基于无人机遥感影像分类方法研究

Research on Classification Method of Remote Sensing Image Based on UAV

以无人机遥感影像作为分类数据,分别采取ENVI软件中的支持向量机分类方法和eCognition软件中的面向对象的分类方法,对所选研究区的农作物以及研究区所存在的其他地物类型进行分类。试验结果表明:eCognition软件中面向对象的分类方法的分类效果较好,总体分类精度达到了90.99%,Kappa系数为0.8737,相对ENVI软件中的支持向量机分类方法,eCognition软件中面向对象的分类方法更加适合用在无人机遥感分类上。

Taking UAV remote sensing image as classification data,the classification methods of support vector machine in ENVI software and object-oriented in eCognition software were adopted respectively to classify the crops in the selected research area and other types of land objects in the research area.The experimental results showed that the object-oriented classification method in eCognition software had a good classification effect,the overall classification accuracy reached 90.99%,and the Kappa coefficient was 0.8737.Compared with the support vector machine classification method in ENVI software,the object-oriented classification method in eCognition software was more suitable for UAV remote sensing classification.