基于高分二号卫星数据的农作物分类方法研究  

Research on Crop Classification Based on GF-2 Satellite

作　　者：曹伟男 王文高 王欣 于亚娇 刘善军[1] CAO Weinan;WANG Wengao;WANG Xin;YU Yajiao;LIU Shanjun(School of Resources and Civil Engineering,Northeastern University,Shenyang 110819,China)

机构地区：[1]东北大学资源与土木工程学院,辽宁沈阳110819

出　　处：《测绘与空间地理信息》2021年第4期158-161,共4页Geomatics & Spatial Information Technology

摘　　要：针对目前高分二号卫星数据(GF-2)有较高的空间分辨率而在农业领域应用较少和农作物分类普遍存在"同谱异物"和"同物异谱"的现象,以辽宁省沈阳市苏家屯区以西的新开河村周边为试验基地,利用最佳波段组合指数法(OIF)对所选取的高分二号(GF-2)卫星数据的纹理特征和植被指数以及波段信息进行筛选,选取最佳的波段组合,以增加分类信息、减少数据冗余。最后,针对筛选后的数据,使用最大似然法进行分类,得到农作物的分类结果。结果表明,利用该方法对农作物进行分类,分类精度得到了一定程度的提高,为目前大规模农作物种植面积的精确、迅速统计提供了一套可行的方案。At present,GF-2 satellite has a high spatial resolution,but it is seldom applied in the agricultural field,and the phenomenon of"same spectral from different materials"and"same material with different spectral"is universal in crop classification.For these problems above,we take the surrounding areaof Xinkaihe village(SujiatunDistrict,Shenyang City,Liaoning Province,China)as the test base,texture features,vegetation index and band information of the GF-2 satellite are screened taking advantage of OIF,In order to increase the classified information and decrease the numeral redundancy in the data.Finally,based on the screened data,classifying crops uses the maximum likelihood method.The classification result indicates that classify crops in this way can improve classification accuracy to a certain degree,it provides a feasible program forcounting the cultivated area of the vast crop accurately and quickly.

关 键 词：高分二号 灰度共生矩阵 纹理特征 OIF 最大似然法 

分 类 号：P237[天文地球—摄影测量与遥感]