论文部分内容阅读
使用Radarsat-2全极化雷达数据,基于面向对象技术对图像进行分割,分析各类地物的散射特征,并对分割对象进行特征提取,依此设计各类地物的提取因子并对鄱阳湖湿地进行地物信息提取,最后利用实地采样数据和同期的高分一号影像数据对分类结果进行验证。研究表明:采用面向对象技术并结合地物散射特征设计的提取因子,可以有效克服雷达图像分类中常见的斑点现象,提高湿地分类精度。
Using Radarsat-2 Polarimetric Radar data, the image is segmented based on object-oriented technique, the scattering characteristics of various types of ground objects are analyzed, and the features of the segmented objects are extracted. Based on this, the extraction factors of various ground objects are designed and the Poyang Lake Wetland for feature extraction, and finally the use of field sampling data and high scores on the same period the image data to verify the classification results. The research shows that adopting the object-oriented technology and combining with the extraction factor of the feature scattering feature can effectively overcome the speckle phenomenon common in radar image classification and improve the classification accuracy of wetlands.