论文部分内容阅读
１９９７ 年６ 月利用ＰＨＩ推帚式成像光谱仪在北京市沙河镇进行了飞行实验，其主要目的是成像光谱技术应用于城市用地和建筑物分类中关键技术研究。对所获取的１５ 个波段的可见光－ 近红外数据进行了分析与处理，试图通过对地面覆盖物质的光谱响应特征的分析来进行城市人工目标的识别与分类。由于城市地物的光谱特征异常复杂，难于应用一般的模式分类算法分离出所有的类型，本文在对地面覆盖类型光谱响应分析的基础上，采用分层复合分类的方法，逐步将不同的地物类型分离开来，实验证明是非常有效的
In June, 1997, a PHI push broom imaging spectrometer was used to conduct flight experiments in Shahe Town, Beijing. The main purpose of this experiment was to apply imaging spectroscopy to the study of key technologies in urban land use and building classification. The acquired 15 bands of visible-near-infrared data were analyzed and processed in an effort to identify and classify urban artificial targets by analyzing the spectral response characteristics of the ground cover. Due to the complexity of the spectral features of urban features, it is difficult to separate all the patterns using the general pattern classification algorithm. Based on the spectral response analysis of the ground cover types, this paper adopts the method of hierarchical compound classification to gradually separate different features Type separation, the experiment proved to be very effective