基于资源三号卫星遥感影像的山区套种烟草面积估测

Estimation of the Intercropped Tobacco Area in Mountainous Area Based on ZY-3 Remote Sensing Images

[目的]针对传统实地测量难以实现山区大面积套种烟草面积监控的问题,研究一种快捷、高精度的估测途径。[方法]利用ENVI5.1软件处理资源三号卫星2.1 m全色影像和5.8 m多光谱影像,导入1∶1万DEM数据进行正射校正以纠正地形起伏引起影像上的误差。野外获得具有代表性的地面控制点(GCP),利用面向对象分类方法,对影像进行智能化的分割和合并,以GCP的光谱、纹理和形状属性创建规则,进而提取烟草面积。[结果]经过分类后处理提取的研究区烟草面积为14 088.46 hm2,分类精度达到94.63%。[结论]面向对象分类方法提取遥感影像信息的实用性较高。

[ Objective ] To find a rapid and high-precision estimation approach for estimation of the intercropped tobacco area in Mountainous Area. [ Method ] ZY-3 2.1 m panchromatic image and 5, 8 m multispectral image were researched by ENVIS. 1 software. Orthorectification was carried out by introducing 1：10 000DEM data, so as to correct the image error induced by topographic relief. Representative ground con- trol points （GCPs） were obtained in outdoor. [ Result＇] Images were intelligently combined and divided by object-oriented classification meth- od. Regulations were established based on the GCP spectrum, texture and shape properties. The tobacco area was extracted, which was 14 088.46 hm2. The classification accuracy reached 94.63%. [ Conclusion] Extracting remote sensing images by object-oriented classifica- tion method had relatively high practicability.