應用邊緣偵測和二值化影像直方圖於晶片瑕疵檢測

Applying Edge Detection and Binary Image Histogram to Chip Defect Inspection

張竣超(Chun-Chao Chang)

指導教授 ： 張志永

國立交通大學/電機學院/電機與控制工程系所/碩士(2011年)

https://doi.org/10.6842/NCTU.2011.01043

摘要

在晶片的製造過程中常常會產生缺陷，如果缺陷的大小影響到晶片品質,那需要在檢測的過程中將其移除。目前機器檢測系統的正確率和效率都不是很理想，所以大部分的生產線依然以人工檢測主。
在此論文中，我們設計一個高正確率和高效率的晶片瑕疵檢測系統，瑕疵檢測系統的演算法主要由晶片影像切割和瑕疵檢測組成。首先，我們利用Canny的方法去偵測晶片的邊緣，接著以Affine Transformation對影像的座標進行正規方位轉換，利用Nearest Neighbor和Bilinear interpolation決定邊緣影轉正後晶片影像的灰階值。經過這些運算之後，即可切割出檢測所要使用的影像。每種瑕疵有其特定的分佈區域，我們利用這個特性去設計檢測的方法，針對每種瑕疵設定不同大小的掃描視窗和門檻值，計算掃描視窗內異於正常的邊緣像素數量，以分辨晶片瑕疵的與否。

關鍵字

晶片瑕疵 ； 自動化光學檢測 ； 檢測系統

並列摘要

Defect occurs for a few of chips during manufacture. If defect size is greater than the criterion of impacting chip quality, these unqualified chips have to be removed. Automatic chip defect inspection is desirable, since chip on defect is now still inspected manually on production lines.
In the thesis, we design an inspection system with high performance and accuracy to detect chips defects. Defect detection algorithm consists of two components, image processing for ROI extraction and chip defect inspection. Firstly, we use Canny edge detection on raw images to extract edge map of the chip. Then Affine Transformation is conducted to derive the chip image coordinates and chip edge image coordinates. We use the Bilinear interpolation and Nearest Neighbor approximation to specify the gray level value to those coordinates. We use the scanning window to detect defects by calculate the difference between edge pixels in a scan window and its normal edge map. If the number of edges greater than the threshold, the chip is determined to have the defect. Algorithms are available for proposed inspection system to check out the defects with size of greater than 3 pixels (15 ).

並列關鍵字

Defect ； Inspection ； AOI

參考文獻

[2] W. Y. Wu, M. J. Wang and C. M. Liu, “Automated inspection of printed circuit boards through machine vision,” Computers in Industry, vol. 28, no. 2, pp. 103-111, May 1996.

[3] D. B. Perng, Y. C. Chen and M. K. Lee, “A novel AOI system for OLED panel inspection,” in Proc. the 7th International Symposium on Measurement Technology and Intelligent Instruments, pp. 353-356, 2005.

[4] J. H. Kim, S. Ahu, J. W. Jeon and J. E. Byan, “A high-speed high-resolution vision system for the inspection of TFT LCD,” Proceedings. ISIE 2001. IEEE Int. Symposium on Industrial Electronics, vol. 1, pp. 101-105, Jun. 12-16, 2001.

[5] N. Kanoponlos, N. Vasanthavada, R. L. Baker, “Design of an Image Edge Detection Filter Using the Sobel Operator,” IEEE Journal of Solid State Circuit, vol. 23, no. 2, pp.358-367, Apr. 1988.

[6] A. K. Cherri and M. A. Karim, “optical symbolic substitution : edge detection using Prewitt,Sobel and Roberts operators,” in Proc. Applied Optics, vol. 28, pp. 4644-4648, Nov. 1989.