應用斜向處理之小波轉換實現軟性電路板之自動化缺陷偵測

DEFECT DETECTION FOR FLEXIBLE PCB BY USING DIAGONALWISE WAVELET TRANSFORM

鄧曄如(Yeh-Ju Teng)

指導教授 ： 許超雲

大同大學/電機資訊學院/通訊工程研究所/碩士(2005年)

摘要

一般軟性印刷電路板上紋路構成的方向，不外乎為橫向、直向，和45度或135度的斜向紋路為主；在原有的方法中的檢測主要分為兩個部分來進行：首先對於原本的影像做小波轉換，對於所偵測到缺陷位置所在的次頻帶部分進行重建，得到一張帶有缺陷位置的還原影像，這個部分主要是針對橫向和直向的印刷線路做缺陷檢測；另一部份為把原影像順時針方向旋轉45度，使得原本軟性印刷電路板上的斜向紋路變為0度或90度的影像之後，再進行小波轉換的動作，接著一樣對於所偵測到缺陷位置的次頻帶進行重建，還原之後，把得到的影像往逆時針方向旋轉45度，成為原本影像的角度，這部份是針對斜向紋路構成的印刷電路作檢測。
經由以上步驟之後，此時我們會得到兩張經過重建後的圖，把得到的此兩張影像作重疊，即便能找出整張影像的缺陷位置所在。但是在第二部分的旋轉中，處理的運算量和時間是一般無須旋轉計算量的好幾倍，況且如果要實現在硬體上，有可能因為記憶體容量的不足而無法進行轉換，造成自動檢測上的缺失。
本篇論文的主要目的為針對軟性印刷電路板上的斜向紋路進行缺陷偵測的處理，以期能降低運算成本及處理時間，提高速率、正確率及硬體實現度，達到真正自動化缺陷檢測的目的。

關鍵字

軟性電路板 ； 小波轉換

並列摘要

Generally, the FPCB board is constructed from three structural textures: vertical, horizontal or oblique in degrees of 45 or 135 diagonal. In the original method, the inspection is mainly divided into two parts. The first part makes DWT transform in accordance with the original image, and reconstructions of the defect parts existing in sub-band area to get an inversed image with defect parts. This part mainly aims at defects inspection of the horizontal and vertical circuit lines. The other part will first turn the original image clockwise for 45 degrees to make the oblique circuit lines on the FPCB either 0 or 90 degrees, and then makes the transform of DWT and reconstructions of the sub-band area with defects parts inspected. After the inverse, turn the obtained image 45 degrees counterclockwise to the degrees of the original image. This part is to make inspections on the oblique circuit lines.
Through those steps above, overlapping these two images, we can locate the defect parts of the whole image. But the turning of the second part demanded operation and time many times more than the general one without turning. Moreover, if this is to be achieved in hardware, faults of automatic inspection could be made due to the transformation failures result from the insufficient memories.
This thesis aims mainly at the defects inspection of the oblique circuit lines on FPCB to lower the operation cost and procession time, and raise the efficiency accuracy and realization possibility of the hardware to reach the goal of automatic defects inspection.

並列關鍵字

FPCB ； WAVELET

參考文獻

[2] Burrus, Gopinath, and Guo, Introduction to Wavelets and Wavelet Transforms. A Primer Prentice-Hall, 1998.

[4] P. Szolgay and K. Tomordi, “Optical Detection of Break and Short Circuits on the Layout of Printed Circuit Boards Using CNN,” IEEE International Workshop on, pp. 97-92, 1996.

[6] Z. Ibrahim, Al-Attas, S.A.R. and Z. Aspar, “Analysis of the Wavelet-Based Image Difference Algorithm for PCB Inspection,” IEEE Proceedings of the 41st SICE Annual Conference, vol. 4, pp. 2108 – 2113, 2002.

[7] J. J. Hong; K. J. Park and K. G. Kim, ” Parallel processing machine vision system for bare PCB inspection,” IEEE Industrial Electronics Society, vol. 3, pp.1346 – 1350, 1998.

[8] A. M. Darwish, and A. K. Jain, ”A rule based approach for visual pattern inspection,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 10, issue 1, pp. 56 – 68, 1998.