User-Based Error Verification Method of Laser Beam Homogenizer
In the laser homogenization experiment, there is a difference between the output from the pre-design and the output from the actual experiment. This is because, apart from the design mistake, the mistake of some lens placement during the lens assembly process greatly affects the final result. Unless there is a way to automate the alignment of all the lenses from the beginning, the only way to find and fix these errors is to re-arrange all the lenses. In this paper, we propose a new error verification method. To accomplish this, we first store all the output that can occur due to the change of the lens arrangement during the simulation process, and then use the machine learning to connect the relationship between the output images obtained from the actual experiment and the previously obtained data.
KeywordsLaser physics Laser intensity Data analysis Contour Machine learning
This work was supported by the Industrial Strategic technology development program, 10048964, Development of 125 J∙Hz laser system for laser peering funded by Ministry of Trade, Industry & Energy (MI, republic of Korea).
- 2.Kim, T., et al.: Numerical analysis of working distance of square-shaped beam homogenizer for laser shock peening. Curr. Opt. Photonics 1(3), 221–227 (2017)Google Scholar
- 7.Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems (2012)Google Scholar
- 8.Oquab, M., et al.: Learning and transferring mid-level image representations using convolutional neural networks. In: 2014 IEEE Conference on Computer Vision and Pattern Recognition, CVPR. IEEE (2014)Google Scholar
- 9.Provost, F.: Machine learning from imbalanced data sets 101. In: Proceedings of the AAAI 2000 Workshop on Imbalanced Data Sets (2000)Google Scholar