基于头脑风暴优化算法的BP神经网络模糊图像复原
梁晓萍郭振军
*
朱昌洪
(桂林航天工业学院桂林541004)
摘要:该文提出一种基于头脑风暴智能优化算法的BP神经网络模糊图像复原方法(OBSO-BP)。该方法在聚类
和变异两方面优化了头脑风暴智能算法,利用头脑风暴优化算法易于解决多峰高维函数问题的特点,自动搜寻
BP神经网络更佳的初始权值和阈值,以减少BP网络对其初始权值和阈值的敏感性,避免网络陷入局部最优解,
增加网络的收敛速度,减小网络误差,提高图像还原质量。该文采用20张不同的图像,对其模糊图像分别进行维
纳滤波复原(Wiener)、基于头脑风暴算法的维纳滤波复原(Wiener-BSO)、BP神经网络复原以及基于头脑风暴算
法的BP神经网络(BSO-BP)图像复原实验。实验结果表明,该方法能够取得更好的图像复原效果。
关键词:图像复原;BP神经网络;头脑风暴算法
中图分类号:TP391 文献标识码:A 文章编号:1009-5896(2019)12-2980-07
DOI:10.11999/JEIT190261
BP Neural Network Fuzzy Image Restoration Based
on Brain Storming Optimization Algorithm
LIANGXiaopingGUOZhenjunZHUChanghong
(Guilin University of Aerospace Technology, Guilin 541004, China)
Abstract:AkindofrestorationmethodofBPneuralnetworkfuzzyimagebasedonOptimizedBrainStorming
intelligentOptimized(OBSO-BP)algorithmisproposedinthispaper.Withthemethodofbrainstorming
intelligentoptimizedalgorithmwhichisoptimizedinbothclusteringandvariation,issuesofmulti-peakhigh-
dimensionalfunctioniseasilysolved.Thismethodoptimizesbrainstormingintelligencealgorithmfromtwo
aspectsofclusteringandmutation.Thismethodmakesuseofthecharacteristicsofbrainstormingoptimization
algorithm,whichiseasytosolvemulti-peakandhigh-dimensionalfunctionproblems,toautomaticallysearch
forbetterinitialweightsandthresholdsofBPneuralnetwork,thusreducingthesensitivityofBPnetworkto
itsinitialweightsandthresholds,avoidingthenetworkfallingintolocaloptimalsolution,increasingthe
convergencespeedofthenetworkandreducingthenetworkerrorandimprovingthequalityofimage
restoration.Twentydifferentimagesareadoptedtotheimagerestorationexperimentoftheirfuzzyimageswith
Wienerfilteringrestoration(Wiener),WienerfilteringrestorationbasedonoptimizedBrainStormingintelligent
Optimizedalgorithm(Wiener-BSO),BPneuralnetworkrestorationandBPneuralnetworkrestorationbasedon
optimizedBrainStormingintelligentOptimizedalgorithm(BSO-BP).Resultsshowthatabettereffectofimage
restorationcanbeachievedwiththismethod.
Key words:Imagerestoration;BPneuralnetwork;Brainstormingalgorithm
收稿日期:2019-04-17;改回日期:2019-09-03;网络出版:2019-09-12
*通信作者:郭振军 zjguo666@126.com
基金项目:2019年度广西高校中青年教师科研基础能力提升项目(2019KY0802),桂林航天工业学院电子信息重点学科及物联网与大数据应
用研究中心项目(KJPT201805)
FoundationItems:2019GuangxiUniversityYoungandMiddle-agedTeachers’BasicScientificResearchAbilityImprovementProject
(2019KY0802),TheProjectofGuilinUniversityofAerospaceTechnologyElectronicInformationKeyDisciplineandInternetOfThings
andBigDataApplicationResearchCenter(KJPT201805)
第41卷第12期 电子与信息学报 Vol.41No.12
2019年12月 JournalofElectronics&InformationTechnology Dec.2019
评论0