多源地理空间数据集成与更新是地理信息科学的前沿课题，而空间目标匹配则是这一课题研究中的关键难题。匹配方法涉及的相似性度量指标选取与计算、阈值选取的不确定性和非一对一的复杂匹配关系，成为制约同名目标匹配技术的应用瓶颈。本项目以GIS空间线目标匹配为研究对象，研究最优化与统计回归理论在地理空间同名目标匹配中的匹配模型构建与应用。具体研究内容包括：在多源多尺度空间数据下，基于中位数的统计思想，研究现有空间数据同名线目标几何距离、方向、形状等相似度指标的稳健计算模型；基于最优化和统计回归理论，提出全局最优化与多变量逻辑回归匹配匹配模型，旨在提高同名目标匹配的正确率与查全率。研究成果为解决当今GIS应用中不同来源、不同精度的数据合并、集成、更新、变化检测等空间分析和操作中的关键问题（同名目标匹配）提供新的思路和解决方法。

Multi-source geospatial data integration and updating is the pivotal frontier in the domain of geographical information science, and the key problem is the object matching,which is to find all sets of corresponding objects, i.e., objects that represent the same real-world entity in distinct sources. The issues such as the similarity measures selection and calculation, the uncertainty in similarity threshold and the none one-to-one matching relationships become a bottleneck restricting the applications of the object matching. This project is to present a comprehensive methodology for linear object matching based on the theories of optimization and logistic regression. The detailed research contents are as follows: based on the median statistics properties, this project studies the robust calculation model of the existing similarity measures of corresponding objects (e.g., geometric distance, direction, shape); based on the theories of optimization and statictical regression, this project proposes the multivariate logistic regression matching model, aiming to improve the matching accuracy and matching recall. The researches are expected to provide new ideas and solutions for solving the key problem (i.e., object matching) in current GIS applications in the geospatial data integration, updating and information sharing.