A key points-based blind watermarking approach for vector geo-spatial data

This paper presents a blind watermarking approach to protecting vector geo-spatial data from illegal use. By taking into account usability, invisibility, robustness, and blindness, the approach firstly determines three feature layers of the geo-spatial data and selects the key points from each layer as watermark embedding positions. Then it shuffles the watermark and embeds it in the least significant bits (LSBs) of the coordinates of the key points. A similar process for selecting the feature layers and the key points in the watermark embedding process is carried out to detect the watermark followed by obtaining the embedded watermark from the LSBs of the coordinates of the key points. Finally, the similarity degrees of three versions of the watermark from three feature layers are calculated to check if the data contains the watermark. Our experiments show that the method is rarely affected by data format change, random noise, similarity transformation of the data, and data editing.