基于Wordview-3数据的面向对象地物分类研究

Research on Object-Oriented Feature Classification Based on Word View-3 Data

利用Wordview-3数据对土地利用类型分类研究具有重要的意义。以云南省勐海县作为研究区域,采用Wordview-3遥感影像为数据源,对影像进行多尺度分割,建立相应地物的分类规则,采用支持向量机(Support Vector Machine, SVM)进行分类。实验结果发现:利用面向对象方法对地物分类的总体精度达到87.46%,Kappa系数为0.85,总体精度比基于像元NDVI的分类提高了7.05%,Kappa系数提高了0.11。

It is of great significance to use Word view-3 data to classify land use types. Taking Menghai County, Yunnan Province as the study area, Word view-3 remote sensing images were used as the data source to segment the images at multiple scales, establish classification rules for the corresponding features, and use Support Vector Machine (SVM) for classification. Experimental results show that the overall accuracy of the object-oriented approach to feature classification is 87.46%, with a Kappa coefficient of 0.85. The overall accuracy is 7.05% higher and the Kappa coefficient is 0.11 higher than that of NDVI-based classification.