“高分一号”卫星数据遥感影像分类方法研究——以内蒙古维拉斯托地区为例

Classification method of GF-1 Satellite data:A case study on the Vilasto district,Inner Mongolia

"高分一号"(GF-1)卫星是中国高分辨率对地观测系统的首发星,突破了高空间分辨率、多光谱与宽覆盖相结合的光学遥感等关键技术,开启了中国对地观测的新时代。文章以内蒙古维拉斯托地区为研究对象,首先对GF-1卫星采集的遥感数据进行正射校正、配准、融合、裁剪等预处理;采取不同的图像分类方法提取研究区地形地貌特征及矿业活动特征,并对不同分类方法提取结果进行合并处理与精度评价。结果显示,利用监督分类中人工神经网络方法可以对研究区中的矿业活动特征进行有效的提取。同时应用此种分类方法对大兴安岭西区景观特征的提取也有指导意义。

GF-1 satellite is China＇s first high-resolution earth observation system satellite. It had made a breakthrough on a number of key technologies, such as high spatial resolution, multi-spectral and wide coverage combined optical remote sensing, and started a new era of China＇s earth observation. The authors carried out the case study in the Vilasto district using remote sensing data collected by GF- 1 satallite by processing of thorectification, registration, fusion and clipping. The topographic features of the study area and mining activities by using different image classification methods were extracted, then the different classification resuhs in Vilasto district, Inner Mongolia, were jointed and evalauted. The results show that the artificial neural network method can effectively extract the mining activity characteristics in the study area. At the same time, the application of this classification method can guide the extraction of landscape features in the western area of the Da Hinggan Mountains.