论文部分内容阅读
以Landsat ETM+为数据来源,基于ENVI平台与ID语言构建的条件分类模块为基础,面对林业应用中时常出现的地类复杂、分布交错、高程变化大的丘陵山地区域,在多尺度的光谱、纹理特征统计的基础上,通过对影像的充分认知,结合地形特征、知识经验,引入多种归一化指数,根据包容性、区分性原则建立阈值规则,得到一次分类结果后,应用优先级理论修正未划分区域,并根据面向地类间的空间关系推理,结合经验常识构建两种搜索模型用于叠合图斑修正.经过多次调整,参数修正,易混区划的生产精度、用户精度分别为85.57%,84.85%.
Taking Landsat ETM + as the data source and based on the condition classification module constructed by ENVI platform and ID language, in the face of the frequently occurring mountainous terrain area with complicated terrain, staggered distribution and high elevation in forestry application, Based on the statistics of the texture features, through the full cognition of the images, combined with the topographic features and knowledge experience, a variety of normalized indices are introduced, threshold rules are established according to the principle of inclusiveness and differentiation, and the result of a classification is obtained. The application priority According to the theory of space-based spatial reasoning and the common sense, two kinds of search models are constructed for the correction of overlapped patches.After many adjustments, the parameters are corrected, the production accuracy of the easy-to-digest partition, the user precision Respectively 85.57%, 84.85%.