多源同名居民地面目标识别及其同名特征点匹配方法

Automatic Recognition of Multi-source Homonymous Residential Area Targets and Their Homonymous Feature Point Matching Method

同名目标匹配是空间数据融合、共享与集成的关键所在,针对多源居民地面目标空间数据融合问题,该文提出一种同名居民地面目标自动识别及其同名特征点自动匹配算法。该算法通过计算面目标质心重叠前后的匹配相似度实现同名居民地面目标多重匹配关系的自动识别;通过构建向量方向相似度、面积比相似度与距离邻近度等特征指标,并将其建模为最优化函数,进而采用基于编辑距离的串匹配算法,有效解决了不同匹配关系下同名居民地面目标特征点的匹配问题,进一步实现了居民地面目标的几何纠正。以实测的不同来源导航电子地图郑州市部分居民地面目标数据对算法进行验证,结果表明:该算法能稳健识别多源居民地面目标间各种匹配类型,且能自动匹配同名特征点,二者准确率均在95%以上,可为面目标位置的精确融合奠定重要基础。

Homonymous target matching is the key to spatial data fusion,sharing and integration.Aiming at the problem of spatial data fusion of multi-source residential area targets,an automatic recognition algorithm of residential area targets and the automatic matching algorithm of their homonymous feature points are proposed.The algorithm calculates the matching similarity before and after the overlap of the area target centroid to achieve the automatic identification of multiple matching correspondences of the homonymous residential area targets.By constructing feature indicators such as vector direction similarity,area ratio similarity,and distance proximity,modeling them as optimization functions,and then using a string matching algorithm based on editing distance,the matching problem of feature points on the homonymous residential area targets under different matching relationships is effectively solved,and the geometric correction of the homonymous residential area targets is further realized.Taking the residential area targets of Zhengzhou from different sources of navigation electronic maps as the experimental dataset,the case study is carried out.The study results indicate that the proposed algorithm can robustly recognize various matching relationships among multi-source residential area targets with an accuracy of more than 95%.For various matching types of multi-source residential area targets,the proposed method can automatically match the homonymous feature points,and the accuracy is over 95%.The study can lay an important foundation for accurate fusion of area target position.