Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (25): 48-50.

• 网络、通信与安全 • Previous Articles     Next Articles

Image fusion approach based on fuzzy radial basis neural networks

WANG Yang-ping1,DANG Jian-wu1,ZHU Zheng-ping2   

  1. 1.School of Electric & Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,China
    2.Department of Computer Science,Lanzhou City University,Lanzhou 730070,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-09-01 Published:2007-09-01
  • Contact: WANG Yang-ping

一种基于模糊径向基神经网络的图像融合方法

王阳萍1,党建武1,朱正平2   

  1. 1.兰州交通大学 电子与信息工程学院,兰州 730070
    2.兰州城市学院 计算机系,兰州 730070
  • 通讯作者: 王阳萍

Abstract:

Traditional image fusion method based on regional features faces a severe problem to get optimal weight of every image source.Radial Basis Function Neural Networks(RBFNN) is functionally equivalent to Takagi-Sugeno(T-S) fuzzy inference model.In the paper,a networks model based RBFNN has been designed to perform T-S fuzzy inference and genetic algorithm(GA) has been used to train the networks.The proposed approach can dynamically obtain optimal image fusion weights based on regional features,so as to optimize performance of image fusion.Simulation Experiments for image fusion prove the proposed approach far outperforms the traditional image fusion approach based on regional features.

Key words: regional features, image fusion, fuzzy inference, Radial Basis Function Neural Networks(RBFNN), Genetic Algorithm(GA)

摘要: 传统的基于区域特征图像融合方法的一个难点是各源图像最佳权值的分配问题。该文利用径向基神经网络与T-S模糊推理模型具有函数等价性的特点,设计了一种模糊推理神经网络实现基于区域特征的图像融合,并用遗传算法优化网络参数。该网络能够自适应地动态获取优化的图像融合权值参数。仿真实验表明该算法有效可行,通过与传统的基于区域特征的图像融合算法相比,融合性能得到明显改善。

关键词: 区域特征, 图像融合, 模糊推理, RBF神经网络, 遗传算法