辅以纹理特征的典型盐渍区地表覆被遥感影像分类  

Land Cover Classification in Salt-affected Zone Supplemented with Textural Features

作　　者：樊彦国[1] 侯春玲[1] 郭若成[1] 钱尊岩[1] 张涛[1] 

机构地区：[1]中国石油大学(华东)地球资源与信息学院,青岛266555

出　　处：《遥感信息》2011年第4期34-37,共4页Remote Sensing Information

基　　金：中央高校基本科研业务费专项资金资助;山东省科技攻关项目(2008GG10009018)

摘　　要：结合实测的盐渍地表光谱数据,采用经验线性法对遥感影像进行大气校正。分析了实测高光谱数据与遥感影像光谱值的关系,然后针对盐生植被光谱的相似性,对影像进行纹理特征提取。结合影像的纹理特征和光谱特征,运用神经网络分类法进行盐渍区地表覆被分类。结果表明辅以纹理特征的分类方法有效提高了盐生植被的分类精度。The paper firstly analyzed the relation between the measured spectral data and multi-spectral image data after atmospheric correction by empirical line calibration method.Then textural features were extracted using the multi-spectral ALOS image.Classification and recognition were carried out by the neural network method combined with the spectral characteristics and spatial characteristics on the synthetic image.The classification accuracy was improved.

关 键 词：经验线性法 纹理特征提取 盐生植被分类 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]