基于面向对象的橡胶分布面积估算研究

Object oriented estimation of rubber plant area using remote sensing data

海南是最大的天然橡胶种植区域,常规的调查方法费时费力。随着高分辨率遥感影像的应用,为橡胶信息提取提供了可能。以高分辨率QuickBird卫星影像为基础,采用面向对象的信息提取方法,用光谱、形状、纹理等构建特征空间,进行橡胶的分类试验,获取橡胶的种植面积信息,结果显示,基于面向对象方法的提取精度高,用面向对象的分类方法进行橡胶信息的提取是可行的。

Hainan is the largest natural rubber plantation area, the normal investigation was laborious and time-consumed method. With the developing of high resolution remote sensing images, it was possible for us to extract rubber plant area.Based on object-oriented methods and using spectral, shape, texture to build space characteristics for the classification of rubber plant, which was used to access rubber planting area from the high-resolution QuickBird satellite image. The results indicated that, Object-Oriented classification method was high precision, which was feasible to extraction rubber information.