基于脆弱水印的轨迹数据篡改检测方法

Tampering detection of trajectory data based on fragile watermarking

轨迹数据是导航与位置服务系统的核心,是一种重要的数据资源。为验证轨迹数据的完整性,检验轨迹数据是否被篡改,提出一种基于脆弱水印的轨迹数据篡改检测方法。该方法包括水印嵌入算法和水印检测算法,其中水印嵌入算法和水印检测算法分别由组水印和轨迹点水印算法组成。嵌入水印时,首先将轨迹数据按照分组时间分为若干个组,然后将由组标识生成的组水印和由轨迹点坐标生成的轨迹点水印分别嵌入到轨迹点坐标的最低有效位与次最低有效位。水印检测时,首先分别对轨迹数据进行组水印和轨迹点水印检测,然后通过水印验证向量的结果来判定轨迹数据的完整性,最后统计篡改检测率并评估算法的脆弱性。结果表明,当轨迹数据以不同的分组和篡改幅度篡改时,其修改检测率均大于99%,而且该算法可以定位数据篡改并识别篡改类型,最后通过实验验证了这一点。该方法可为同类轨迹数据篡改检测算法的设计提供参考。

Trajectory data is the important infras tructure of Navigation and location service system, and is an important data resource. In order to verify the integrity of trajectory data and determine whether the trajectory data is tampered or not, a fragile watermarking scheme is proposed.The fragile watermarking scheme includes watermark embedding algorithm and watermark detection algorithm. The watermark embedding algorithm and the watermark detection algorithm are composed of the group algorithm watermark and the track point watermark algorithm respectively. When embedding watermark,first of all the trajectory data is divided into several groups by the group time, then trajectory point coordinates of least significant bit and next least significant bit are embedded by group watermark generated by group identification and the trajectory point watermark generated by a group of trajectory point coordinates. Watermark detection algorithm consists of group watermark detection algorithm and trajectory points watermark detection algorithm. When detecting the watermark, first of all the integrity of trajectory data can be verified by the watermark verification vector＇s result pattern, and then according to the result of watermark verification vector the detection rate can be revised and the vulnerability of the watermarking algorithm can be evaluated. The results demonstrate that when the trajectory data is tampered with by different group and tamper amplitude, all the modified detection rates are greater than 99%. What＇s more, the modifications made to trajectory points can be detected and localized,which is also checked by the experimental results. This method can provide a reference for the design of similar methods for trajectory data tamper detection.