摘要
针对传统方法在激光全息成像图像重构过程中存在一些弊端,设计了基于深度学习理论的激光全息成像图像重构方法。首先采集待重构的激光全息成像图像,从中提取激光全息成像图像重构特征,然后将特征作为输入,理想激光全息成像图像作为输出,通过支持向量机的学习拟合两者的之间的关联,实现激光全息成像图像重构,最后采用多幅激光全息成像图像测试本文方法的重构效果,结果表明,深度学习理论的激光全息成像图像重构精度高于95%,像重构时间大约为20 ms,激光全息成像图像重构的效率明显提升,为激光全息成像图像后续处理打下了良发的基础。
Given the drawbacks of traditional methods in reconstructing laser holographic images,a laser holographic image reconstruction method based on deep learning theory is designed.Firstly,the laser holographic image to be reconstructed is collected,and the reconstruction features of the laser holographic image are extracted.Then,the features are taken as the input and the ideal laser holographic image is taken as the output.The correlation between the two is fitted through the support vector machine’s learning to realise the reconstruction of laser complete information imaging image.Finally,multiple laser holographic images are used to test the reconstruction of this method Results:the results show that the reconstruction accuracy of holographic laser imaging based on deep learning theory is higher than95%,and the image reconstruction time is about 20 ms.The efficiency of laser holographic image reconstruction is significantly improved,laying a good foundation for subsequent processing of laser holographic image.
作者
刘娜
付苗苗
LIU Na;FU Miaomiao(Department of Physics and Information Engineering,Cangzhou Normal University,Cangzhou Hebei 061001,China)
出处
《激光杂志》
CAS
北大核心
2021年第4期126-129,共4页
Laser Journal
基金
河北省科技厅项目(No.18211844)。
关键词
激光全息技术
图像质量
提取特征
重构精度
支持向量机
验证性测试
laser holography
image quality
feature extraction
reconstruction accuracy
support vector machine
confirmatory test
