CN-122023097-A - Geometric attack resistant vector geographic point data lossless watermark generation and extraction method

CN122023097ACN 122023097 ACN122023097 ACN 122023097ACN-122023097-A

Abstract

The invention discloses a geometric attack resistant vector geographic point data lossless watermark generation and extraction method, which comprises the steps of firstly generating a Thiessen polygon of vector point data before watermark embedding, wherein each point element corresponds to a unique Thiessen polygon, secondly determining the storage direction of a point according to the relationship of the Thiessen polygon circumferences of adjacent point elements, establishing a hash mapping relationship with watermark information by normalizing the sum of the Thiessen polygon circumferences of the adjacent points, and finally rearranging the storage sequence of the point elements to embed the watermark. When the watermark is detected, generating a Thiessen polygon of vector point data to be detected, and extracting the watermark according to the storage directions of adjacent point elements. Experimental results show that the algorithm has good robustness to translation, scaling and rotation, and as coordinate values are not changed, lossless embedding of watermarks is realized, the requirement of high-precision data is met, and the algorithm has good practicability.

Inventors

LI YUJIAN
LV WENJUN
Xu Xiaozuo
Shi Pixu
GAO YING
Han Jiangyong
HOU JUNLING

Assignees

中铁第一勘察设计院集团有限公司

Dates

Publication Date: 20260512
Application Date: 20251201

Claims (10)

1. A method for generating a geometric attack resistant vector geographic point data lossless watermark, the method comprising: acquiring an original binary watermark image, scrambling the watermark image and converting the scrambled watermark image into a one-dimensional watermark sequence; vector geographic point data are obtained, and unique Thiessen polygons of all point elements are constructed; determining the storage direction values of two adjacent point pairs according to the comparison result of the Thiessen polygon circumferences of the two adjacent point pairs; acquiring normalized values of the sum of the Thiessen polygon circumferences of two adjacent point pairs, establishing a hash mapping relation between the normalized values and watermark bits in a one-dimensional watermark sequence, and determining watermark embedding positions; And rearranging the storage sequence of the two adjacent point pairs according to whether the storage direction value of the two adjacent point pairs is consistent with the value of the corresponding watermark bit, and finally generating vector geographic point data containing the watermark.
2. The method for generating the geometric attack resistant vector geographic point data lossless watermark according to claim 1, wherein the method comprises the steps of obtaining an original binary watermark image, scrambling the watermark image, and converting the scrambled watermark image into a one-dimensional watermark sequence, and specifically comprises the following steps: Using Arnold transformation to disorder watermark image and transform into one-dimensional watermark sequence Wherein Is the length of the watermark, Is the value of a single watermark bit and i is the position-location index to the one-dimensional sequence.
3. The method for generating the geometric attack resistant vector geographic point data lossless watermark according to claim 1, wherein the determining the storage direction of the points according to the comparison result of the sizes of the Thiessen polygon circumferences of two adjacent point pairs specifically comprises: if the Thiessen polygon perimeter of the previous point is larger than that of the subsequent point, the direction value is stored as 1, otherwise, the direction value is stored as 0.
4. The method for generating the vector geographic point data lossless watermark resistant to geometric attacks according to claim 1, wherein obtaining a normalized value of a sum of the circumferences of Thiessen polygons of two adjacent point pairs, and establishing a hash mapping relation between the normalized value and watermark bits in a one-dimensional watermark sequence, and determining a watermark embedding position comprises the following steps: the sum of the Thiessen polygon circumferences of two adjacent point pairs is subjected to minimum maximum normalization, and the normalized value is amplified Establishing a hash mapping relation with watermark bits in a one-dimensional watermark sequence, and determining a watermark embedding position; The value of (2) is determined by the watermark length; The mapping relation is composed of Determining, wherein Amplifying the normalized value The number of times of the number of times, Is the watermark length.
5. The method for generating the geometric attack resistant vector geographic point data lossless watermark according to claim 1, wherein the storage sequence of two adjacent point pairs is rearranged according to whether the storage direction values of the two adjacent point pairs are consistent with the values of corresponding watermark bits, and finally the vector geographic point data containing the watermark is generated, which specifically comprises the following steps: when the storage direction value of the two adjacent point pairs is consistent with the value of the watermark bit mapped by the two adjacent point pairs, the storage sequence of the adjacent point pairs is not exchanged, otherwise, the storage sequence of the adjacent point pairs is exchanged.
6. The method for extracting the vector geographic point data lossless watermark resistant to geometric attack is characterized by comprising the following steps of: obtaining vector geographic point data containing watermarks, and constructing unique Thiessen polygons of all point elements; acquiring normalized values of the sum of the Thiessen polygon circumferences of two adjacent point pairs, establishing a hash mapping relation between the normalized values and watermark bits in a one-dimensional watermark sequence, and determining watermark embedding positions; Extracting the value of a single watermark bit in sequence according to watermark embedding positions, wherein the value of a final watermark bit is determined by adopting a voting principle aiming at the situation that a plurality of adjacent point pairs are mapped to the same watermark bit, so as to obtain a one-dimensional watermark sequence; and carrying out dimension lifting processing and scrambling on the extracted one-dimensional watermark sequence to obtain a final binarization watermark image.
7. The geometric attack resistant vector geographic point data lossless watermark extraction method according to claim 6, wherein obtaining a normalized value of a sum of Thiessen polygon circumferences of two adjacent point pairs, and establishing a hash mapping relation between the normalized value and watermark bits in a one-dimensional watermark sequence, and determining a watermark embedding position, specifically comprises: the sum of the Thiessen polygon circumferences of two adjacent point pairs is subjected to minimum maximum normalization, and the normalized value is amplified Establishing a hash mapping relation with watermark bits in a one-dimensional watermark sequence, and determining a watermark embedding position; The value of (2) is determined by the watermark length.
8. The method for extracting the lossless watermark of the vector geographic point data resistant to the geometric attack as defined in claim 6, wherein the value of a final watermark bit is determined by adopting a voting principle aiming at the condition that a plurality of adjacent point pairs are mapped to the same watermark bit, so as to obtain a one-dimensional watermark sequence, and the method specifically comprises the following steps: A one-dimensional sequence of equal length to the length of the embedded watermark is first defined, Wherein Is the watermark length; For the initial value of each watermark bit, the initial value is 0;i which is the position locating index of the defined one-dimensional sequence; Watermark values mapped to the same location use an accumulation formula To make statistics in which Is composed of The rule is determined as if Then Otherwise ; The value is the value of each watermark extraction, and the value is only 0 and 1; From the following components Assigning values, wherein the values are only 1 and-1, and finally if The corresponding watermark bit has a value of 1, or else, the corresponding watermark bit has a value of 0, and the calculation formula is as follows: Wherein the method comprises the steps of The perimeter of the Thiessen polygon generated for the previous point element in the two adjacent point pairs, Generating a Thiessen polygon perimeter for a next dot element in two adjacent dot pairs; The storage direction value of two adjacent point pairs is used; For the final value of the watermark bit, only 0 and 1;i are the position location indexes of the defined one-dimensional sequence.
9. A geometric attack resistant vector geographic point data lossless watermark generation system, the system comprising: the one-dimensional sequence generation module is used for acquiring an original binary watermark image, scrambling the watermark image and converting the scrambled watermark image into a one-dimensional watermark sequence; The Thiessen polygon generation module is used for acquiring vector geographic point data and constructing unique Thiessen polygons of all point elements; the storage direction determining module is used for determining the storage direction values of the two adjacent point pairs according to the comparison result of the Thiessen polygon perimeter sizes of the two adjacent point pairs; the watermark mapping module is used for obtaining the normalized value of the sum of the Thiessen polygon circumferences of two adjacent point pairs, establishing a hash mapping relation between the normalized value and watermark bits in the one-dimensional watermark sequence, and determining the watermark embedding position; and the sequence exchange module is used for rearranging the storage sequence of the two adjacent point pairs according to whether the storage direction value of the two adjacent point pairs is consistent with the value of the corresponding watermark bit, and finally generating vector geographic point data containing the watermark.
10. A geometric attack resistant vector geographic point data lossless watermark extraction system, the system comprising: The Thiessen polygon generation module is used for acquiring vector geographic point data containing the watermark and constructing unique Thiessen polygons of all point elements; the watermark mapping module is used for obtaining the normalized value of the sum of the Thiessen polygon circumferences of two adjacent point pairs, establishing a hash mapping relation between the normalized value and watermark bits in the one-dimensional watermark sequence, and determining the watermark embedding position; the one-dimensional sequence generating module is used for sequentially extracting the values of single watermark bits according to watermark embedding positions, wherein the values of final watermark bits are determined by adopting a voting principle aiming at the situation that a plurality of adjacent point pairs are mapped to the same watermark bit, so as to obtain a one-dimensional watermark sequence; and the watermark acquisition module is used for carrying out dimension lifting processing and scrambling on the extracted one-dimensional watermark sequence to obtain a final binarization watermark image.

Description

Geometric attack resistant vector geographic point data lossless watermark generation and extraction method Technical Field The invention relates to the technical field of vector geographic data copyright protection, in particular to a geometric attack-resistant vector geographic point data lossless watermark generation and extraction method. Background Vector geographic data is taken as important data in the fields of mapping, remote sensing, GIS and the like, and is an indispensable strategic resource in national economy and national defense construction. With the increasing sharing demand of geographic data, illegal use of vector geographic data and infringement events frequently occur, so that effective technical means are urgently needed for copyright protection and tracing. At present, the security of protecting vector geographic data by utilizing a digital watermarking technology has many research results at home and abroad, and the watermark algorithm can be divided into a space domain, a transformation domain and a geometric domain according to different embedded domains. The spatial domain watermark algorithm is used for directly embedding watermark information into the spatial coordinates of data, has strong capability of resisting adding and deleting points and cutting attacks, the transform domain watermark algorithm is used for embedding watermark information into coefficients of a transform domain, can resist noise, translation and rotation attacks to a certain extent, and has good robustness against geometric attacks when the geometric domain watermark algorithm is used for embedding the watermark into the geometric characteristics of geographic data. However, as one of the common attack modes of vector geographic data, many scholars have proposed different algorithms at present, and although the geometric attack can be resisted to a certain extent, the accuracy of the spatial data is lost. In the prior art, a lossless watermark algorithm for reordering storage sequence is proposed, watermark is embedded by changing the storage sequence of adjacent line elements or surface elements, watermark information lossless embedding is realized because the watermark embedding process does not change the coordinate value, but the definition of storage direction in the algorithm is to use the vertex number of adjacent line data or surface data, so the algorithm is only suitable for line data and surface data, and point data cannot be used. Based on the above problems, the invention provides a lossless watermarking algorithm suitable for vector point data, which utilizes vector point data to be embedded with watermark to generate Thiessen polygons, determines the storage direction according to the size relation of the circumferences of the Thiessen polygons, and embeds watermark information by rearranging the storage sequence, thereby not only resisting any geometric transformation, but also realizing lossless embedding of the watermark information. Disclosure of Invention The application provides a geometric attack resistant vector geographic point data lossless watermark generation and extraction method, which aims to solve the problems that in the prior art, space data precision is lost, the method is only suitable for line data and surface data, point data cannot be used and the like. According to a first aspect, in one embodiment, there is provided a geometric attack resistant vector geographic point data lossless watermarking method, the method comprising: acquiring an original binary watermark image, scrambling the watermark image and converting the scrambled watermark image into a one-dimensional watermark sequence; vector geographic point data are obtained, and unique Thiessen polygons of all point elements are constructed; determining the storage direction values of two adjacent point pairs according to the comparison result of the Thiessen polygon circumferences of the two adjacent point pairs; acquiring normalized values of the sum of the Thiessen polygon circumferences of two adjacent point pairs, establishing a hash mapping relation between the normalized values and watermark bits in a one-dimensional watermark sequence, and determining watermark embedding positions; And rearranging the storage sequence of the two adjacent point pairs according to whether the storage direction value of the two adjacent point pairs is consistent with the value of the corresponding watermark bit, and finally generating vector geographic point data containing the watermark. Further, the method comprises the steps of obtaining an original binary watermark image, scrambling the watermark image and converting the scrambled watermark image into a one-dimensional watermark sequence, and specifically comprises the following steps: Using Arnold transformation to disorder watermark image and transform into one-dimensional watermark sequence WhereinIs the length of the watermark,Is the value of a single watermark bit a