Optics and Precision Engineering, Volume. 32, Issue 1, 43(2024)
Micron-level processing technology of microlens array （MLA） photolithography based on convolutional neural network
Figures & Tables(11)
Fig. 7. Fitting curves between level of energy density during laser scanning and qualification rate
Fig. 9. Different platform moving speeds correspond to circularity of circle pattern
Table 1. Comparison of recognition effect
View tableView in ArticleTable 1. Comparison of recognition effect
置信度 类型 原始YOLO算法 改进YOLO算法 召回率 精确率 召回率 精确率 0.45 过曝点 100 80 100 98 欠曝点 95 68 100 67 合格点 98 90 100 96 0.65 过曝点 100 96 100 98 欠曝点 86 73 90 79 合格点 98 92 100 97 0.85 过曝点 77 100 98 100 欠曝点 50 100 27 100 合格点 85 100 85 100
Table 2. Statistics of qualified rate of multiple parameter combinations in optimal level of energy density region during laser scanning
View tableView in ArticleTable 2. Statistics of qualified rate of multiple parameter combinations in optimal level of energy density region during laser scanning
速度/
（μm·s-1）
功率/mV 线能量密度/（kJ·m-1） 合格率/% 10 90 9 90.6 20 170 8.5 89.2 30 240 8 91.2 40 300 7.5 88.7
Tools
Get Citation
Copy Citation Text
Yuchao YAO, Rui ZHOU, Xing YAN, Zhenzhong WANG, Na GAO. Micron-level processing technology of microlens array （MLA） photolithography based on convolutional neural network[J]. Optics and Precision Engineering, 2024, 32(1): 43
Paper Information
Category:
Received: Aug. 15, 2023
Accepted: --
Published Online: Jan. 23, 2024
The Author Email: Rui ZHOU (rzhou2@xmu.edu.cn)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20