GF-1影像和OLI影像协同土地利用模糊分类方法研究

发布时间：2019-01-19 21:46

【摘要】：针对高分辨率遥感数据进行土地利用类型分类时出现的"同谱异物"现象,以及中分辨率遥感数据划分土地利用类型时受空间分辨率限制产生的"混合象元"问题,本文以高分一号数据(GF-1)和Landsat-8数据(OLI)为例,提出了一种协同利用高分辨率遥感数据和中分辨率遥感数据进行土地利用类型模糊分类的方法。首先,利用主成分变换的方法分别对GF-1纹理信息和OLI光谱信息进行压缩和增强,并将增强后的纹理信息和光谱信息进行特征协同;然后,根据各地物类型的光谱、纹理特征,对特征协同数据进行60、80、100共3个尺度的分割;最后,根据地物类型间的光谱特征和纹理特征的差异,构建各地物类型的模糊逻辑隶属度函数,实现对影像土地利用类型的模糊分类。实验结果表明,主成分变换的方法有效地将研究区GF-1和OLI数据的光谱、纹理信息压缩、增强,为面向对象分类中分类特征的选取提供了一种思路;同时,本文方法成功划分了研究区土地利用类型,并获得了较高分类精度,总体分类精度达到93.52%,对其它高空间分辨率与高光谱分辨率遥感数据协同分类研究具有一定借鉴意义。
[Abstract]:The phenomenon of "isospectral foreign bodies" in the classification of land use types with high resolution remote sensing data and the problem of "mixed pixels" caused by spatial resolution constraints in the classification of land use types with medium resolution remote sensing data are discussed. In this paper, a fuzzy classification method of land use type is proposed by using high-resolution remote sensing data and medium-resolution remote sensing data in cooperation with high-resolution remote sensing data (GF-1) and Landsat-8 data (OLI) as examples. Firstly, the GF-1 texture information and the OLI spectral information are compressed and enhanced by principal component transform (PCA), and the enhanced texture information and spectral information are combined with each other. Then, according to the spectral and texture features of the local object types, the feature cooperative data are segmented on three scales of 60,80100. Finally, based on the differences of spectral and texture features between the types of objects, the fuzzy logical membership function of the types of objects is constructed to realize the fuzzy classification of the types of land use in the image. The experimental results show that the principal component transform method can effectively compress and enhance the spectral and texture information of GF-1 and OLI data in the study area, which provides a way of thinking for the selection of classification features in object-oriented classification. At the same time, the method of this paper has successfully divided the land use types in the study area, and obtained the higher classification accuracy, the overall classification accuracy has reached 93.52. It can be used for reference to other cooperative classification of high spatial resolution and high spectral resolution remote sensing data.
【作者单位】：山东女子学院信息技术学院;湖北省国土测绘院;国土资源部地学空间信息技术重点实验室;国家测绘地理信息局四川基础地理信息中心;
【基金】：山东省高等学校科技计划项目(J12LN42) 全国统计科学研究计划(2012LY022) 山东省自然科学基金项目(ZR2012DL01) 山东高等学校科技计划项目(J15LN11)
【分类号】：P237

【相似文献】