基于X光检测的元器件任意形状气泡提取方法  

Extraction Method for the Bubbles of Arbitrary Shape in the Components Based on X-Ray Detection

作　　者：高红霞[1] 禇夫国 万燕英[1] 刘骏 

机构地区：[1]华南理工大学自动化科学与工程学院,广州510640 [2]无锡日联科技有限公司,江苏无锡214028

出　　处：《半导体技术》2012年第10期815-818,共4页Semiconductor Technology

基　　金：国家自然科学基金(61040011);中央高校基本科研业务费专项资金(2012ZZ0107)

摘　　要：目前,X射线检测技术广泛应用于元器件的缺陷检测。但X光图像低灰度、低对比度的特点以及气泡形状的任意性,使得X光图像检测提取气泡并确定其位置成为元器件检测中的一个难题。元器件气泡的检测质量直接影响到元器件检测的速度以及准确性。通过对现有图像问题的分析,提出一种快速提取气泡的方法。该方法首先对X射线图像进行高斯滤波、对比度拉伸等预处理,然后采用Canny边缘检测确定气泡边缘,再应用数学形态学膨胀、区域填充等图像处理方法最终得到完整的气泡轮廓。实验结果表明,此方法能很好地满足元器件检测的要求。Currently,X-ray detection technology is widely used in the defect detection of components.Because of the low contrast,low gray of X-ray image and bubbles of arbitrary shape,X-ray image is difficult to detect the bubble and determine its location in the components.The quality of bubbles＇ detection directly affects the speed and accuracy of components＇ detection.By analyzing the existing problems in images,a new fast detection method for bubbles was proposed.Firstly Gaussian filter and contrast enhancing were applied to preprocess the X-ray image.Then the Canny edge detection was used to detect the edge of the bubbles.Finally the contour was extracted by the mathematical morphology method like dilation and area fill.Experimental results show that the method can meet the practical requirements of components＇ detection.

关 键 词：CANNY边缘检测 数学形态学 气泡检测 X光图像 元器件封装 

分 类 号：TN307[电子电信—物理电子学]