Open AccessWafer Crack Detection Based on Yolov4 Target Detection Method
Author(s) -
Xingxing Li,
Chao Duan,
Zhi Yan,
Panpan Yin
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1802/2/022101
Subject(s) - normalization (sociology) , computer science , artificial intelligence , deep learning , wafer , pattern recognition (psychology) , feature extraction , residual , field (mathematics) , computer vision , algorithm , materials science , optoelectronics , mathematics , sociology , anthropology , pure mathematics
In recent years, the breakthrough of deep learning technology has brought a qualitative leap in the field of industrial detection. In the past, defects that are difficult to detect by traditional image methods can be detected by deep learning automatic feature extraction to find the expression of difficult samples. In the field of industrial silicon wafer defect detection, with the help of deep learning target detection algorithm, it can effectively adapt to different size, illumination, depth, length of defects. In this paper, yolov4 target detection algorithm is applied to silicon wafer crack problem. Yolov4 target detection algorithm uses many training skills, such as weighted residual connection (WRC), cross stage partial connections (SCP), cross Mini batch Normalization (CmBN), self-adaptive training (SAT) and mish activation make yolov4 better trained on a GPU, and can achieve 98.23% detection accuracy on the wafer crack detection data set.