运用多算子加权的面要素几何匹配方法

Geometric Matching Method of Area Feature Based on Multi-weighted Operators

针对面要素几何匹配算法,提出反映距离、尺寸、方向、形状4种主要几何度量指标的多算子加权匹配方法,建立了质心距离、Hausdorff距离、面积重叠度、对称差、轮廓紧致度和正切空间的量化算子,对量化结果进行加权评分计算总相似度确定面要素匹配值。运用多版本实测数字线划图数据进行匹配实验,结果表明该方法匹配结果可靠,并且通过算子权重和匹配阈值的调节设置可以筛选受关注度量指标的高匹配结果。

By analyzing existing disadvantages for spatial object matching methods, multi- weighted operator matching method for areal feature geometric matching algorithm is pro- posed in this paper. This method can reflect four main geometric measuring targets： dis- tance, size, orientation and shape. This proposed algorithm builds six quantitative opera- tors. centroid distance, Hausdorff distance, area overlap, symmetric difference, outline compactness and tangential space. Then we calculate the total similarity for quantitative re- suits by weighted scores. The experimentation used multi-version digital line graphs for matching, and test results show that this method is credible. Furthermore, it can also filter better matching results to more concerned targets by adjusting weight and threshold of the operators.