涉密矢量数字地图中敏感要素几何信息量的测度方法

Quantitative Measurement of Geometrical Information for Sensitive Features in Secret-related Vector Digital Maps

涉密矢量数字地图中敏感要素几何信息的定量测度,是实现定量化密级评定的基础和前提。如何进行敏感要素几何信息量的计算,目前尚缺乏相关研究。在现有地图信息论研究成果的基础上,论文主要开展了3方面的研究工作:首先,从地理对象的敏感性入手,探讨了地图要素敏感性的定义与要素集合的敏感性表征指标;然后,基于任意发生元的Voronoi构建方法,实现了对点要素集、线要素集、部分覆盖面要素集及综合要素集的信息单元自动化剖分;最后,基于长度系数、面积系数和角度系数,提出了线、面要素图形复杂度的计算方法,并在此基础上形成了涉密矢量数字地图中敏感几何信息量的测度方法。实验表明,该方法的计算结果遵循了信息量的非负性、连续性、可加性等特点,较为合理地反映了相关因素(地图比例尺、要素空间分布、要素图形复杂度、要素敏感系数等)对计算结果的影响。相关成果将支持矢量数字地图的定量化密级评定,并有助于完善地理信息安全监管的理论和方法体系。

Quantitative measurement for sensitive geometrical information of secret-related vector digital maps is the basis and precondition of achieving the quantitative evaluation of classification level. At present, there are rare related researches on how to measure the sensitive geometrical information of a map. Based on the findings of map information measurement, we focus on the measurement method of the sensitive geometrical information for secret-related vector digital maps. The methodology is composed of three phases. Firstly, we discuss the definition of map feature sensitivity and the sensitivity index of feature sets. Then, we aim to the automatic division of information-unit for point feature set, line feature set, polygon feature set and comprehensive feature set. The method of constructing voronoi diagram with general generators is used. Finally, based on length coefficient, area coefficient and angle coefficient, we propose the methods for figure complexity of line feature and polygon feature. The results demonstrate that it follows the principles of information theory, such as nonnegative, continuous and additive, and reasonably reflects some factors＇ （map scale, features spatial distribu- tion, features figure complexity and features sensitivity coefficient） impact on the results. The results will support the quantitative evaluation for classification level of vector digital maps, and help to improve the theory and method systems of geographical information security supervision.