面实体多重表达下的谱特征量化

Quantification of spectral features under multiple representations of face entities Quantification of spectral features under multiple representations of face entities

在对地理信息的数字化表达存储中,面实体在点、线、面、体不同维度的实体中占据最多的份额,而面实体在多重表达下的几何、属性、知识方面的特征最为复杂,针对面实体多重表达下的复杂性及其形态差异,提出谱特征量化描述空间实体本身的形态。优化矢量面状轮廓的形状分割比,然后建立形状内距分割比矩阵和形状外距分割比矩阵,从而组建谱特征集,并应用其计算面实体多重表达下的形态相似度模型和形态复杂度模型,计算结果表明了计算模型的可行性,最终为无级化地物表达提供描述地理实体的定量指标。

In the digital representation and storage of geographic information,surface entities occupy the largest share in entities with different dimensions of point,line,surface and volume,while the geometric,attribute and knowledge features of surface entities under multiple representations are the most complex.Aiming at the complexity and morphological differences of surface entities under multiple representations,spectral features were proposed to quantitatively describe the shape of space entities,optimize the shape segmentation ratio of vector areal contour,and then establish the shape internal distance segmentation ratio matrix and the shape external distance segmentation ratio matrix,so as to build the spectral feature set,and use it to calculate the shape similarity model and shape complexity model under the multiple representation of surface entities.The calculation results showed the feasibility of the calculation model,and finally provided quantitative indicators to describe the geographical entities for the hierarchical representation of ground objects.