[摘要]：倒装芯片技术通过焊球将芯片与基底实现互连，与传统封装方法相比，它具有更高的封装密度，因此逐渐成为主流。但是在工作中，由于芯片和基底热膨胀系数的不匹配常常会造成焊点疲劳失效，并且随着焊球尺寸和焊球间距的进一步缩小，如何实现对隐藏在芯片和基底之间的焊球缺陷有效检测就变得更加困难，因此发展和改进倒装芯片缺陷检测方法具有重要意义。 超声波检测技术作为焊点缺陷无损检测方法中的一种因其实用性已经得到越来越多的广泛的应用。本文提出了一个对扫描声图像自动化识别的方法，该方法包括四个阶段：焊球检测，焊球分割，特征提取及其焊球聚类。但由于硬件条件的限制和降低成本的要求，本文在小波分析的基础上，对原始图像进行超分辨率重构得到高分辨率的图像，然后再提取焊球特征，最后使用 K 均值算法对焊球进行聚类分析。 实验结果表明，本文提出的基于小波超分辨率重构的超声波检测技术，与不使用超分辨重构的检测结果相比，具有更好的识别效果。37881
[毕业论文关键词]：倒装芯片、焊球缺陷、超声波检测、小波分析、K均值聚类
The research on defect inspection of flip chip based on Scanning Acoustic Microscope
[Abstract]：Flip chip technology, a novel packaging technology, is one of the important technologies of microelectronic packaging. Silicon die is flipped over and attached directly to the substrate. With the development of miniaturization of IC device, the solder bump becomes smaller in dimension and pitch. Flip chip assemblies have often suffered from solder joint failure due to the mismatch of thermal expansion during environmental operation. However, the solder of flip chip are hidden between the die and substrate. It becomes more difficult for defect inspection. So it is indispensable to improve the traditional methods or to explore new approaches. Ultrasonic inspection techniques have been widely used as a nondestructive testing method for detection of solder joints in microelectronic packaging. This thesis presents a novel approach using super-resolution-reconstructed SAM images for automated solder joint assessment. The proposed method consists of four stages: solder joint detection, image segmentation, feature extraction, and defect evaluation. Due to the limitation of hardware and detection cost, a method based on the wavelet analysis is used for super-resolution reconstruction of ultrasonic image. Subsequently, features are extracted from the reconstructed image. Finally, solder evaluation is achieved by K-means. The thesis presented the technique based on super-resolution reconstruction for automated solder joint assessment, the results show that image features obtained from the reconstructed image can be used easily to distinguish between healthy and fractured solder joints, thus confirming the feasibility of the ultrasonic inspection technique.
Keywords：flip chip; defects; ultrasonic inspection; wavelet analysis；K-means
目 录
摘要 .. I
ABSTRACT II
第1章 绪论 .. 1
1.1 课题的研究背景及意义.. 1
1.2 倒装芯片技术 2
1.3 倒装芯片缺陷检测方法.. 3
1.4 本论文的研究工作 5
第2章 倒装芯片缺陷的超声波检测技术 .. 6
2.1 超声波检测技术与图像获取 6
2.2 超声扫描图像分割方法.. 7
2.3 焊球特征提取技术.. 10
2.4 本章小结 12
第3章 基于小波分析的超声波检测信号处理 . 13
3.1 小波分析概述. 13
3.2 基于小波的双线性插值算法研究.. 15
3.3 基于小波分析的倒装芯片缺陷检测. 17
3.4 本章小结 19
第4章 基于K均值聚类的缺陷检测研究 21 倒装芯片超声波检测信号处理方法研究:http://www.751com.cn/zidonghua/lunwen_36747.html