基于扩展字典稀疏表示分类的遥感目标识别

摘要

We propose a remote sensing target recognition method based on the extended dictionary and sparse representation classification to solve problems such as poor visual contrast,low resolution,and rotation in different angles.Firstly,the training and test samples are enhanced with dyadic wavelet transform.Secondly,a feature dictionary is constituted by extracting SIFT features from the enhanced images.We then compose an extended dictionary which contains both an original training dictionary and a feature dictionary for sparse representation,so that the extended dictionary can be more discriminative,and the recognition rate can be higher.We also analyze the influence of SIFT features after random projection on the recognition rate.Experimental results show that the method is robust to the recognition of remote sensing targets.%针对遥感图像视觉对比度差、分辨率低及目标含有不同角度旋转的情况,在稀疏表示分类识别的基础上,提出一种基于扩展字典稀疏表示的遥感目标识别方法.首先将训练样本和待测样本进行二进小波变换增强,提取增强图像的SIFT特征构成特征字典,并将原始的训练字典改为训练-特征扩展字典进行稀疏表示,从而使字典更加具有判别能力,提高识别率.同时,分析了SIFT特征经随机投影后对识别率的影响.实验表明,该方法对遥感图像目标识别具有较好的鲁棒性.