无人机高空间分辨率影像分类研究

Classification of high resolution imagery by unmanned aerial vehicle

本文利用无人机影像进行土地利用类型研究,面向对象方法对影像分割,获取了最佳分割尺度;根据各土地类别的特征信息建立分类定义,提出了快速、准确获取土地利用类型的方法。研究结果表明,运用面向对象方法能很好地解决无人机高分辨率影像分类问题,其中关键是影像分割尺度的选择和影像对象特征信息的提取。

It is difficult to obtain remote sensing data and aerial photography images in multi-cloud region, so higher spatial resolution remote sensing data （unmanned aerial vehicle image） was selected for the research in land use classification; image segmentation scale was researched by applying object-oriented method, and the best segmentation scale was obtained, the classification categories definition was established based on the feature information, and a quick and exact method of obtaining land use types was put forward. Experimental results showed that the object-oriented technology could solve classification of unmanned aerial vehicle images very well, and the key is the choice of image segmentation scale and the extraction of feature information.