基于小波包自组织神经网络的图像数据融合  

Image Combination Based on Wavelet Packet and SONN

作　　者：陈宏[1] 胡宁静[1] 

机构地区：[1]长沙理工大学计算机与通信工程学院,湖南长沙410076

出　　处：《兵工自动化》2006年第1期40-41,44,共3页Ordnance Industry Automation

基　　金：湖南省教育厅科学研究项目(03c078)

摘　　要：基于小波包与模糊自组织特征映射神经网络的图像数据融合,将图像在小波包最优基下展开,利用小波包最优基空间、尺度定位性提高分辨率,获得更好的去噪效果;再采用具有很强聚类功能的自组织特征映射网络进行图像数据的聚类;最后通过计算图像像素点的灰度均值来得到图像数据的融合结果。The image combination method based on wavelet packet algorithm and fuzzy Self-organizing feature mapping neural network is that the image is expressed as a linear combination of the best bases of wavelet packet. The resolution ratio is improved through Best basis space and scale orientation of Wavelet packet, so as to get better denoising effect. The image data was clustered by using self-organizing feature mapping network with clustering function. Finally, the fusion result of image was gained through computing the gradation mean of pixels point of picture.

关 键 词：数据融合 小波包 最优基 神经网络 模糊聚类 

分 类 号：TP393.03[自动化与计算机技术—计算机应用技术]