A Robust Zero-Watermarking Algorithm for Spatiotemporal Trajectory Data Protection

Abstract

As an important type of spatiotemporal big data, trajectory data faces severe challenges such as illegal copying and dissemination, which infringes upon the legitimate rights of copyright owners. Existing copyright protection methods for trajectory data often regard it as a collection of points for watermark embedding, ignoring its inherent spatiotemporal characteristics. These methods exhibit limitations in both robustness and specificity. To address this, this paper introduces a zero-watermarking algorithm for trajectory data that effectively integrates spatiotemporal features. The proposed algorithm begins by segmenting the trajectory into sub-trajectories using stay areas as key nodes, extracting feature points accordingly. Then, a watermark index is constructed based on the movement speed of each sub-trajectory in the X-direction, with the watermark information generated through a voting mechanism. Finally, an exclusive-ORing (XOR) operation is performed between the watermark information and the scrambled copyright image to produce the zero-watermark image. Experimental results demonstrate that the proposed method exhibits strong robustness, effectively resisting various attacks such as time shifting, cropping, and coordinate point deletion. Specifically, the normalized correlation (NC) value remains above 0.95 even when 50% of the coordinate points are removed, and achieves an NC value of 1.00 under geometric attacks including translation and scaling. Compared to existing watermarking schemes for trajectory data, the proposed approach exhibits significantly enhanced robustness.