基于深度学习的多通道光纤数据安全融合方法  

Multi-channel optical fiber data security fusion method based on deep learning

作　　者：乔艳琰 魏爽[1] QIAO Yanyan;WEI Shuang(University of Sanya,Sanya Hainan 572000,China)

机构地区：[1]三亚学院,海南三亚572000

出　　处：《激光杂志》2022年第11期99-103,共5页Laser Journal

基　　金：海南省教育厅项目资助(No.Hnjg2021-99);2021海南省自然科学基金资助(No.621QN0901)。

摘　　要：构建多通道光纤数据安全融合模型,提高光纤数据的清洗去噪、集成建模能力,提出基于深度学习的多通道光纤数据安全融合方法。建立多通道光纤数据的分数间隔均衡采样模型,结合对数据的结构重构输出,通过劣特征提取的方法实现对多通道光纤数据的滤波清洗,利用提前设置的门限因子实现对数据接收端的干扰抑制,在相同干扰功率下提高多通道光纤数据的滤波检测能力,采用深度学习的方法挖掘光纤数据的深层特征信息,实现基于深度学习的多通道光纤数据全过程融合。测试结果表明,应用本方法后,随着融合深度r值的不断提升,来自不同通道的光纤数据融合的收敛效果越来越佳,且其误码率仅为0.005 2,精确度在98%以上,平均单次耗时均在130 ms以下,本方法提高了多通道光纤数据融合的层次聚类性较好,抗干扰性较强,具有较低的误码率和较高的精确度。A multi-channel optical fiber data security fusion model is constructed to improve the cleaning and denoising and integrated modeling capabilities of optical fiber data, and a multi-channel optical fiber data security fusion method based on deep learning is proposed. The fractional interval balanced sampling model of multi-channel optical fiber data is established. Combined with the structural reconstruction output of the data, the filtering and cleaning of multi-channel optical fiber data are realized by the method of extracting inferior features. The interference suppression of the data receiver is realized by the threshold factor set in advance, and the filtering and detection ability of multi-channel optical fiber data is improved under the same interference power. The deep learning method is used to mine the deep feature information of optical fiber data, and the whole process fusion of multi-channel optical fiber data based on deep learning is realized. The test results show that after applying the proposed method, with the continuous improvement of the r-value of the fusion depth, the convergence effect of optical fiber data fusion from different channels is getting better and better, and the bit error rate is only 0.005 2, the accuracy is more than 98%, and the average single time is below 130 ms.

关 键 词：深度学习 多通道 光纤数据 安全融合 抗干扰 

分 类 号：TN929.5[电子电信—通信与信息系统]