基于面向对象的QUICKBIRD数据和SAR数据融合的地物分类  

An Object-based Fusion of QUICKBIRD Data and RADARSAT SAR Data for Classification Analysis

作　　者：王宇航[1] 范文义[1] 刘超逸[1] 

机构地区：[1]东北林业大学,哈尔滨150040

出　　处：《东北林业大学学报》2016年第9期44-49,共6页Journal of Northeast Forestry University

基　　金：国家高技术研究发展计划(2012AA102001)

摘　　要：为了实现精确植被类型信息提取,以福建省三明市将乐林场Quickbird影像和Radarsat-2全极化影像作为基础数据,探讨高空间分辨率光学遥感影像与SAR(合成孔径雷达)全极化影像融合进行地表覆盖及森林类型识别的可行性。采用面向对象多尺度分割方法对Quickbird全色与多光谱的融合影像进行处理,SAR影像采用Gram-Schmidt融合方法处理,运用处理的Quickbird与SAR的融合影像,分类提取植被的光谱、纹理和几何特征信息,建立类层次结构,并对分类结果进行比较分析。结果表明:基于对象与知识的方法对高空间分辨率影像分类取得了较好的分类效果,多源遥感数据分类的总体精度为0.903。Based on the Quickbird data and Radarsat-2 full polarization data of San Ming City, Fujian Province, the object-ori- ented method was adopted to identify the land cover types from the fusion of Quickbird panchromatic and multi-spectral im- age, SAR images and the fusion of Quickbird image and SAR image which acquired by using the Gram-Schmidt method. Classification factors including spectral, texture and geometric features were used to establish a class hierarchy, and the classification results were compared. The knowledge-based and object-based methods was effective in the identification and classification of a high spatial resolution images, and vegetation types were effectively identified. The accuracy of muhi- source remote sensing data was up to 0.903 with some improvements.

关 键 词：多源遥感 面向对象 尺度分割 合成孔径雷达(SAR) 数据融合 

分 类 号：S757.2[农业科学—森林经理学]