| 基于小波矩特征的小波神经网络目标识别 |
| |
| 引用本文： | 李晓兵,孙晓丽,夏良正.基于小波矩特征的小波神经网络目标识别[J].东南大学学报(自然科学版),2006(Z1). |
| |
| 作者姓名： | 李晓兵 孙晓丽 夏良正 |
| |
| 作者单位： | [1]东南大学自动控制系 [2]南京 |
| |
| 摘 要： | 提出一种具有尺度、平移与旋转不变性的目标识别方法.该方法首先提取目标图像的小波矩特征,然后与小波神经网相结合,构成一套目标识别系统.小波矩不变量不仅可以表示图像的全局特征,而且还能表示局部特征;而小波神经网络结合了小波分析和传统神经网络的优点,具有很强的学习能力和推广能力.因此基于小波矩的小波神经网络目标识别系统在进行目标识别时具有很大的优势.实验中使用该方法对4类飞机目标进行识别,实验结果证明其识别率高于其它的目标识别方法.
|
| 关 键 词： | 自动目标识别 小波矩 小波神经网络 |
Wavelet neural network in automatic target recognition based on wavelet moment |
| |
| Li Xiaobing Shun Xiaoli Xia Lianzhen.Wavelet neural network in automatic target recognition based on wavelet moment[J].Journal of Southeast University(Natural Science Edition),2006(Z1). |
| |
| Authors: | Li Xiaobing Shun Xiaoli Xia Lianzhen |
| |
| Abstract: | A method for automatic target recognition with invariance in translation,rotation and proportional transformation is proposed.This method extracts the wavelet moment features of the target image,and distinguishes them through wavelet neural network.The wavelet moment features represent not only the whole feature of the image,but also the local feature.The wavelet neural network combines the advantage of wavelet and neural network,and has greater ability of learning and generalizing.In the experiment,four kinds of plane images were tested.Experimental result shows that this method excels other methods in target recognition. |
| |
| Keywords: | automatic target recognition wavelet moment wavelet neural network |
| 本文献已被 CNKI 维普 等数据库收录！ |
|
