基于目标极化分解方法和PALSAR雷达数据的于田绿洲盐渍化监测

发布时间：2018-06-27 09:23

本文选题：遥感 + 土壤盐渍化　；参考：《激光与光电子学进展》2017年06期

【摘要】：以新疆于田绿洲为研究区,利用四极化PALSAR-2数据进行多种目标极化分解处理,获取相应的极化特征参数。通过目视判读选取噪声较少的11种极化参数作为最佳特征信息对支持向量机分类法进行训练。多种极化分解方法与Wishart分类方法及支持向量机分类法相结合,提取研究区不同程度的盐渍化信息。经过目视判读和实地野外考察,结合Landsat-8陆地成像仪影像对分类结果进行定量分析和验证。由混淆矩阵的计算分析可知,相比Wishart分类方法,支持向量机分类法将分类精度从80.48%提高到88.00%,将Kappa系数从0.73提高到0.83。结果表明,单独的相干分解不能充分挖掘PALSAR-2数据包含的丰富信息,将目标极化分解参数用于特征信息分类处理,可以达到较好的分类效果;利用全极化PALSAR-2数据,结合目标极化分解方法和支持向量机分类法提取盐渍化信息有一定的优势。
[Abstract]:Taking Yutian Oasis in Xinjiang as the research area, the four-polarization PALSAR-2 data are used to process the multi-target polarization, and the corresponding polarization characteristic parameters are obtained. The SVM classification method is trained by visual interpretation with 11 polarization parameters with less noise as the best feature information. Several polarization decomposition methods combined with Wishart classification method and support vector machine classification method were used to extract the salinization information of different degrees in the study area. The classification results are quantitatively analyzed and verified by visual interpretation and field investigation combined with Landsat-8 land imager. Compared with Wishart classification method, support vector machine can improve the classification accuracy from 80.48% to 88.00 and increase the Kappa coefficient from 0.73 to 0.83. The results show that the rich information contained in the PALSAR-2 data can not be fully mined by single coherent decomposition, and the target polarization decomposition parameters can be used to classify the feature information, and the classification effect can be achieved by using the fully polarized PALSAR-2 data. It is advantageous to extract salinization information by combining target polarization decomposition method and support vector machine classification method.
【作者单位】：新疆大学资源与环境科学学院;新疆大学绿洲生态教育部重点实验室;
【基金】：国家自然科学基金重点基金联合项目(U1138303);国家自然科学基金(41561089,41361016)
【分类号】：TN957.52

【参考文献】