一种格网划分的三维点云数据数字水印算法

A digital watermarking algorithm for 3D point cloud data based on grid division

针对现有的三维点云数据数字水印算法对随机增点、简化和裁剪等攻击鲁棒性不足的问题,该文提出一种格网划分的三维点云数据数字水印算法。首先,将三维点云数据分别沿X、Y和Z轴进行投影,并在对应的投影平面上进行均匀格网划分;其次,对分块后每一个格网中的顶点坐标分别进行归一化处理;然后,运用映射方法建立归一化坐标与水印信息之间的对应关系;最后,运用QIM的方法,分别通过修改归一化后的X、Y和Z坐标以嵌入水印信息。利用分块重复嵌入和映射方法,算法的鲁棒性得到了增强,并能实现水印的盲检测。实验表明,该算法对常见的平移、缩放、随机增点、简化和裁剪等攻击具有良好的鲁棒性,并且具有良好的不可见性,为三维点云数据的版权保护提供了一种新的解决方案。

Aiming at the lack of robustness against attacks such as random point increment,simplification and cropping in existing 3D point cloud data digital watermarking algorithms,this paper proposed a 3D point cloud data digital watermarking algorithm based on grid partitioning.First,project the 3D point cloud data along the X,Y,and Z axes respectively,and perform uniform grid division on the corresponding projection plane Secondly,the coordinates of the vertices in each grid are normalized separately after the chunking process.Then,the mapping method is used to establish the correspondence between the normalized coordinates and the watermark information.Finally,the QIM method is used to embed the watermark information by modifying the normalized X,Y and Z coordinates respectively.Using chunked embedding and mapping methods,the robustness of the algorithm was enhanced and blind detection of the watermark was achieved.Experiments show that the algorithm is robust to common translation,scaling,random increment,simplification and cropping attacks,and has good invisibility.The proposed algorithm provided a new solution for copyright protection of 3D point cloud data.