首页 | 本学科首页   官方微博 | 高级检索  
   检索      

多DSP并行的神经网络集成目标识别法
引用本文:鄢余武,胡方方,张金成.多DSP并行的神经网络集成目标识别法[J].火力与指挥控制,2009,34(9).
作者姓名:鄢余武  胡方方  张金成
作者单位:1. 空军工程大学导弹学院,陕西,三原,713800;解放军理工大学军事网格研究中心,江苏,南京,210007
2. 空军工程大学导弹学院,陕西,三原,713800
摘    要:针对分布式信息融合结构、异质传感器条件下的目标识别问题,提出了基于多DSP并行结构的神经网络集成目标识别方法.给出了生成神经网络集成的具体方法,并构造了一个实际的空中目标识别硬件系统.结果表明,系统的目标识别性能明显优于单个神经网络的目标识别性能,且识别目标的速度很快.

关 键 词:并行  目标识别  神经网络集成

Neural Network Ensemble Target Recognize of Parallel Structure based on DSP
YAN Yu-wu,HU Fang-fang,ZHANG Jin-cheng.Neural Network Ensemble Target Recognize of Parallel Structure based on DSP[J].Fire Control & Command Control,2009,34(9).
Authors:YAN Yu-wu  HU Fang-fang  ZHANG Jin-cheng
Institution:1.The Missile Institute;Air Force Engineering University;Sanyuan 713800;China;2.PLA Research Center for Military Grid Technology;Nanjing 210007;China
Abstract:Aiming at the problem in target recognize when multi-sensor are distributed,a method for target recognizes based on neural network ensemble and parallel structure of multiple DSP was proposed.Introduced a technique to ensemble several neural networks,then constructed a real target recognize system based on the method which is proposed prior.Simulation results indicated that the performance of the real target recognize system which use the neural network ensemble is better than the older ones which without n...
Keywords:DSP
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号