Fragile watermark generation, embedding and verification method for HD maps based on geometric features

Abstract

The present application provides a method for generating, embedding, and verifying fragile watermarks for high-precision maps based on geometric features. This relates to the field of vector geographic data processing technology. The watermark generation method includes: calculating the watermark embedding position based on the geometric features in the high-precision map document, generating watermark reference content, generating the salt value required for the hash function, generating Argon2 and Blake2b hash values, and generating road feature watermarks, document watermarks, and non-road feature watermarks. The watermarks generated and embedded by this application will not cause display anomalies or file size changes, and the introduced errors will not affect the normal use of the data.

Description

Technical Field

[0001] The present application relates to the field of vector geographic data processing technology, and in particular to a method for generating, embedding and verifying fragile watermarks for high-precision maps based on geometric features. Background Art

[0002] High-precision map data is a specialized electronic map data primarily used for high-level assisted driving and intelligent driving. It is critical infrastructure necessary for autonomous driving. As a digital twin of the driving environment, HD map data meets the needs of both autonomous vehicles and humans. Its high precision, richness, and dynamism make it irreplaceable for autonomous driving. During the distribution and sharing process, HD map data can be tampered with. Data recipients need to promptly and efficiently verify the integrity and authenticity of the data and, in emergency situations, determine the location of tampered content within the data context and coordinate system so that the untampered portion can be temporarily used. Therefore, effective technical means are urgently needed to authenticate the integrity and authenticity of HD map data and locate the location of tampered content.

[0003] While achieving data integrity and authenticity authentication, fragile watermarking technology not only integrates authentication information with data, making it an inseparable part of the data, but also can effectively infer where the data has been tampered with and to what extent. At present, although a fragile watermarking method suitable for high-precision maps in the OpenDRIVE format has been proposed, this method embeds authentication information into high-precision map data based on Unicode zero-width characters and the MD5 hash method, and has the ability to locate tampering to features. However, under browsing conditions that do not support Unicode characters, watermarked data will display abnormalities. At the same time, watermarks based on zero-width characters cause changes in file size, and measures to control file size changes limit the watermark's anti-collision ability and the accuracy of tampering positioning. Therefore, it is necessary to study a fragile watermarking method for high-precision maps that has a wider range of applicable scenarios, greater multi-format applicability, and does not cause changes in file size.

[0004] Static road network data for HD maps is vector geographic data, and the design of fragile watermarking methods for it can refer to existing fragile watermarking methods for vector geographic data. For example, some proposed fragile watermarking methods can detect whether data has been tampered with, but cannot determine the location of the tampering. Other proposed methods can locate the location where the tampering occurred, but are not resistant to the deletion of entire features or entire blocks. Others have independently proposed fragile watermarking methods that are sensitive to the deletion of entire features or entire blocks. These methods use measures such as cross-validation between block regions or cross-validation of neighboring features to infer the approximate location of the deleted elements. Furthermore, all of these methods embed watermarks without causing a change in file size. In summary, existing fragile watermarking methods for vector data are well developed. However, when directly applied to HD maps, these methods suffer from insufficient available watermark space due to limited coordinate points and insufficient representation of feature locations. Furthermore, these methods lack mechanisms for verifying the integrity of topological information.

[0005] In the face of the shortcomings of existing methods, it is necessary to take into account the geometric characteristics, topological characteristics and accuracy characteristics of high-precision maps and design a new fragile watermarking method to meet the needs of high-precision map data integrity authentication. Summary of the Invention

[0006] This application provides a method for generating, embedding and verifying fragile watermarks for high-precision maps based on geometric features. It takes OpenDRIVE high-precision map static road network data as the research object, takes into account the properties of high-precision map data, selects geometric features with controllable influence on the geometric information of high-precision maps as watermark embedding locations, generates Argon2 and Blake2b hash values ​​based on document and feature content, generates watermark information and embeds it.

[0007] In a first aspect, the present application provides a method for generating a fragile watermark for a high-precision map based on geometric features, the method comprising:

[0008] Obtaining a high-precision map document, and determining a watermark embedding position based on geometric features in the high-precision map document;

[0009] Generate a corresponding copy based on the high-precision map document, replace the watermark embedding position in the copy with 0, and obtain the document watermark reference content. Based on the document watermark reference content: separate the road element content by ID to obtain the road element watermark reference content. ; By feature tag name Separate non-road feature content and obtain non-road feature watermark reference content ; Separate file header node Content, get the document tag reference content ;

[0010] According to the road element watermark reference content , Non-road element watermark reference content and document tag reference content , get the road element watermark hash salt value , document watermark hash salt value And the non-road feature watermark hash salt value ;

[0011] According to the road element watermark hash salt value , document watermark hash salt value And the non-road feature watermark hash salt value , determine the document watermark hash value , road feature hash value , document tag hash value and non-road feature hash values ;

[0012] According to the document watermark hash value , road feature hash value , document tag hash value and / or non-road feature hash values , determine the road feature watermark, document watermark and non-road feature watermark.

[0013] Furthermore, determining the watermark embedding position according to the geometric features in the high-precision map document includes:

[0014] The geometric features in the high-precision map document are represented as , Road feature to which the geometric feature belongs ID number, For geometric features The storage order in ;

[0015] Will After the decimal point From the beginning, take The bit is used as the watermark embedding position; wherein, it is determined by the following method and :

[0016] like The value of is expressed in scientific notation, that is, , for decimal part, for The index part is based on the node label name , geometric feature name and regulatory factors to determine Value, according to The number of decimal places and to determine The calculation formula of the adjustment factor is:

[0017] .

[0018] Furthermore, according to the road element watermark reference content , Non-road element watermark reference content and document tag reference content , the road element watermark hash salt value is calculated by the following formula , document watermark hash salt value And the non-road feature watermark hash salt value :

[0019] ;

[0020] in, g Indicates taking the first row of operations, l Indicates the operation of obtaining the length of text.

[0021] Furthermore, the road element watermark is determined by the following formula:

[0022] ;

[0023] in, Indicates the road feature watermark, Indicates the file header watermark, represents the sequence concatenation operation, Indicates the element’s own watermark, Indicates the neighbor marker on the east side edge. Indicates the neighbor marker on the west edge. Indicates the north edge neighbor marker, Indicates the neighbor marker on the south edge. Indicates the hexadecimal hash value sequence of the document tag After converting the 1st to 4th digits to decimal, take the first 4 digits and reverse them to get an 8-digit decimal sequence. Indicates taking road features Hexadecimal hash value sequence After converting the 5th to 12th digits to decimal, take the first 8 digits and reverse them to get the 8-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the east edge sort After converting the 33rd to 36th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the east edge sort After converting the 49th to 52nd digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the west edge sort After converting the 41st to 44th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the west edge sort After converting the 57th to 60th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the north edge sort After converting the 37th to 40th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the north edge sort After converting the 53rd to 56th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the south edge sort After converting the 45th to 48th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the south edge sort After converting bits 61 to 64 to decimal, take the first 4 bits and reverse them to obtain a 4-digit decimal sequence.

[0024] Furthermore, the document watermark is determined by the following formula:

[0025] ;

[0026] in, Indicates a document watermark. Indicates that the document watermark hash value The resulting sequence converted to decimal.

[0027] Furthermore, the non-road feature watermark is determined by the following formula:

[0028] ;

[0029] in, Represents the non-road feature watermark, Indicates the watermark content corresponding to the last type of non-road features. Indicates the The watermark content corresponding to the non-road feature, Indicates taking the Hexadecimal hash value of non-road features Convert to base nine, take the 1st to Reverse the bit and add one bit by bit to make it non-zero digit decimal sequence, Indicates the The total number of non-road features.

[0030] In a second aspect, the present invention provides a method for embedding fragile watermarks in high-precision maps based on geometric features, the method comprising:

[0031] Obtain a high-precision map document, and generate a road element watermark, a full document watermark, and a non-road element watermark based on the above method;

[0032] Road features The first multiple watermarks are replaced by and record the watermark position occupied by the road element watermark;

[0033] Repeatedly embed the full document watermark and non-road watermark in sequence, divide the road features into continuous blocks according to the storage order, and each road feature block uses a tuple express, Indicates the sequence number of the first road feature in the block. Indicates the total number of nodes in the block;

[0034] by for The total number of watermark bits remaining in the unoccupied road feature watermarks, for Each road feature block in ,in is the document watermark length, The length of the watermark for non-road elements. For the road elements that are not enough to form a block that meets the requirements, the full document watermark and non-road watermark are not embedded. The remaining watermark positions in the document are replaced with the full document watermark and non-road feature watermarks If the content of the block ,have ,exist and The watermark bits between are filled with 0.

[0035] Furthermore, before the watermark embedding begins, the watermark embedding method includes: judging the acquired data, parsing the acquired data as high-precision map data, and if the data does not conform to the XML syntax specification during the parsing process, or if the parsing is successful but the data contains node names that do not conform to the high-precision map data format, then the data is judged to be non-high-precision map data, a prompt is output, and the document processing is refused, and the watermark embedding is terminated.

[0036] In a third aspect, the present application provides a fragile watermark verification method for high-precision maps based on geometric features, the watermark verification method comprising:

[0037] Extract existing watermarks from high-precision map documents ; Wherein, the existing watermark is embedded into the high-precision map document by the method described above;

[0038] Generate reference watermark ;

[0039] Compare the extracted watermarks with the reference watermarks one by one and generate an exception record; wherein the exception record includes the abnormal situation record value , abnormal situation record value The initial value is 0, which means no exception. According to the inconsistency of the records, the abnormal situation of the road element is recorded by bitwise OR operation with the abnormal type value and added to the abnormal record set. Retain for future use;

[0040] Locate tampering based on abnormal records.

[0041] Furthermore, tampering is located based on abnormal records, including:

[0042] Locate the added road features based on the document mark exceptions and traverse , separate document tags ,statistics , get the document tag mode and mode frequency ,if , correct , , compare and ,if , then determine the road elements To be added elements; search 、 、 and , remove from ,if ,renew ;

[0043] For road features that are not determined to be added , extract watermark space All suspected Value ,statistics , get the mode of non-road feature watermarks and mode frequency , assign non-road feature watermark extraction value , , compare and If the total number of 0s is inconsistent, it is inferred that the type of non-road feature node has changed. Otherwise, compare and ,set up Length is ,if , then the inferred label name is If the non-road feature nodes are added or missing, output this inference and the specific quantity changes. , then the inferred label name is The content of the non-road feature node is changed, and this inference is output.

[0044] Locate missing road feature nodes:

[0045] 1) Calculation 、 Coordinate range, 、 Coordinate range 、 The calculation formula is as follows:

[0046] ;

[0047] ;

[0048] in Indicates the east edge of the last element in the east edge sorting coordinate, Indicates the west edge of the first element in the west edge sorting coordinate, Indicates the north edge of the last feature in the north edge sorting coordinate, Indicates the north edge of the first element in the south edge sorting coordinate;

[0049] 2) Calculate and record the alternative unilateral boundary of the inferred location of the deleted element;

[0050] 3) Integrate the boundaries to obtain the inferred position. The boundary equations for the west, east, south, and north sides are expressed by the following formulas (a), (b), (c), and (d), respectively:

[0051] ;

[0052] Where min and max represent the minimum and maximum values ​​in the set respectively.

[0053] The method for generating, embedding, and verifying fragile watermarks for high-precision maps based on geometric features provided in this application has at least the following beneficial effects:

[0054] The watermark embedded in this application will not cause display anomalies, will not cause changes in file size, and the introduced errors will not affect the normal use of the data. When the watermarked data is tampered with, the tampered road elements can be accurately located, and the tampering of non-road elements can be accurately inferred. In the case of deletion of an entire element, the position of the deleted road elements can be inferred, and the change in the number of non-road elements can be determined. The inferred position of the deleted elements is more accurate than previous studies. This application provides a new solution for the integrity authentication of high-precision map data, and also provides a useful reference for the integrity authentication of high-precision map data in other formats. It has certain practical value for the security of high-precision map data. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0056] Figure 1 Flowchart of the method for generating fragile watermarks for high-precision maps based on geometric features provided in an embodiment of the present application;

[0057] Figure 2 A flowchart of a fragile watermark embedding method for high-precision maps based on geometric features provided in an embodiment of the present application;

[0058] Figure 3 A flowchart of a method for verifying fragile watermarks on high-precision maps based on geometric features provided in an embodiment of the present application;

[0059] Figure 4 This is a visualization of the original experimental data provided in the examples of this application;

[0060] Figure 5 A comparison chart of multi-scale visualization effects provided in the embodiments of this application;

[0061] Figure 6 A comparison chart of the text display effects of high-precision map data provided in the embodiments of this application;

[0062] Figure 7 A diagram showing the results of a simulated driving experiment provided in an embodiment of the present application;

[0063] Figure 8 This is a graph showing the results of an experiment on internal tampering of road elements provided in an embodiment of the present application;

[0064] Figure 9 This is a graph showing the comparative experimental results of deleting the entire road element provided in the embodiment of the present application;

[0065] Figure 10 This is a graph showing the experimental results of adding road elements provided in an embodiment of the present application;

[0066] Figure 11 This is a diagram of the non-road element tampering verification results provided in an embodiment of the present application.

[0067] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0068] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0069] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of information such as financial data or user data involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0070] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned. They should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0071] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0072] The embodiment of the present application provides a fragile watermarking method for high-precision maps based on geometric features, which is a fragile watermarking method for high-precision maps that embeds watermark information into parameters related to geometric features. It regards road elements with geometric features as the smallest unit for generating, embedding and detecting watermark information. Elements that do not have geometric features are verified and protected as ancillary information of the data. Research has found that modifying specific decimal places of specific geometric feature values ​​will not cause deformation or offset of the element position beyond the accuracy requirements, and can be used to embed watermark information. These specific decimal places are called watermark bits. The set of all watermark bits in an element with geometric features is called the watermark space of the element. In order to verify the integrity of the file and identify the authenticity of the file while ensuring that the embedding of the watermark does not affect the stability of the watermark itself, the watermark information contained in the watermark space changes with the content of the element after removing all watermark bits.

[0073] The fragile watermarking method for HD maps based on geometric features consists of three steps: watermark generation, watermark embedding, and watermark verification. The geometric fragile watermarking algorithm focuses on detecting and locating tampering of features with geometric features, as well as inferring the location of deleted features, without introducing deformation or offset that exceeds accuracy requirements. At the same time, it also takes into account tampering detection and location of features without geometric features.

[0074] Example 1:

[0075] The embodiment of the present application provides a method for generating fragile watermarks for high-precision maps based on geometric features. Figure 1 As shown, it is a flowchart of a method for generating fragile watermarks for high-precision maps based on geometric features provided in an embodiment of the present application. The method for generating fragile watermarks for high-precision maps based on geometric features includes the following steps S10-S50.

[0076] S10: Obtain a high-precision map document, and determine a watermark embedding position based on geometric features in the high-precision map document.

[0077] In some embodiments, the watermark embedding position is determined by:

[0078] Before the watermark generation process begins, it is necessary to check whether all the geometric features in the document can be embedded in the watermark and how many bits can be embedded in the watermark. The geometric features in the document can be represented as , Road feature to which the geometric feature belongs ID number, For its The storage order in . The total number of geometric features is . After the decimal point From 1, there are a total of bits can be used to embed a watermark. If in a document, The value of is in scientific notation, i.e. is recorded in the form of for decimal part, for The exponential part of . Then The value is also affected by the node label name , geometric feature name and regulatory factors The common influence of the value by The number of decimal places and Otherwise, Only affected 、 The value is shown in formula (1).

[0079] (1);

[0080] The specific rules for value selection are shown in Table 1. If the value is not expressed in scientific notation, Substitute into the calculation. If The name and the tag name cannot be in Table 1, then .

[0081] Table 1 Comparison table of watermark embedding starting positions

[0082]

[0083] In particular, in the cases listed in Table 1, if ,but .

[0084] The total number of available watermark bits As shown in formula (2).

[0085] (2);

[0086] In this way, all available locations for embedding watermarks are obtained.

[0087] S20: Generate a corresponding copy according to the high-precision map document, replace the watermark embedding position in the copy with 0, and obtain the document watermark reference content; according to the document watermark reference content: separate the road element content by ID, and obtain the road element watermark reference content ; By feature tag name Separate non-road feature content and obtain non-road feature watermark reference content ; Separate file header node Content, get the document tag reference content .

[0088] S30: Based on the road element watermark reference content , Non-road element watermark reference content and document tag reference content , get the road element watermark hash salt value , document watermark hash salt value And the non-road feature watermark hash salt value .

[0089] In some embodiments, the road feature watermark hash salt value , document watermark hash salt value , non-road feature watermark hash salt value It can be calculated according to the following formula:

[0090] (3);

[0091] (4);

[0092] (5);

[0093] in, Indicates taking the first row of operations, Indicates taking a string From bit, length A substring of .

[0094] S40: hashing salt value according to the road element watermark , document watermark hash salt value And the non-road feature watermark hash salt value , determine the document watermark hash value , road feature hash value , document tag hash value and the document tag hash value .

[0095] For example, this embodiment uses an existing hash algorithm to determine the hash value of the road element , document tag hash value and the document tag hash value (Argon2, Blake2b hash value).

[0096] The Argon2 hash algorithm is a memory-hard hash algorithm. To reduce its memory overhead will greatly increase its time overhead. This embodiment uses the Argon2 algorithm to generate the hash value contained in the document watermark. When applied to this algorithm, the default time cost is 8 iterations, the memory cost is 64 megabytes, the parallelism is 4 threads, the type is "Argon2_d", the version number is 19, and the output format is a 256-bit hexadecimal number. The input value is 、The salt value is Output document watermark hash value .

[0097] The Blake2 hash algorithm is a fast and secure hash algorithm. This embodiment uses the Blake2b algorithm to generate the hash values ​​contained in the road element watermark and the non-road element watermark. When using this algorithm, the default output format is a 64-bit hexadecimal number. The input value is 、The salt value is Output road feature hash value ; Input value is , the salt value is Output document token hash value ; Input value is , the salt value is Output document token hash value .

[0098] S50: Based on the document watermark hash value , road feature hash value , document tag hash value and / or document token hash value , determine the road feature watermark, document watermark and non-road feature watermark.

[0099] In some embodiments, a road feature watermark is generated by:

[0100] Calculate all road elements The boundary polygon Coordinate maximum, minimum and The maximum and minimum coordinates are sorted in ascending order according to the calculation results, and the road elements are obtained. Sorting by maximum coordinate value 、 Sorting by minimum coordinate value 、 Sorting by maximum coordinate value and Sorting by minimum coordinate value . 、 、 and Road elements exist 、 、 and To ensure that watermark embedding does not destroy the stability of the above sorting, it is stipulated that when the difference between the maximum coordinate values ​​of two elements is less than 0.02 meters, the road elements are re-sorted in ascending order of their ID numbers in the ascending order of their maximum coordinate values.

[0101] road elements Corresponding road feature watermark It can be calculated by formula (6):

[0102] (6);

[0103] in, Indicates the road feature watermark, Indicates the file header watermark, represents the sequence concatenation operation, Indicates the element’s own watermark, Indicates the neighbor marker on the east side edge. Indicates the neighbor marker on the west edge. Indicates the north edge neighbor marker, Indicates the neighbor marker on the south edge. Indicates the hexadecimal hash value sequence of the document tag After converting the 1st to 4th digits to decimal, take the first 4 digits and reverse them to get an 8-digit decimal sequence. Indicates taking road features Hexadecimal hash value sequence After converting the 5th to 12th digits to decimal, take the first 8 digits and reverse them to get the 8-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the east edge sort After converting the 33rd to 36th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the east edge sort After converting the 49th to 52nd digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the west edge sort After converting the 41st to 44th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the west edge sort After converting the 57th to 60th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the north edge sort After converting the 37th to 40th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the north edge sort After converting the 53rd to 56th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the south edge sort After converting the 45th to 48th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the south edge sort After converting bits 61 to 64 to decimal, take the first 4 bits and reverse them to obtain a 4-digit decimal sequence.

[0104] in is a hexadecimal number. It can be calculated by the following formula:

[0105] (7);

[0106] in Indicates that the hexadecimal digit sequence Treat it as a number and transcribe it in decimal form. Represents a sequence of digits Start from the last digit and take the next digit forward bits, as a new sequence of digits. and is the nearest neighbor function, and its value is:

[0107] (8);

[0108] (9);

[0109] in, is the total number of road elements. Document tag, called are self-markers, respectively 、 、 、 Mark the neighboring edges on the north, south, east, and west sides.

[0110] In some embodiments, for document watermarks , which can be calculated by the following formula:

[0111] (10);

[0112] In some embodiments, non-road feature watermarks are generated by:

[0113] Non-road feature watermark It can be calculated by formula (11):

[0114] (11);

[0115] in For the document The name of the non-road feature class. Non-road elements total, is the total number of non-road feature types in the document, is calculated as follows:

[0116] (12);

[0117] in, Indicates that the hexadecimal digit sequence is treated as a number and the number is transcribed in base 9. Represents a sequence multiplication operation, such as . To traverse the base 9 digit sequence , and are decimal digit sequences of equal length Bitwise assignment operation, if No. Position ,for No. Bit Assignment .

[0118] Example 2:

[0119] The embodiment of the present application provides 7. A fragile watermark embedding method for high-precision maps based on geometric features, such as Figure 2 As shown, the watermark embedding method includes the following steps:

[0120] Step 1: Before watermark embedding begins, the data is evaluated. Parsing is performed on the assumption that the data conforms to the standard HD map data. If the data is found to be non-compliant with XML syntax during parsing, or if parsing succeeds but the data contains node names that do not conform to the HD map data format, the data is determined to be non-HD map data, a warning is output, and the document is rejected, and watermark embedding is terminated.

[0121] Step 2: Generate watermark. Read the HD map document and replace all watermark bits with 0 to get . Split by nodes ,get 、 and Extract the salt value and get 、 and . Solve the road element boundary polygon, obtain the maximum horizontal and vertical coordinates of the road elements and sort them, and obtain the sorting results 、 、 and Generate road element watermarks according to the solution described in Example 1 , Full document watermark and non-road feature watermarks .

[0122] Step 3: embed the road element watermark into the corresponding road element node. The first 44 watermark bits are replaced by Record the watermark position occupied by the road feature watermark to ensure that the existing content will not be overwritten in the next step.

[0123] Step 4, repeat embedding the full document watermark and non-road watermark in sequence. Divide the road features into continuous blocks according to the storage order. Each road feature block can be represented by a tuple express, Indicates the sequence number of the first road feature in the block. Represents the total number of nodes in the block. for The total number of watermark positions remaining in the image that are not occupied by road feature watermarks. Calculated by formula (13):

[0124] (13);

[0125] for Each road feature block in ,in is the document watermark length, The length of the watermark for non-road elements. For the road elements that are not enough to form a block that meets this requirement, this step of watermark embedding is not performed. The remaining watermark bits are replaced by and If the content of the block ,have ,exist and The watermark bits between are filled with 0.

[0126] Step 5: Save the changes to the document. In this process, to prevent the file size from changing due to the change in the line terminator format, keep the line break type consistent with the original document.

[0127] Example 3:

[0128] The present invention provides a method for verifying fragile watermarks on high-precision maps based on geometric features. Figure 3 As shown, the watermark verification method includes the following steps:

[0129] Step 1: Before watermark verification begins, the data is evaluated. The data is parsed as if it complies with the standard HD map data. If the data is found to be non-compliant with XML syntax during parsing, or if the parsing succeeds but the data contains node names that do not conform to the HD map data format, the data is determined to be non-HD map data, a warning is output, and the document verification is rejected, aborting the watermark verification.

[0130] Step 2: Extract the existing watermark. Read the high-precision map document, determine the node watermark space from the road feature node according to Table 1, and obtain the node extraction watermark. .

[0131] Step 3: Generate a reference watermark. As described in Example 1, generate a watermark , and The watermark is then reconstructed into a form suitable for direct embedding The watermark sequence in , called the reference watermark.

[0132] Step 4: Generate abnormality records. Compare and extract watermarks one by one With reference watermark Abnormal situation record value The initial value is 0, which means no abnormal situation. According to the inconsistent records, the abnormal situation of the road element is recorded by bitwise OR operation with the abnormal type value and added to the abnormal record set. Keep it for future use.

[0133] The values ​​and meanings of geometric feature fragile watermark anomaly types are shown in Table 2:

[0134] Table 2 Geometric feature fragile watermark outliers and their corresponding meanings

[0135]

[0136] Step 5: Locate tampering based on abnormal records.

[0137] ① Changes to the content of the positioning road elements. If any , then the road elements are inferred is tampered with and outputs this inference.

[0138] ② Locate the added road elements based on the document marking anomalies. , separate document tags ,statistics , get the document tag mode and mode frequency .if , correct , , compare and ,if , infer road features Output this inference for the added element; retrieve 、 、 and , remove from .if ,renew .

[0139] ③ Locate the tampering of non-road elements. For the road elements that were not determined to be added in step ② , extract watermark space All suspected Value ,statistics , get the mode of non-road feature watermarks and mode frequency Assign non-road feature watermark extraction value , Comparison and If the total number of 0s is inconsistent, it is inferred that the type of non-road feature node has changed, and this inference and the specific type change are displayed. Otherwise, compare and .set up Length is ,if , then the inferred label name is If the non-road feature nodes are added or missing, output this inference and the specific quantity changes. Otherwise, if , then the inferred label name is The content of the non-road feature node is changed, and this inference is output.

[0140] ④ Locate missing road element nodes.

[0141] 1) Calculation 、 Coordinate extremes. 、 Coordinate range 、 The calculation formula is as follows:

[0142] (14);

[0143] (15);

[0144] in 、 express 、 The serial number is A little bit 、 coordinate.

[0145] 2) Calculate and record the alternative unilateral boundary of the deleted element’s inferred location. 、 、 or , for the alternative west, east, south, and north boundary sets 、 、 、 ,have 、 、 or .like 、 、 or ,for 、 、 、 ,have 、 、 or If there is 、 , then for the alternative east boundary, If there is 、 , then for the alternative western boundary, there is If there is 、 , then for the alternative north boundary, there is If there is 、 , then for the alternative south boundary, there is .

[0146] 3) Integrate the boundaries to obtain the inferred position. The boundary equations for the west, east, south, and north sides are expressed by formulas (a), (b), (c), and (d), respectively:

[0147] (16);

[0148] in 、 Respectively represent the set The rectangle defined by these four boundary equations is used as the inferred result for the approximate location of the deleted feature. The four intersection points are output to inform the user that the deleted feature exists and its approximate location is the rectangle defined by the lines connecting these four points.

[0149] ⑤ If the above inspection does not reveal any tampering, but there are road elements ,satisfy ,Right now If the watermark content is abnormal, it means that tampering has occurred, but the algorithm cannot determine where the tampering occurred. The output indicates that the tampering could not be successfully located.

[0150] ⑥ Output the verification result of the document. If it is empty, the data integrity is verified and the output document is prompted to pass the watermark verification. Otherwise, the output document is prompted to fail the watermark verification. And according to the user's choice, it is output or not. .

[0151] Example 4:

[0152] In order to verify whether the methods proposed in Examples 1-3 of this application are applicable to OpenDRIVE format high-precision map data, the high-precision map data containing fragile watermarks of geometric features are analyzed for imperceptibility, applicability of watermarked data, sensitivity to tampering, and accuracy in locating tampering. This example selects three copies of high-precision map data, code-named a, b, and c. The file sizes of the three copies are 2252 kilobytes, 2066 kilobytes, and 26722 kilobytes respectively, and the number of road elements contained is 279, 242, and 1055 respectively. The visualization effects of the data are as follows: Figure 4 As shown in (a), (b) and (c).

[0153] The operating system used in the experiment is Windows 11; the programming language is C++20; the development environment is Visual Studio 2022; the high-precision map data viewer is ODRViewer for VS Code; and the simulation software is the Driving Scenario Designer, an autonomous driving simulation experiment tool included with MATLAB R2024b.

[0154] In order to verify the imperceptibility of the fragile watermark with geometric features, the experiment in this embodiment analyzes the visualization effect and the text display effect respectively.

[0155] First, observe and compare the visualization effects of the data before and after watermark embedding with the naked eye. The visualization effects of data a at multiple scales are as follows: Figure 5 The numerical scale is only an approximation and does not have the meaning of quantitative analysis. Figure 5 (b), (d) and (f) show the local visualization effects of the original data at different scales. Figure 5 (a), (c) and (e) show the local visualization effects of watermark data at corresponding scales.

[0156] from Figure 5 From (a) and (b), we can see that the geometric shapes of lane lines and road boundaries containing watermark data are well maintained, with no abnormalities visible to the naked eye. Figure 5 From (c) and (d), we can see that the details of the watermarked data are basically consistent with the original data, with no abnormalities visible to the naked eye. Figure 5 As can be seen from (e) and (f), due to the slight interference when embedding the watermark, a slight misalignment can be observed at the junction of road elements. Therefore, in terms of data visualization effect, the geometric feature fragile watermark has good imperceptibility.

[0157] Secondly, observe and compare the text display effect of the data before and after the watermark is embedded. The local text display effect of data a before and after the watermark is embedded is as follows: Figure 6 As shown in (a) and (b) in . Figure 6 It can be seen that the geometric feature fragile watermark only modifies the geometric feature value. Without knowing the original value, it is difficult to detect the anomaly. Therefore, in terms of text display effect, the watermark is imperceptible.

[0158] Furthermore, since all write operations to the HD map during the watermark embedding process only involve replacing geometric parameters, and the line breaks in the watermarked document remain consistent with the original file, the file size containing the watermarked data is exactly the same as the original data. This makes the watermark imperceptible in terms of file size consistency.

[0159] In order to verify the imperceptibility of the fragile watermark with geometric features, the experiment in this embodiment is evaluated from two aspects: error statistics and simulation experiments.

[0160] First, the error introduced by the geometrically fragile watermark is analyzed. Considering that the coordinates of the road elements in the HD map data are implicit in the geometric parameters, it is difficult to represent the overall impact of the watermark on the geometric features of the HD map data by only considering the error of the starting point of the road centerline with coordinates. Therefore, the offset of the road's solution boundary point should be used as the criterion for judging the error. Since many geometric parameters are changed during the watermark embedding process, the offsets of different solution boundary points of the same element are also inconsistent. For the convenience of analysis and calculation, it is assumed that the point closest to the boundary point before solution in the solution boundary point of the data after embedding the watermark is the point that arrives after the offset of the original data solution boundary point, and an error statistical analysis is performed. In this embodiment, the step size for solving the polygon boundary is 0.01 meters. The error statistics are shown in Table 3, where the average value is the average absolute error before and after the offset, and the maximum value is the maximum absolute error between the two points.

[0161] Table 3 Error statistics (meters)

[0162]

[0163] Secondly, the Driving Scenario Designer software is used to simulate the autonomous driving scenario. Simulated driving experiments are conducted on the original data a and the watermarked data a. The same vehicles are set to travel at the same speed. The results are as follows: Figure 7 As shown, Figure 7 (a) is the test result with watermark, and (b) is the test result without watermark, which correspond to the test results of the simulated driving experiment conducted on the original data a and the watermarked data a respectively.

[0164] Depend on Figure 7 As can be seen, the simulated driving experiments were all successfully completed, with a driving time of 14.06 seconds. Therefore, it can be determined that the geometric feature fragile watermark has no impact on the function of assisting autonomous driving with the watermarked data.

[0165] Analysis of sensitivity to tampering and accuracy in locating tampering.

[0166] (1) Analysis of internal tampering positioning of road elements

[0167] Add a parking lane to the road feature with ID 23 in data a, and the lane ID is 6; delete the lane on one side of the road with ID 24, and the deleted lane ID is -1; change the length of the road feature with ID 52, extending it by 20 meters on the original basis. The effect after tampering is as follows Figure 8 (a) is shown in Figure 2. The watermark verification is performed on the tampered data a, and the verification result is as follows: Figure 8 Comparing the verification results with the actual tampering, we can see that when tampering occurs within the road elements in the data, the text watermark algorithm detects the data tampering and accurately locates the tampered road elements.

[0168] (2) Road element deletion and tampering positioning and comparative analysis

[0169] Delete the road element with ID 34 in data b and verify the tampered data. The tampering situation is as follows: Figure 9 As shown in (a), the verification results are as follows Figure 9 As shown in (b), Figure 9 (c) is the verification result of the document [7] when encountering the same tampering. The approximate location of the deleted element, the positioning result of the geometric feature fragile watermark algorithm and the positioning result of the text fragile watermark algorithm are shown as follows: Figure 9 As shown in the red, green, and orange boxes in (a). Comparing the verification results with the actual tampering, it can be seen that when the road elements are completely deleted, the text watermark algorithm and the geometric feature fragile watermark algorithm can both detect the data tampering and infer the approximate location of the tampered elements. Moreover, the geometric feature fragile watermark algorithm has improved the positioning logic based on the position relationship of the adjacent elements, and the positioning result of the deleted elements is more accurate than the text fragile watermark algorithm. Among them, the reference [7] is "Yuan Tianyang, Zhu Changqing, Chen Huixian, et al. Fragile watermark method for high-precision maps in OpenDRIVE format [J / OL]. Journal of Wuhan University (Information Science Edition): 1-14 [2024-07-04]. https: / / doi.org / 10.13203 / j.whugis20230500."

[0170] (3) Analysis of road element tampering and location

[0171] Add the road feature with ID 92 from data c to data b. The situation after tampering is as follows Figure 10 As shown in (a), the tampered data is verified, and the verification result is as follows Figure 10 Comparing the verification results with the actual tampering, it can be seen that when road elements from other data appear in the data, the text watermark algorithm detects that the data has been tampered with and accurately locates the added road elements.

[0172] (4) Non-road element tampering detection and analysis

[0173] ①Add non-road elements

[0174] Add the intersection feature from data a with ID value 831 to data b and verify the tampered data. The verification results are as follows: Figure 11 As shown in (a).

[0175] ②Delete non-road elements

[0176] Delete the intersection element with ID value 863 in data b and verify the tampered data. The verification results are as follows: Figure 11 As shown in (b).

[0177] ③Change the characteristic value of non-road elements

[0178] Change the connection relationship feature value of the intersection element with ID value 804 in data b. Verify the tampered data. The verification results are as follows: Figure 11 As shown in (c).

[0179] ④Change the file header node feature value

[0180] Change the date feature value of the data b file header node, and the non-road feature tampering verification result is as follows Figure 11 As shown in (d).

[0181] By comparing the verification results in the above experiment with the actual tampering, it can be seen that when non-road elements are tampered with, the method proposed in this application can detect that the data has been tampered with and determine the specific type of element that has been tampered with.

[0182] In summary, this embodiment addresses the integrity protection of HD map data. Using OpenDRIVE HD map data as an example, this paper applies the principle of fragile watermarking for vector geographic data. Taking into account the geometric characteristics of HD map data, this paper selects geometric features with controllable impact on HD map geometry as watermark embedding locations. Argon2 and Blake2b hash values ​​are generated based on document and feature content, and watermark information is generated and embedded. This paper proposes a fragile watermarking method for HD map data based on geometric characteristics. Experimental results show that this method embeds watermarks without causing display anomalies or file size changes, and the introduced errors do not affect the normal use of the data. When watermarked data is tampered with, it can accurately locate the tampered road features and accurately infer the tampering of non-road features. In the case of entire feature deletion, the location of the deleted road features can be inferred, and the change in the number of non-road features can be determined. The inferred location of deleted features is more accurate than previous studies. This proposed algorithm provides a new solution for HD map data integrity verification and offers a useful reference for integrity verification of HD map data in other formats. It has certain practical value for HD map data security.

[0183] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0184] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for generating fragile watermarks for high-precision maps based on geometric features, characterized in that: The method comprises: Obtaining a high-precision map document, and determining a watermark embedding position based on geometric features in the high-precision map document; Generate a corresponding copy based on the high-precision map document, replace the watermark embedding position in the copy with 0, and obtain the document watermark reference content. Based on the document watermark reference content: separate the road element content by ID to obtain the road element watermark reference content. ; By feature tag name Separate non-road feature content and obtain non-road feature watermark reference content ; Separate file header node Content, get the document tag reference content ; According to the road element watermark reference content , Non-road element watermark reference content and document tag reference content , get the road element watermark hash salt value , document watermark hash salt value And the non-road feature watermark hash salt value ; According to the road element watermark hash salt value , document watermark hash salt value And the non-road feature watermark hash salt value , determine the document watermark hash value , road feature hash value , document tag hash value and non-road feature hash values ; According to the document watermark hash value , road feature hash value , document tag hash value and / or non-road feature hash values , determine the road feature watermark, document watermark and non-road feature watermark.

2. The method for generating fragile watermarks for high-precision maps based on geometric features according to claim 1 is characterized in that: Determining a watermark embedding position according to geometric features in the high-precision map document includes: The geometric features in the high-precision map document are represented as , Road feature to which the geometric feature belongs ID number, For geometric features The storage order in ; Will After the decimal point From the beginning, take The bit is used as the watermark embedding position; wherein, it is determined by the following method and : like The value of is expressed in scientific notation, that is, , for decimal part, for The index part is based on the node label name , geometric feature name and regulatory factors to determine Value, according to The number of decimal places and to determine The calculation formula of the adjustment factor is: 。 3. The method for generating fragile watermarks for high-precision maps based on geometric features according to claim 1 is characterized in that: According to the road element watermark reference content , Non-road element watermark reference content and document tag reference content , the road element watermark hash salt value is calculated by the following formula , document watermark hash salt value And the non-road feature watermark hash salt value : ； in, g Indicates taking the first row of operations, l Indicates the operation of obtaining the length of text.

4. The method for generating fragile watermarks for high-precision maps based on geometric features according to claim 1 is characterized in that: The road feature watermark is determined by the following formula: ； in, Indicates the road feature watermark, Indicates the file header watermark, represents the sequence concatenation operation, Indicates the element’s own watermark, Indicates the neighbor marker on the east side edge. Indicates the neighbor marker on the west edge. Indicates the north edge neighbor marker, Indicates the neighbor marker on the south edge. Indicates the hexadecimal hash value sequence of the document tag After converting the 1st to 4th digits to decimal, take the first 4 digits and reverse them to get an 8-digit decimal sequence. Indicates taking road features Hexadecimal hash value sequence After converting the 5th to 12th digits to decimal, take the first 8 digits and reverse them to get the 8-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the east edge sort After converting the 33rd to 36th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the east edge sort After converting the 49th to 52nd digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the west edge sort After converting the 41st to 44th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the west edge sort After converting the 57th to 60th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the north edge sort After converting the 37th to 40th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the north edge sort After converting the 53rd to 56th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the previous road feature in the south edge sort After converting the 45th to 48th digits to decimal, take the first 4 digits and reverse them to get a 4-digit decimal sequence. Indicates taking road features Sequence of hexadecimal hash values ​​of the next road feature in the south edge sort After converting bits 61 to 64 to decimal, take the first 4 bits and reverse them to obtain a 4-digit decimal sequence.

5. The method for generating fragile watermarks for high-precision maps based on geometric features according to claim 1 is characterized in that: The document watermark is determined by the following formula: ； in, Indicates a document watermark. Indicates that the document watermark hash value The resulting sequence converted to decimal.

6. The method for generating fragile watermarks for high-precision maps based on geometric features according to claim 1 is characterized in that: The non-road feature watermark is determined by the following formula: ； in, Represents the non-road feature watermark, Indicates the watermark content corresponding to the last type of non-road features. Indicates the The watermark content corresponding to the non-road feature, Indicates taking the Hexadecimal hash value of non-road features Convert to base nine, take the 1st to Reverse the bit and add one bit by bit to make it non-zero digit decimal sequence, Indicates the The total number of non-road features.

7. A fragile watermark embedding method for high-precision maps based on geometric features, characterized by: The watermark embedding method comprises: Obtaining a high-precision map document, and generating a road element watermark, a full-document watermark, and a non-road element watermark based on the method according to any one of claims 1 to 6; Road features The first multiple watermarks are replaced by and record the watermark position occupied by the road element watermark; Repeatedly embed the full document watermark and non-road watermark in sequence, divide the road features into continuous blocks according to the storage order, and each road feature block uses a tuple express, Indicates the sequence number of the first road feature in the block. Indicates the total number of nodes in the block; by for The total number of watermark bits remaining in the unoccupied road feature watermarks, for Each road feature block in ,in is the document watermark length, The length of the watermark for non-road elements. For the road elements that are not enough to form a block that meets the requirements, the full document watermark and non-road watermark are not embedded. The remaining watermark positions in the document are replaced with the full document watermark and non-road feature watermarks If the content of the block ,have ,exist and The watermark bits between are filled with 0.

8. The fragile watermark embedding method for high-precision maps based on geometric features according to claim 7 is characterized in that: Before watermark embedding begins, the watermark embedding method includes: judging the acquired data, parsing the acquired data as high-precision map data, and if the data does not conform to the XML syntax specification during the parsing process, or if the parsing is successful but the data contains node names that do not conform to the high-precision map data format, then the data is determined to be non-high-precision map data, a prompt is output, and the document processing is refused, and the watermark embedding is terminated.

9. A fragile watermark verification method for high-precision maps based on geometric features, characterized in that: The watermark verification method comprises: Extract existing watermarks from high-precision map documents ; Wherein, the existing watermark is embedded into the high-precision map document by the method described in claim 7 or 8; Generate reference watermark ; Compare the extracted watermarks with the reference watermarks one by one and generate an exception record; wherein the exception record includes the abnormal situation record value , abnormal situation record value The initial value is 0, which means no exception. According to the inconsistency of the records, the abnormal situation of the road element is recorded by bitwise OR operation with the abnormal type value and added to the abnormal record set. Retain for future use; Locate tampering based on abnormal records.

10. The method for verifying a fragile watermark for a high-precision map based on geometric features according to claim 9, characterized in that: Locate tampering based on abnormal records, including: Locate the added road features based on the document mark exceptions and traverse , separate document tags ,statistics , get the document tag mode and mode frequency ,if , correct , , compare and ,if , then determine the road elements To be added elements; search 、 、 and , remove from ,if ,renew ; For road features that are not determined to be added , extract watermark space All suspected Value ,statistics , get the mode of non-road feature watermarks and mode frequency , assign non-road feature watermark extraction value , , compare and If the total number of 0s is inconsistent, it is inferred that the type of non-road feature node has changed. Otherwise, compare and ,set up Length is ,if , then the inferred label name is If the non-road feature nodes are added or missing, output this inference and the specific quantity changes. , then the inferred label name is The content of the non-road feature node is changed and the inference is output; Locate missing road feature nodes: 1) Calculation 、 Coordinate range, 、 Coordinate range 、 The calculation formula is as follows: ； ； in Indicates the east edge of the last element in the east edge sorting coordinate, Indicates the west edge of the first element in the west edge sorting coordinate, Indicates the north edge of the last feature in the north edge sorting coordinate, Indicates the north edge of the first element in the south edge sorting coordinate; 2) Calculate and record the alternative unilateral boundary of the inferred location of the deleted element; 3) Integrate the boundaries to obtain the inferred position. The boundary equations for the west, east, south, and north sides are expressed by the following formulas (a), (b), (c), and (d), respectively: ； Where min and max represent the minimum and maximum values ​​in the set respectively.

Images