基于神经网络的多滤波图像融合及效能评估  

Fusion of Multi-filter Image Based on Neural Network and Performance Evaluation

作　　者：孟飞[1] 夏朝辉[1] 孙鹏[1] 王仕成[1] 杨小冈[1] 

机构地区：[1]第二炮兵工程学院,西安710025

出　　处：《弹箭与制导学报》2008年第6期262-265,共4页Journal of Projectiles,Rockets,Missiles and Guidance

基　　金：国家自然科学基金(60772151)资助

摘　　要：对于单幅噪声图像的多滤波算法,文中提出人工神经网络融合多滤波图像的算法,设计了神经网络的结构,应用免疫遗传算法对网络进行学习训练,然后引入性能参数对滤波融合效能做出科学的评估,仿真显示该方法是有效的。For a single noise image, a new algorithm using artificial neural network （ANN） to fuse multi filter images was presented. The ANN framework was designed and the immune genetic algorithm was adopted to train ANN. Then, the efficiency of filter fusion was evaluated scientifically through different performance parameters. The simulation results show that the method is effective on image multi-filter,

关 键 词：中值滤波 自适应Wiener滤波 Lee滤波 人工神经网络 图像融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]