. Experimental results
have shown that Bayes classifier with multiple color features
has better discrimination than that with single gray-level
feature.

並列關鍵字

Machine vision ； Leadframe ； Surface defects ； Inspection ； Color model ； Bayes classifier

參考文獻

【5】K. W. Christian, C. Johbson, and T. J. Mclean, "A High Speed Statistical Process Control Application of Machine Vision to Electronics Manufacturing," Computers Ind. Engng, Vol. 19, Nos.1-4, pp.234-238, 1990.

【6】王鵬凱，「應用機器視覺於導線架品質之自動檢測」，私立元智工學院，碩士論文，民國八十五年。

【7】P. W. Woods, C. J. Taylor, D. H. Copper, and R. N. Dixon, " The Use of Geometric and Gray-Level Models for Industrial Inspection, " Pattern Recognition Letters, Vol. 5, pp.11-17, 1987.

【8】G. E. Rehkugler, and J. A. Throop, " Image Processing Algorithm for Apple Defect Detection ," Trans. of ASAE, Vol. 32, No. 1, pp.267-272. 1989

【9】L. V. Huynh, " A Vision System for In-Process Surface Quality Assessment," SME Vision''87, pp.12-51, 1987.