GIS 支持下三峡库区主要地类自动分类方法研究

针对三峡库区遥感图像变异大的特征, 设计建立了主要地类自动分类系统并在其子系统中完成了地形光谱纠正、辅助数据生成及数据合成等预处理工作; 提出了一种可交互作业的计算机全自动土地利用类型分类方法。应用专业的图像处理软件eCognition 为平台, 通过图像图斑化、选取样本、寻找最佳波段并生成判别函数来进行自动分类, 在很大程度上克服了“同物异谱”、“异物同谱”、“边界过渡”等现象对分类结果的影响。在应用地理信息系统和自动分类方法来获取波谱特征复杂地区的土地利用信息的技术路线和方法方面作了有效探索。

Expoloration on Method of Auto-classification Based on Gis for the Main Ground Objects of the Three Gorges Reservoir Area

Aiming at resolving the problem of great variances of remote sensing image in the Three Gorges Reservoir Area,this paper not only builds up the automatically classifying system of land cover and finishes such tasks as correcting the topographical spectrum,producing auxiliary data,forming the needed image in its subsystem but puts forward a kind of manual controlled computer-automatic categorized method.Taking the specialized image analysis software-eCognition as the platform,the research not only achieves the goal of classifying through choosing samples,picking out the best Wave bands,producing the identifying functions,but also dispels the influence of such phenomena as the same thing with different spectrums,different things with the same spectrum,and border transitions,etc. The research does certain exploration in the aspect of technological route and method of the use of GIS and automatic classification of the remote sensing image to obtain the information of land cover for the regions whose ground objects have complicated spectrum as well.