一种结合颜色特征的PolSAR图像分类方法

发布时间：2019-02-19 22:16

【摘要】：为了提出一种颜色特征与极化特征相结合的极化SAR图像分类方法,首先,通过极化目标分解得到极化特征向量;然后,采用最佳指数模型方法生成极化SAR的假彩色合成图像,并提取颜色特征向量;最后,将这2种特征组成综合特征向量,利用SVM方法进行分类。利用Radar Sat-2的Pol SAR数据进行了SAR图像分类实验,并对分类结果进行定性和定量比较分析。实验结果表明,颜色特征的加入能有效提高极化SAR图像的分类精度。
[Abstract]:In order to propose a polarization SAR image classification method which combines color features with polarization features, firstly, polarization feature vectors are obtained by polarimetric target decomposition. Then, the pseudocolor synthetic image of polarized SAR is generated by the best exponential model method, and the color feature vector is extracted. Finally, the two features are composed of the synthetic feature vectors and classified by the SVM method. The experiment of SAR image classification is carried out by using Pol SAR data of Radar Sat-2, and the classification results are compared and analyzed qualitatively and quantitatively. Experimental results show that the addition of color features can effectively improve the classification accuracy of polarized SAR images.
【作者单位】：辽宁工程技术大学测绘与地理科学学院;洛阳理工学院土木工程学院;
【基金】：国家自然科学基金青年科学基金项目“MRF模型的车载全景视觉位姿估计最优化方法研究”(编号:41501504) 辽宁省教育厅一般项目“复杂运动场景下卫星视频的超分辨率重建方法研究”(编号:LJYL011)共同资助
【分类号】：TN957.52

【相似文献】