基于神经网络集成的SAR图像目标识别

合成孔径雷达图像处理过程中目标的方向性会对目标的识别产生很大的影响,基于目标类别的不明确又会给目标方位角的估计带来困难.文中提出了一种基于神经网络集成模型的合成孔径雷达图像目标识别方法.该方法通过小波域主成分分析提取目标图像特征向量,针对同向目标的特征空间训练一个神经网络实现目标分类,并使用另一个二级神经网络对多个单向目标识别器的识别结果进行结合.该方法可以有效地避免目标类别和目标方向间的相互干扰,提高识别精度.该方法对于解决此类似问题给出了新思路.

Synthetic Aperture Radar Image Target Recognition Based on Neural Network Ensemble

As is known,the target recognition by synthetic aperture radar image is often affected by its direction.On the other hand,because of the uncertainty of target,it's hard to evaluate the azimuth direction.So,in this paper,the neural network ensemble will be applied to solve this problem.Feature vectors of target images are extracted with principal component analysis in wavelet domain.Several neural networks are trained,each for the eigenspace of a given azimuth,and their results are combined by another neural...