基于虚拟现实的分类模板定义技术在水土保持遥感调查中的应用  

Application research of technique with defining classification signatures based on virtual reality on remote sensing investigation of soil and water conversation

作　　者：吴卿[1] 王冬梅[1] 周心澄[1] 秦志强 丁素荣[3] 

机构地区：[1]北京林业大学水土保持学院水土保持与荒漠化防治教育部重点实验室 [2]河北省馆陶县水利局,河北馆陶057750 [3]内蒙古赤峰市农业科学研究所,内蒙古赤峰024031

出　　处：《中国水土保持科学》2006年第6期56-60,共5页Science of Soil and Water Conservation

基　　金：黄河水利委员会创新项目"伊洛河流域郑州市(巩义)项目区水土保持生态环境建设3S动态监测方法研究"

摘　　要：为提高水土保持遥感调查中的分类精度,基于虚拟地理信息系统(Virtual GIS)软件平台,构建伊洛河流域的虚拟现实环境,以SPOT5卫星2.5 m空间分辨率的影像为数据源,在卫星影像的三维可视化观察的基础上,选择训练样地并定义分类模板。通过与传统二维环境下的分类模板定义技术进行比较,发现前者分类模板定义整体精度可以达到95.48%,而后者最高达88.19%;认为虚拟现实技术可增强解译人员对遥感影像的认知能力,有助于提高分类精度,表明其在遥感影像分类方面具有较大的应用潜力。For improving the classification accuracy on remote sensing investigation of soil and water conversation, the Virtual Reality Environment for watershed of Yiluohe River was generated based on Virtual GIS software platform. Using the satellite images of SPOT5 2.5 m, the training tion signatures was defined under the 3D visible viewer. Comparing with sites were selected and the classificathe traditional method under 2D viewer, the results shown the total precision of the former can reach at 95.48%, and the latter only 88.19%. Our research has illuminated that the technique of Virtual Reality can enhanced the capacity of perceivability on satellite images and increase the classification accuracy. Finally, it presented that the Virtual Reality technique have the huge potentiality on satellite images classification.

关 键 词：虚拟现实 遥感 水土保持 分类模板 

分 类 号：S157.1[农业科学—土壤学]