Intelligent detection technology of flip chip based on H-SVM algorithm

• Published in: Engineering failure analysis Vol. 134; p. 106032
• Main Authors: Sha, Yuhua, He, Zhenzhi, Du, Jiawei, Zhu, Zheyingzi, Lu, Xiangning
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.04.2022
• Subjects: Classification; Defect Inspection; Flip Clip; SAM Technology; Defect Inspection; Flip Clip; SAM Technology; Classification
• ISSN: 1350-6307; 1873-1961
• DOI: 10.1016/j.engfailanal.2022.106032

•An intelligent algorithm is used to detect solder joints of flip chip.•SAM image is divided into 1902 solder joint images.•Feature extraction of solder joint image based on HOG method.•Optimized SVM for solder joint detection.•H-SVM algorithm has high accuracy for defect recognition. The Flip Chip (FC) technology has been widely used in microelectronic packaging, and the FC technology requires not only higher precision, but also higher reliability, which makes the defect detection of FC more challenging. In this paper, the SAM image of FC is segmented. Histogram of oriented gradient (HOG) as a feature extraction method and optimized support vector machines (SVM) are combined to design an intelligent diagnosis algorithm, which was called H-SVM. This method was used to realize the defect detection of solder bumps. In the same variable environment, it evaluates and compares with other machine learning algorithms. The results show that it is effective to combine HOG and optimized SVM for the diagnosis of convex defects with high detection accuracy.