面向对象分类法在唐山南湖遥感信息提取中的优越性分析

摘要

The object-oriented classification method and pixel-oriented classification in remote sensing information extraction are used respectively by taking Tangshan Nanhu as study object and different classification results are achieved. The experiment shows that, in the process of Nanhu remote sensing image information extraction, object-oriented classification overall accuracy (91.3%) is much larger than the overall accuracy of pixel-oriented classification (72.1%). Therefore, an object-oriented classification method has obvious advantages. It may provide some reference for other areas of remote sensing information extraction.%选取唐山南湖作为研究对象，分别采用面向对象分类法与面向像元分类法对研究区进行遥感信息提取，得到不同的分类结果。实验证明，在南湖遥感影像的信息提取中，采用面向对象分类法总体精度（91.3%）要远大于面向像元法的总体精度（72.1%），因此，面向对象分类法具有明显优势，对于将面向对象分类方法应用于其它领域的遥感信息提取能够提供一定的借鉴意义。