矢量地理数据密级标识方法研究

密级标识是用于标识信息产品安全等级的隐蔽信息,是信息分级管理、控制与应用的基础和前提。作为涉密地理信息产品主要组织格式的矢量地理数据,其密级标识方法的研究具有重要意义。结合矢量地理数据在遭受几何攻击时,其要素空间关系不变性特征,以及密级标识方法的不可感知性、高鲁棒性要求,本文提出一种基于距离比值调制的矢量地理数据密级标识方法,以实现矢量地理数据的分级管理与保护。其主要研究思路为：① 基于距离比值调制的方法,能够较好地应对几何攻击;② 在每个线面要素内,通过重复嵌入密级标识,有效提高算法对于编辑操作的鲁棒性;③ 运用道格拉斯-普克算法提取特征点,在此基础上进行密级标识的嵌入,能够较好地对抗压缩攻击。实验结果表明：该方法具有良好的不可感知性;同时,对嵌入密级标识后的矢量地理数据进行几何变换、数据编辑、数据压缩等攻击后,其密级标识的提取率能够达到80%以上,表明方法具有较好的鲁棒性。

Study on Secret Classification Marking Method for Vector Geographic Data

Secret classification marking is hidden information used to mark the security level of information products and is the basis and precondition for classification management, control and application of information. Since vector geographic data is the main organization format of classified geographic information products, it is of great significance to study secret classification marking method for vector geographic data. When suffered from geometry attack, the spatial relationship among elements of vector geographic data is invariable and secret classification marking method requires good imperceptibility and high robustness. In order to realize classification management and protection of vector geographic data, this paper presents a secret classification marking method for vector geographic data based on distance ratio modulating. In essence, using the digital watermark technology, the method embeds secret classification marking as watermarking information into vector geographic data, which has important application value for security protection of vector geographic data. The main ideas related to the study are as follows: Firstly, the method is based on distance ratio modulating and it can deal with geometry attack very well. Secondly, the method effectively improves the robustness against editing attack by repeatedly embedding secret classification marking within each line and area. Lastly, through embedding secret classification marking into feature points which are extracted by using Douglas-Pecker algorithm, the method can also effectively resist data compression attack. The experimental results show that the proposed method has good imperceptibility. In addition, when vector geographic data is suffered from the attacks, such as geometric transformation, data editing and data compression, the extraction ratio of its secret classification marking can reach almost more than 80%, indicating the robustness of this method.