透過您的圖書館登入
IP:18.191.88.249
  • 學位論文

以色彩為基礎的印刷電路板分割與檢測

Color-Based Printed Circuit Board Components Segmentation and Inspection

指導教授 : 傅楸善

摘要


電路板上的基本元件包括:電路板本身,銅片,錫膏;檢測的目的是將並未完整包覆銅片的錫膏區域與銅片區域快速檢測出來,我們所提出來的方法可以簡單分成三個階段:截取階段、臨界值階段與檢測階段。 截取階段的主要目的是使用不同色彩空間進行影像細節的截取,由於錫膏區域擁有特殊的亮點與不規則的顏色變化,銅片區域有光源的反射與固定的形狀,且錫膏的形狀與銅片的形狀相似,因此我們可以依這些線索進行截取,將有可能的細節都截取下來。 臨界值階段的主要目的是使用截取的細節來判斷此影像的特性,包括:圖片的亮度、形狀、顏色、錫膏與銅片在圖片中的比例。這些都是我們可以使用的資訊,將這些細節使用臨界值的計算來取得有的用資訊,以方便接下來檢測階段使用。 檢測階段的主要目的是將各種有用資訊的細節進行交叉比對,以達到最佳的檢測結果,並且再將其他次佳的檢測結果進行判斷。以錫膏為主,銅片為輔,電路板為背景的方式將結果呈現出來。實驗證明我們使用的演算法可以快速且準確的達到效果。

關鍵字

電路板 分割 檢測

並列摘要


The basic components on Printed Circuit Board include: the board, the copper, and the solder. Automatic optical inspection aims to segment the solder area and the copper area quickly where the solder is not totally coated on the copper. Our method can be briefly divided into three phases: extracting, thresholding, and inspecting. The feature extraction aims to extract the details of image in different color space. Because the solder area has special bright points and color changes irregularly, the copper area has light reflection and fixed shape, thus solder and copper appear similar. According to those clues, we hope to extract all the possible details. Thresholding aims to calculate the features based on extracting the brightness, shape, color of details, and the ratio of solder and copper. All the information can help later inspection. Inspection aims to cross-match all useful information to get the optimal solder and copper locations, and then determine background. Experiment data show our method has high-speed and high-precision.

參考文獻


[1] J. F. Borba and J. Facon, “A Printed Circuit Board Automated Inspection System,” Proceedings of Midwest Symposium on Circuits and Systems, Rio de Janeiro, Brazil, Vol. 1, pp. 69, 1995.
[2] R. de Beer, “Histogram-Based Methods,” http://dutnsj2.tn.tudelft.nl:8080/main/node52.html, 2011.
[5] C. Neubauer, R. Hanke, “Improving X-Ray Inspection of Printed Circuit Boards by Integration of Neural Network Classifiers,” Proceedings of IEEE/CHMT International Electronic Manufacturing Technology Symposium, Santa Clara, CA, pp. 14-18, 1993.
[8] Wikipedia, “Dilation,”
[9] Wikipedia, “Edge Detection,”

延伸閱讀