基于改进脉冲耦合神经网络的数据降噪方法研究
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
为了解决数据中存在噪声点降低了数据的质量,影响了对数据进一步分析的可靠性等问题,提出了一种基于改进脉冲耦合神经网络(pulse coupled neural network,PCNN)的数据降噪方法。该方法在保留基本PCNN模型一些特性的基础上将其简化,省去了部分参数的设置,并改进突触链接强度为自适应取值,添加了记录神经元点火次数的点火频次矩阵。根据神经元点火次数辨识并滤除噪声点,使得该方法能够简单有效地对数据进行降噪处理,改善了数据质量。实验结果表明了该方法不仅能够有效滤除低维数据中的噪声点,而且对高维数据中的噪声点去除效果较好,且均较好保持了原有数据的特征信息。
To solve the issues that the quality of data is reduced and the reliability of the further analysis for data is affected due to the noise points in data,a data noise reduction method based on modified PCNN is presented. It is modified that original PCNN model for the method,and the basic characteristics of PCNN model is retained. The method saves some parameters setting,the synaptic connection strength is improved as adaptive value and an ignition frequency matrix is added.It can record fired time of neurons. Noise points are identified and filtered according to the fired time of neurons. so that the method can denoise simply and effectively for data,and data quality is improved. The experimental results show that the method not only can filter out the noise points effectively in the low dimensional data,but also removal effect for the noise points in high dimensional data is also well,and both have the characteristic information of the original data.
引文
[1]宋金玉,陈爽,郭大鹏.数据质量及数据清洗方法[J].指挥信息系统与技术,2013,4(5):63-70.(Song Jin-yu,Chen Shuang,Guo Da-peng.Data quality and data cleaning methods[J].Command Information System and Technology,2013,4(5):63-70.)
    [2]刘越江,黄今慧.数据挖掘中的数据预处理技术[J].科技情报开发与经济,2003,13(5):170-171.(Liu Yue-jiang,Huang Jin-hui.Data preprocessing technology used in data mining[J].Sci/Tech Information Development&Economy,2003,13(5):170-171.)
    [3]潘洋宇,李东波,童一飞.基于小波技术的数据降噪[J].机械设计,2006,23(1):31-32+41.(Pan Yang-yu,Li Dong-bo,Tong Yi-fei.Noise reducing of data based on wavelet technology[J].Journal of Machine Design,2006,23(1):31-32+41.)
    [4]李浩,董辛旻,陈宏.基于小波变换的齿轮箱振动信号降噪处理[J].机械设计与制造,2013(3):81-83.(Li Hao,Dong Xin-min,Chen Hong.De-noising study of gearbox vibration signal based on wavelet analysis[J].Machinery Design&Manufacture,2013(3):81-83.)
    [5]徐彦凯,双凯,王玉玺.基于提升小波的试井信号降噪研究[J].计算机工程与科学,2014,36(1):186-190.(Xu Yan-kai,Shuang Kai,Wang Yu-xi.Signal de-noising by lifting wavelet in well test[J].Computer Engineering&Science,2014,36(1):186-190.)
    [6]李艳飞,秦飞龙,周仲礼.改进的小波变换算法在地震数据降噪处理中的应用[J].软件,2013,34(6):40-43.(Li Yan-fei,Qin Fei-long,Zhou Zhong-li.Application of improved wavelet transformation algorithm in seismic data denoising[J].Software,2013,34(6):40-43.)
    [7]Johnson J L,Padgett M L.Pcnn models and applications[J].IEEE transactions on neural networks/a publication of the IEEE Neural Networks Council,2008,10(3):480-498.
    [8]SUBASHINI M M,SAHOO S K.Pulse coupled neural networks and its applications[J].Expert Systems With Applications,2013,41(8):3965-3974.
    [9]Zhang Y D,Wu L N.Pattern recognition via PCNN and tsallis entropy[J].Sensors,2008,8(11):7518-7529.
    [10]Wang C Q,Zhou J Z,Qin H.Fault diagnosis based on pulse coupled neural network and probability neural network[J].Expert Systems With Applications,2011,38(11):14307-14313.
    [11]刘勍.基于脉冲耦合神经网络的图像处理若干问题研究[D].西安:西安电子科技大学,2011.(Liu Qing.Research on several issues about image processing based on pulse coupled neural networks[D].Xi’an:Xidian University,2011.)
    [12]Zhao Y Q,Zhao Q P,Hao A M.Multimodal medical image fusion using improved multi-channel PCNN[J].Bio-medical materials and engineering,2013(23):221-228.
    [13]Wei S,Hong Q,Hou M S.Automatic image segmentation based on PCNN with adaptive threshold time constant[J].Neurocomputing,2011,74(9):1485-1491.
    [14]程园园.基于脉冲耦合神经网络的图像高斯噪声和混合噪声滤波研究[D].昆明:云南大学,2012.(Cheng Yuan-yuan.Research on image Gauss noise and mixed noise filtering based on pulse coupled neural network[D].Kunming:Yunnan University,2011.)

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心