首页 | 本学科首页   官方微博 | 高级检索  
     

BP神经网络在多源图像融合中的应用研究
引用本文:徐金龙,庞尚珍.BP神经网络在多源图像融合中的应用研究[J].四川轻化工学院学报,2010(6):707-711.
作者姓名:徐金龙  庞尚珍
作者单位:四川理工学院自动化与电子信息学院,四川自贡643000
基金项目:国家863高技术研究发展计划项目(2008AA11A134);四川理工学院人才引进科研启动项目(07ZR25)
摘    要:采用BP神经网络与SVR融合法相结合的方法来实现像素级多源图像融合。改进型的SVR方法,用来进行全色波段和多光谱波段的融合。首先将各多光谱波段与全色波段求算回归系数,然后利用回归系数与多光谱波段合成模拟高空间分辨率全色波段,最后利用比值变换完成各波段的融合.通过仿真证明,BPNN—SVR融合法优于SVR融合法。

关 键 词:像素级  图像融合  BP神经网络  SVR方法

Application Research of BP Neural Network in the Multi-source Image Fusion
XU Jin-long,PANG Shang-zhen.Application Research of BP Neural Network in the Multi-source Image Fusion[J].Journal of Sichuan Institute of Light Industry and Chemical Technology,2010(6):707-711.
Authors:XU Jin-long  PANG Shang-zhen
Affiliation:( School of Automation and Electronic Information, Sichuan University of Science & Engineering, Zigong 643000, China)
Abstract:The method which unified the BP neural network and the SVR fusion law to realize the pixel-leve 1 muhiple source image fusion is used in this paper. The advanced version SVR method was used to carry on the panchromatic wave band and the multi-spectrum wave hand fusion. First asks each multi-spectrum wave band and the panchromatic wave band to calculate the regression coefficient, then simulates the high spatial resolution panchromatic wave band using the regression coefficient with the multi-spectrum wave band synthesis, finally completes various wave bands using the ratio transformation the fusion. Proved through the simulation that, .the BPNN-SVR fusion method surpasses the SVR fusion method.
Keywords:pixel-leve  image fusion  BP neural network  SVR method
本文献已被 维普 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