A Zero-Watermarking Algorithm for High-Precision Maps Based on Feature Invariant
As a core component of vehicle navigation and autonomous driving,high-precision maps are characterized by high accuracy,sensitive information,and high cost.Ensuring their security and confidentiality is crucial.Digital watermarking,an important tool in the field of information security,can provide authentication and protection for data.However,due to the current lack of watermarking algorithms specifically designed for high-precision maps,these maps face a series of security issues during transmission and application.Aiming at the characteristics of high-precision maps,this paper proposes a zero-watermarking algorithm for high-precision maps.First,the watermarking index is determined based on the number of parametric cubic curves used to represent lane information in high-precision maps.Secondly,the watermarking information is determined based on the parity(odd or even)of the number of parametric cubic curves in two lane information nodes.Finally,combined with the watermarking index,a feature matrix is constructed according to the voting principle and the XOR operation is performed with the encrypted watermarking image to generate a zero-watermarking.Experimental results show that this algorithm can still extract relatively complete watermarking information after being subjected to high-level geometric attacks and addition and deletion attacks,demonstrating the good robustness.
万方数据
维普
