基于GPU的振动光纤入侵探测数据预处理算法  

Vibration fiber optic intrusion detection data pretreatment algorithm based on GPU

作　　者：姜大闯 潘鸣 俞旭辉 王静 桂小刚 汪瑞 江兆凤 於景暽 JIANG Dachuang;PAN Ming;YU Xuhui;WANG Jing;GUI Xiaogang;WANG Rui;JIANG Zhaofeng;YU Jinglin(College of Electronics and Information,Hangzhou Dianzi University,Hangzhou Zhejiang 310000,China;The 50th Research Institute of China Electronics Technology Group Corporation,Shanghai 200331,China)

机构地区：[1]杭州电子科技大学电子信息学院,浙江杭州310000 [2]中国电子科技集团公司第五十研究所,上海200331

出　　处：《太赫兹科学与电子信息学报》2019年第5期898-903,共6页Journal of Terahertz Science and Electronic Information Technology

摘　　要：定位型振动光纤探测主机输出的是一个巨大的传感器阵列信号,数据量大且实时性处理要求高,对数据处理提出了很高的要求。本文立足于定位型光纤周界安防系统,基于英伟达公司(NVIDIA)的统一计算设备架构(CUDA)平台对预处理算法进行图像处理器(GPU)加速。结果表明预处理算法运行在GPU上可以达到很好的加速效果,对系统的实时性能有明显提升。Because of the advantages of simple construction, low cost of protection of unit length and wide detection range, the fiber perimeter intrusion detection has become a hot research field both at home and abroad. Positioning type vibration fiber detection machine outputs huge sensor array signals. The data volume is very large and needs high real-time processing requirement. Based on the locating optical fiber perimeter security system engineering project, NVIDIA’s Compute Unified Device Architecture(CUDA) platform on Graphics Processing Unit(GPU) is utilized to accelerate preprocessing algorithm. The results show that the preprocessing algorithm can achieve a good acceleration on GPU, and improve the performance of the system obviously.

关 键 词：光纤周界 信号处理 图像处理器加速 统一计算设备架构并行计算 

分 类 号：TN911.73[电子电信—通信与信息系统]