TM图像的矿区土地资源分类体系的建立与信息提取

以多时相TM图像为数据源，分析矿区土地资源的主要演变形式，参考国家标准分类体系，提出了适合于沈阳矿区的土地资源监测分类体系.结合前人的研究成果，分析地物在TM图像上的灰度变化、空间关系及结构变化，对矿区水体、建设用地、绿地和耕地、塌陷地分别进行了模型提取研究.并采用GIS支持下的分层变化信息提取方法，通过建立知识及决策树模型对信息进行变化检测，检测结果完全能够满足对矿区土地资源演变分析的精度要求，且均采用半自动的提取模型与知识决策树提取方法，简单易行，方便GIS数据管理与分析.

Building of land resources classification system and information extraction based on TM image in mining area

Using multi temporal TM images as data source to analyze the main evolution type of land resources,built land resources classification system for Shenyang mining area.On this basis,through the spectral analysis,spatial relation analysis,structure analysis of TM images,researched the extraction model of mining land information include water,building land,green land,plantation,finally discussed the abstraction method of subsidence ground.In the actual abstraction work,adopted the knowledge method to extract subsidence ground under the support of radar remote sensing images.With the help of GIS spatial analysis,extracted the land information from every temporal TM image and changed information from adjoining,building knowledge and decision tree model for changed detection of land,precision of detection result is satisfied.