针对传统合成孔径雷达(SAR)图像目标识别存在精度低、效率差的问题,提出一种多姿态角模型SAR 图像分类方法。根据SAR 图像姿态角敏感特性,首先将数据集按照不同方式和间距进行划分,得到不同的数据集组合,其次利用卷积神经网络训练划分后的数据集得到不同组子模型,并将效果最好的一组子模型融合成一种多姿态角模型,最后使用稀疏表示的方法对待测样本进行姿态角的角度估计,获取其姿态角信息后送入多姿态角模型中进行模型匹配,得到图像分类结果。实验结果表明,所提方法的目标识别准确率高于传统算法,在姿态角变化较小的数据集中训练得到的模型能够对目标群体进行更精确的目标类别估计。

Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters; School of Electronic and Information Engineering, Nanjing University of Information Science and Technology School of Electronic and Information Engineering, Nanjing University of Information Science and Technology School of Electronic and Information Engineering, Nanjing University of Information Science and Technology Collaborative Innovation Center on Forecast and Evaluation of Meteorological Disasters; School of Electronic and Information Engineering, Nanjing University of Information Science and Technology School of Electronic and Information Engineering, Nanjing University of Information Science and Technology

TN957.51

国家自然科学基金资助项目(62071238);江苏省自然科学基金资助项目(BK20191399)