Abstract: Aiming at the building information modeling (BIM) model copyright problem, basing on the BIM data characteristics, designs a zero watermarking algorithm for BIM data, which will not change data precision. Because different modeling has its own primitives and constraint relation, the proposed algorithm must be taking full account of the difference. Firstly, the lowly constraints of the primitive points are removing and the other primitive location points are obtaining. The distance between each point benchmark and the feature point is the carrier for watermarking by choosing the feature point of the monomer precise modeling. Secondly, the algorithm bubbles sort the eigenvalues and obtains the change rate of consecutive eigenvalues, then removing the high change rate eigenvalues. Finally, dividing the eigenvalues and construct a series of zero watermarking sequence. The results of experiment showed that the algorithm had strong robustness to usual attacking such as translation, rotation and slightly deleting point.

Key words: building information model, zero watermarking, distance partition, robustness, uniqueness