CN109903213B - Digital watermark system for three-dimensional model file of photocuring stereolithography - Google Patents

Abstract

A digital watermarking system for light-curing stereoscopic three-dimensional model files belongs to the field of multimedia content security. The invention discloses a digital watermarking system for STL three-dimensional model files, which can effectively resist rotation, scaling, and patch replay. Sorting and other attacks, the digital watermarking system consists of two parts: 1) a digital watermark embedding part; 2) a digital watermark extraction part, the present invention can successfully extract the watermark STL from the watermark STL that is attacked by rotation, scaling, triangular facet reordering, etc. Extracting digital watermarks from 3D models can be effectively used for copyright protection.

Description

A digital watermarking system for 3D model files of stereolithography

technical field

The invention belongs to the field of multimedia content security, and in particular relates to a digital watermark system for three-dimensional model files of photocuring three-dimensional modeling.

Background technique

The rapid development of technologies such as the Internet and digital media makes it easier for people to collaborate and publish various forms of multimedia products through the network, including text, images, audio and video, as well as three-dimensional (referred to as 3D) models. While the convenient network technology brings convenience to the majority of users, the infringements of illegal possession, copying, modification, and dissemination of pirated products have also become easier. These infringements have seriously infringed upon the legitimate rights and interests of copyright owners. As one of the important means of digital copyright protection, digital watermarking technology has been widely studied since the 1990s. Its basic idea is to embed copyright information, product identification codes, and buyer information into original data files. Under normal circumstances, this information is not perceived by people and does not affect the use of digital products. It is only extracted by watermark extraction algorithm to identify copyright when necessary. At the same time, the digital watermark must be able to withstand certain attacks without being cleared.

In recent years, the rapid development of technologies such as computer-aided design, virtual reality, 3D image processing, and graphics computing has made 3D models widely used in many fields such as architectural design, cultural relic protection, and game model design. With the maturity and popularization of 3D printing technology, the creation and production of personalized products become possible, and the operation mode of the whole society and the direction of industrial development are undergoing tremendous changes. Therefore, the copyright protection of 3D models is becoming more and more important.

3D printing is a kind of rapid prototyping technology. The STL 3D model file used by it is a standard file between software design and printer. The STL format file is composed of a series of triangular surface information units, and each information unit includes The coordinate values of the outer normal vector of the triangle facet starting with "facet normal" and the coordinate values of the three vertices starting with "vertex", as shown in Figure 1. As an interface protocol for collaboration between 3D model design and 3D printers, STL 3D model files are a very popular digital product. In addition, there are very few digital watermarking techniques and systems for STL 3D model files. Therefore, it is particularly important and urgent to design a digital watermarking system for STL 3D model files.

Contents of the invention

The invention aims at the deficiencies of the prior art, and aims to provide a digital watermark system for three-dimensional model files of photocuring stereoplasty. The digital watermarking system of the present invention can effectively resist attacks such as rotation, scaling, triangular patch reordering and the like.

In order to solve the above-mentioned technical problems, the technical solution adopted in the present invention is: a digital watermarking system for three-dimensional model files of photocuring stereoplasty, which is characterized in that it includes two parts:

The first part: digital watermark embedding part

The watermark embedding part specifically includes the following three main steps:

Step 1. Redundant coding of watermark sequence

1) A watermark sequence of length L representing copyright information is expressed as W={w ₁ ,w ₂ ,w ₃ ,..., _wi ,...,w _L }, where W _i ∈[0,1] , if L is an odd number, add a "0" after w _L ;

Adjacent two bits form a group to encode the watermark sequence W, where "00" is encoded as "1", "01" is encoded as "2", "10" is encoded as "3", and "11" is encoded as "4". ", to get the encoded watermark sequence

其中W_i∈[0,4]；

where W _i ∈ [0,4];

2) Add check digits for the first time: add check digits between adjacent repetitions or neighbors (repeated numbers are not adjacent but within 4 digits) in the watermark sequence W ₁ , and adjacent repetitions Next, first add a check digit between the two repeated numbers, then change the second repeated number to its paired number (the definition of the paired number is given in the specific implementation plan); in the case of adjacent repetitions, first add a check digit, and then repeat the last number of the upper unit to obtain a verified watermark sequence W2;

3) Adding the check digit for the second time: After adding the check digit for the first time, the complete and continuous watermark sequence W ₁ is separated by the check digit, and here the set of digital elements between the two check digits is called a unit, And the number of digital elements in a unit does not exceed 4;

3.1 In the watermark sequence W ₂ , add different numbers of check digits after the units with 2 digital elements, 3 digital elements and 4 digital elements;

3.2 If and only if the numbers before and after the check digit are paired numbers, first insert the end number of the unit repeatedly before the original check digit, and then add a check digit after the newly inserted number to obtain the watermark sequence of the second check W ₃ , at this time, there are one or more check digits between the units, and the unit is still only a collection of digital elements between the check digits, and the number of digital elements in a unit does not exceed 5, no matter what In either case, the check digit itself is not included in the cell range;

4) Code binarization: restore the digital elements in the watermark sequence W ₃ of the secondary verification to two adjacent binary code sequences as a group, and obtain the binary sequence W ₄ containing watermark information;

Step 2: Sort the triangles of the STL 3D model; because the STL 3D model may suffer from the reordering attack of the triangles, but the reordering will neither change the geometric structure of the 3D model nor affect the use of the 3D model. Therefore, in order to eliminate the impact of reordering on the watermarking system, it is necessary to sort the triangles according to the properties of the triangles.

Step 3: Redundant coded watermark sequence embedding: adopt the method of reordering the vertices of the triangular faces of the STL three-dimensional model to realize the embedding of the watermark sequence;

The second part: digital watermark extraction part

The watermark extraction part specifically includes the following three main steps:

Step 1. Sorting the triangular patches of the STL 3D model;

Step 2, digital watermark sequence extraction, the specific steps are as follows:

1) Compare the size of the vertex coordinates of the triangular surface. The comparison method is to compare the x coordinates first. When the x coordinates are the same, compare the y coordinates. direction, and the other two vertices are named B and C vertices in order to determine the A, B and C vertices of the triangle face;

2) Extract the binary sequence Q ₁ containing watermark information according to the vertex sorting, and encode the binary sequence information Q ₁ every two adjacent bits into a group, where "00" is coded as "1" and "01" is coded as "2 ", "10" are coded as "3", "11" are coded as "4", and the watermarked information sequence Q ₂ is obtained;

Step 3. Digital watermark sequence verification and correction

1) Check the information sequence Q ₂ from the two aspects of the number itself and the number of check digits, locate the changed place and correct it, the steps are as follows:

The number itself: the number itself: the numbers connected at the head of the upper unit and the end of the lower unit of the adjacent unit must be consistent (consistency includes the two cases where the connected numbers are the same number and paired numbers), and the numbers in the unit cannot be repeated (except for adjacent two numbers). When the number of the first and last units of a unit is a pairing number, the last two digits of the previous unit must be repeated), and the last two digits of the first unit must be repeated when the numbers of two units are a pairing number;

In terms of the number of check digits: check whether the number of check digits is consistent with the value of the individual unit digital elements. When the number of unit digital elements is 1, there is only one check digit; when the number of unit digital elements is 2 or 3, the check digit The number of check digits is consistent with the number at the end; when the number of unit digital elements is 3 or 4, the number of check digits is consistent with the sum of the second and third digits; when the number of unit digital elements is 4 or 5 and the second When the first and third digits are arranged in sequence, there are two check digits, and when the second and third digits are arranged in reverse order, there is only one check digit;

2) Judging and correcting the non-conforming numbers according to the two standards, and obtaining the corrected watermarked information sequence Q ₃ .

Furthermore, for the STL 3D model triangle patch sorting in the digital watermark embedding part and the digital watermark extraction part, no matter what reordering attack the STL 3D model suffers from, the sorting here can ensure that the triangle patch sorting is consistent. Theoretically speaking, there are many sorting methods for triangular patches that meet the above requirements. Here we only take the method based on PCA as an example. The specific steps are as follows:

1) PCA (Principal Component Analysis, abbreviation for principal component analysis) preprocessing is performed on the triangular surface of the STL three-dimensional model.

2) Calculate the coordinate value of each triangle surface in the PCA coordinate system, sort the corresponding triangle surfaces from large to small according to their values, and obtain the triangle surface sequence F;

2.1) In the Cartesian coordinate system, the model is translated so that the center of gravity of the model coincides with the origin of the coordinate system.

2.2) Perform PCA processing on the coordinate values of the triangular surfaces in the model to obtain three eigenvectors.

2.3) Using the three eigenvectors as the coordinate axes of the PCA space, map the model from the Cartesian coordinate system to the PCA coordinate system.

2.4) Obtain a known triangular patch sequence F by comparing the center coordinates of the triangular faces in the model. The comparison method is to first compare the center coordinates of the triangle patches. If the center coordinates are the same, compare the x coordinates. If the x coordinates are the same, compare the y coordinates. If the y coordinates are the same, compare the z coordinates.

Further, the step three of the digital watermark embedding part, the specific steps of embedding the redundant coded watermark sequence are as follows:

1) Compare the size of the vertex coordinates of the triangular surface, and determine the A, B, and C vertices of the triangular surface. The vertex with the largest vertex coordinate value is named A vertex. Since the vertices of the STL three-dimensional model can only be sorted counterclockwise along the outer normal vector, Therefore, when the A, B, and C vertices of the triangular surface are determined, there are only three ways to sort the vertices. According to the counterclockwise direction, the other two vertices are named B and C vertices in turn;

2) Encoding each sorting method;

3) By modifying the order of vertices of each triangle in the triangle sequence F and embedding the binary sequence W4 containing watermark information, the STL 3D model embedded with the redundant coded watermark sequence is obtained.

Further, the coding rules for the sorting method are: the order of the vertices of the triangle is A, B, C, and the code is 0; the order of the vertices of the triangle is B, C, A, and the code is 1, and the order of the vertices of the triangle is C, A , B, coded as a check digit.

In order to verify the effectiveness of the present invention, the STL 3D model with embedded watermarks has been rotated, scaled, randomly reordered triangle facets and random triangle facet vertex reordered attacks, and extracted from the attacked STL 3D model watermark sequence. Experimental results show that this method can completely and correctly detect embedded digital watermark sequences from embedded watermarked STL 3D models subjected to rotation, scaling, random triangle patch reordering attacks, or vertex reordering attacks within any triangle patch.

Instructions attached

Figure 1 is a schematic diagram of an STL 3D model file.

Fig. 2 is a structural block diagram of the digital watermarking system of the present invention.

Figure 3 shows the coding rules for vertex sorting.

Detailed ways

In order to make the purpose, features and advantages of the present invention more comprehensible, specific implementations of the present invention will be described in detail below.

A digital watermarking system for three-dimensional model files of stereolithography, as shown in Figure 2, includes the following two parts:

The first part: digital watermark embedding part

The watermark embedding part specifically includes the following three main steps:

Step 1. Redundant coding of watermark sequence

1) A watermark sequence of length L representing copyright information is expressed as W={w ₁ ,w ₂ ,w ₃ ,..., _wi ,...,w _L }, where W _i ∈[0,1] , if L is an odd number, add a "0" after w _L ;

Adjacent two bits form a group to encode the watermark sequence W, where "00" is encoded as "1", "01" is encoded as "2", "10" is encoded as "3", and "11" is encoded as "4". ", to get the encoded watermark sequence

其中W_i∈[0,4]；

where W _i ∈ [0,4];

2) The check digit is added for the first time. A check bit is added between adjacent repeated or adjacent repeated numbers in the watermark sequence W ₁ (the repeated numbers are not adjacent but within a distance of 4 bits). In the case of adjacent repetitions, first add a check digit between the two repeated numbers, and then change the latter repeated number to its paired number (due to rearranging the order of vertices of any triangular facet, the random watermark sequence W1 The numbers "1" and "4", "2" and "3" cannot be changed and replaced with each other, and they are called paired numbers here. "1" and "4" are paired numbers with each other, and "2" and "3" are paired with each other. is a paired number.); in the case of adjacent repetition, first add a check digit before the adjacent repeated number, and then add the last digit of the upper unit to obtain a verified watermark sequence W ₂ .

3) The check digit is added for the second time. After the check digit is added for the first time, the complete and continuous watermark sequence W ₁ is separated by the check digit. Here, the set of digital elements between the two check digits is called a unit, and the number of digital elements in a unit does not exceed 4 .

3.1) In the watermark sequence W ₂ , different numbers of check digits are added after the units with 2 digital elements, 3 digital elements and 4 digital elements according to the situation. When the number of unit digital elements is 2, re-add the number of check digits that match the number of elements at the end of the unit; when the number of unit digital elements is 3, re-add the number of check digits that match the sum of the second and third digits of the unit Bit; when the number of unit digital elements is 4, if the middle two digits are arranged in sequence, that is, "1,2", "1,3", "1,4", "2,3", "2,4" , "3, 4", add a check digit again, otherwise it will not be added.

3.2) If and only when the numbers before and after the check digit (one or more check digits) are paired numbers, insert the ending number of the unit again before the first check digit. The watermark sequence W ₃ of the secondary verification is obtained. At this time, there are one or more check digits between the units, and the unit is still only a set of digital elements between the check digits, and a unit digital element The number does not exceed 5, and the check digit itself is not included in the cell range in any case.

4) Encoding binarization. The digital elements in the watermark sequence W ₃ of the secondary verification are restored to a binary coded sequence of two adjacent ones, and the binary sequence W ₄ containing watermark information is obtained.

Step 2: Sorting the triangular patches of the STL 3D model

There are many kinds of triangular facet sorting methods that can be selected in the present invention, as long as it can satisfy any reordering attack that the STL three-dimensional model has suffered, the sorting here can ensure that the sorting of the triangular facets is consistent. The present invention only uses PCA-based method as an example.

1) PCA (Principal Component Analysis, abbreviation for principal component analysis) preprocessing is performed on the triangular surface of the STL three-dimensional model.

2) Calculate the coordinate value of each triangle in the PCA coordinate system, sort the corresponding triangles from large to small according to their values, and obtain the triangle sequence F.

2.1) In the Cartesian coordinate system, the model is translated so that the center of gravity of the model coincides with the origin of the coordinate system.

2.2) Perform PCA processing on the coordinate values of the triangular surfaces in the model to obtain three eigenvectors.

2.3) Using the three eigenvectors as the coordinate axes of the PCA space, map the model from the Cartesian coordinate system to the PCA coordinate system.

2.4) By comparing the center coordinates of the triangular faces in the model and arranging them from large to small, a known triangular face sequence F is obtained. Specifically, if the center coordinate values are the same, compare the x coordinate values, if the x coordinate values are the same, compare the y coordinate values, and if the y coordinate values are the same, then compare the z coordinate values.

Step 3: Redundant encoded watermark sequence embedding

1) Embed the watermark sequence W3 of the secondary verification into the vertex sorting of the triangular patch of the STL three-dimensional model. The code of the watermark sequence W3 after the secondary verification is binarized, and each digital element is restored to a group of two adjacent binary codes to obtain the binary sequence W4 containing watermark information.

2) Compare the coordinate values of the vertices of the triangle face, and determine the A, B, and C vertices of the triangle face. Among them, the vertex with the largest vertex coordinate value is named A vertex. Due to the characteristics of the STL three-dimensional model, the vertices can only be counterclockwise along the outer normal vector Sort in a counterclockwise direction, and the remaining two vertices are named B and C vertices in turn, as shown in Figure 3.

3) After the vertices A, B, and C of the triangular face are determined, there are only three sorting ways of the vertices, and each sorting way is coded. In particular, the encoding criteria shown in the table in Fig. 3 can be used to encode the triangle patch by adjusting the vertex sorting, so that a triangle patch can accommodate 1 bit information.

4) By modifying the order of vertices of each triangle in the triangle sequence F, and embedding the binary sequence W4 containing watermark information, the STL 3D model embedded with redundant coded watermark sequences is obtained. For example, the first bit of the binary sequence W ₄ is "0", then adjust the order of the first triangle patch of the triangle patch sequence F to "A, B, C", so that the STL 3D with watermark information W can be obtained Model.

The second part: digital watermark extraction part

The watermark extraction part specifically includes the following three main steps:

Step 1. Sorting the triangular patches of the STL 3D model

As mentioned in the "Part 1: Digital Watermark Embedding Part" of the Summary of the Invention, there are many sorting methods for triangular patches that can be selected here, and here only the method based on PCA is taken as an example.

1) PCA (Principal Component Analysis, abbreviation for principal component analysis) preprocessing is performed on the triangular surface of the STL three-dimensional model.

2) Calculate the coordinate value of each triangle in the PCA coordinate system, sort the corresponding triangles from large to small according to their values, and obtain the triangle sequence F, including the following 4 sub-steps:

2.1) In the Cartesian coordinate system, the model is translated so that the center of gravity of the model coincides with the origin of the coordinate system.

2.2) Perform PCA processing on the coordinate values of the triangular surfaces in the model to obtain three eigenvectors.

2.3) Using the three eigenvectors as the coordinate axes of the PCA space, map the model from the Cartesian coordinate system to the PCA coordinate system.

2.4) Obtain a known triangular patch sequence F by comparing the center coordinates of the triangular faces in the model. The comparison method is to first compare the center coordinates of the triangle patches. If the center coordinates are the same, compare the x coordinates. If the x coordinates are the same, compare the y coordinates. If the y coordinates are the same, compare the z coordinates.

Step 2. Digital watermark sequence extraction

Extract watermark information from STL 3D models that may contain watermark information.

1) Compare the coordinate values of the vertices of the triangular surface, and determine the A, B, and C vertices of the triangular surface. The vertex with the largest vertex coordinate value is named as the A vertex. According to the counterclockwise direction, the other two vertices are named as the B vertex and C vertex, as shown in Figure 3.

2) Extract the binary sequence Q ₁ containing watermark information according to the vertex sorting. For example, in the triangle patch sequence F, the order of vertices of the first triangle patch is "A, B, C", then the embedded binary watermark information is "0". The binary sequence information Q ₁ is coded as a group of adjacent two bits. Among them, "00" is coded as "1", "01" is coded as "2", "10" is coded as "3", and "11" is coded as "4", and the watermarked information sequence Q ₂ is obtained.

Step 3. Digital watermark sequence verification and correction

1) Check the information sequence Q ₂ from the two aspects of the number itself and the number of check digits, locate the changed place and correct it, the steps are as follows:

2.1) The number itself: the last number of the upper unit of the adjacent unit must be the same as the first number of the lower unit (consistency includes the two cases where the adjacent numbers are the same number and paired numbers), and the numbers in the unit cannot be repeated (except for the numbers on the adjacent unit). When the last number of the unit and the first number of the lower unit are paired numbers, the last two digits of the upper unit are repeated), and when the last number of the upper unit of the adjacent unit and the first number of the lower unit are paired numbers, the last two digits of the upper unit must be repeat.

2.2) The number of check digits: check whether the number of check digits (the number of check digits in this unit always refers to the number of check digits connected to the unit) is consistent with the value of the digital elements of a single unit. When the number of unit digital elements is 1, there is only one check digit; when the number of unit digital elements is 2 or 3, the number of check digits matches the end number; when the number of unit digital elements is 3 or 4, The number of check digits is consistent with the sum of the second and third digits; when the number of unit digital elements is 4 or 5 and the second and third digits are arranged in sequence, there are two check digits, and when the second When the first digit and the third digit are arranged in reverse order, there is only one check digit.

2) Judging and correcting the non-conforming numbers according to the two standards, and obtaining the corrected watermarked information sequence Q ₃ .

In order to verify the effectiveness of the present invention, a random triangle patch reordering attack and a random triangle patch vertex reordering attack are carried out on the embedded watermarked STL 3D model, and the watermark sequence is extracted from the attacked STL 3D model. Experimental results show that this method can completely and correctly detect embedded digital watermark sequences from embedded watermarked STL 3D models subjected to random triangle patch reordering attack or vertex reordering attack within any triangle patch.

The embodiments of the present invention have been described in detail above, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge of those skilled in the art without departing from the gist of the present invention.

Claims (2)

1. A digital watermarking system for stereolithography three-dimensional model files, comprising two parts:

a first part: digital watermark embedding section

The watermark embedding part specifically comprises the following 3 main steps:

step one, redundant coding of watermark sequence

1) A watermark sequence of length L representing copyright information is denoted as W = { W₁,w₂,w₃,...,w_i,...,w_LIn which W is_i∈[0,1]If L is an odd number, then at w_LThen, one bit is supplemented to be 0;

two adjacent bits are a group, the watermark sequence W is coded, wherein 00 is coded into 1, 01 is coded into 2, 10 is coded into 3, 11 is coded into 4, and the coded watermark sequence is obtained

Wherein W_i∈[0,4]；

2) Check bit is added for the first time: in the watermark sequence W₁In the method, check bits are added between adjacent repeated or adjacent repeated numbers, and under the condition of adjacent repetition, a check bit is added between two repeated numbers, and then a second repeated number is changed into a pairing number; under the condition of adjacent repetition, firstly adding a check bit, and then repeating the last digit of the last unit to obtain a watermark sequence W checked once₂；

3) Adding check bits for the second time: after the check bit is added for the first time, the complete and continuous watermark sequence W₁The digital elements are separated by the check bits, the number of the digital elements in the middle of the two check bits is called as a unit, and the number of the digital elements in one unit is not more than 4;

3.1 in the watermark sequence W₂Different numbers of check bits are respectively added behind units with the number of digital elements being 2, the number of digital elements being 3 and the number of digital elements being 4;

3.2 if and only if the digits before and after the check digit are the paired digits, firstly, repeatedly inserting the ending digit of the unit before the original check digit, then adding a check digit after the newly inserted digit to obtain a watermark sequence W verified for the second time₃At the moment, one or more check bits exist between the units, the units are still only digital element sets between the check bits, the number of one unit digital element does not exceed 5, and the check bits are not listed in the unit range under any condition;

4) Code binarization: watermark sequence W to be checked twice₃The digital elements are reduced into two adjacent binary coding sequences in a group to obtain a binary sequence W containing watermark information₄；

Step two, sorting triangular patches of the STL three-dimensional model;

step three: embedding a redundant coding watermark sequence, namely realizing the embedding of the watermark sequence by adopting a mode of reordering the vertexes of a triangular patch of an STL three-dimensional model, and specifically comprising the following steps of:

1) Comparing the vertex coordinate values of the triangular patches, determining the vertexes A, B and C of the triangular patches, wherein the vertex with the largest vertex coordinate value is named as the vertex A, and because the vertexes of the STL three-dimensional model can only be sorted anticlockwise along the external normal vector, only three vertex sorting modes are provided after the vertexes A, B and C of the triangular patches are determined, and the other two vertexes are named as the vertex B and the vertex C in sequence according to the anticlockwise direction;

2) For each sort mode coding, the sort mode coding rule is as follows: the vertex sequence of the triangular patch is A, B and C, the code is 0, the vertex sequence of the triangular patch is B, C and A, the code is 1, the vertex sequence of the triangular patch is C, A and B, and the code is check bits;

3) Embedding a binary sequence W4 containing watermark information by modifying the vertex sequence of each triangular patch in the triangular patch sequence F to obtain an STL three-dimensional model embedded with a redundant coding watermark sequence;

a second part: digital watermark extraction section

The watermark extraction part specifically comprises the following 3 main steps:

step one, sorting triangular patches of an STL three-dimensional model;

step two, extracting the digital watermark sequence, which comprises the following specific steps:

1) Comparing the coordinate values of the vertexes of the triangular surface patch, wherein the comparison method comprises the steps of comparing x coordinates, comparing y coordinates when the x coordinates are the same, comparing z coordinates when the y coordinates are the same, naming the vertex with the largest vertex coordinate value as an A vertex, and sequentially naming the other two vertexes as a B vertex and a C vertex according to the anticlockwise direction to determine the vertices A, B and C of the triangular surface patch;

2) Extracting binary sequence Q containing watermark information according to vertex ordering₁A binary sequence information Q₁Every two adjacent bits are coded into a group, wherein the '00' code is '1', the '01' code is '2', the '10' code is '3', the '11' code is '4', and a water-containing printing information sequence Q is obtained₂；

Step three, checking and correcting digital watermark sequence

1) For information sequence Q from both the number itself and the number of check bits₂Checking, locating the changed place and correcting the changed place, and the steps are as follows:

number itself aspect: the head and tail connected numbers of the upper unit and the lower unit of the adjacent units are consistent, the numbers in the units cannot be repeated, and the two head and tail connected numbers of the two units are matched numbers, so that the two digits behind the front unit are repeated;

check bit quantity aspect: detecting whether the number of check bits is consistent with the numerical value of the single unit digital element, and when the number of the unit digital element is 1, only one check bit is available; when the number of the unit digital elements is 2 or 3, the number of the check bits is consistent with that of the tail digit; when the number of the unit digital element is 3 or 4, the number of the check bit is consistent with the sum of the second bit number and the third bit number; when the number of the unit digital elements is 4 or 5 and the second bit and the third bit are arranged in sequence, the check bit has two bits, and when the second bit and the third bit are arranged in a reverse sequence, the check bit has only one bit;

2) Judging and correcting the non-conforming numbers according to two-aspect standards to obtain a corrected watermark-containing information sequence Q₃。

2. The digital watermarking system for stereolithography three-dimensional model files according to claim 1, wherein: the sorting of the STL three-dimensional model triangular patches in the digital watermark embedding part and the digital watermark extracting part adopts a PCA-based method, and the method comprises the following specific steps:

1) Carrying out PCA (principal component analysis) pretreatment on a triangular patch of the STL three-dimensional model;

2) Calculating the coordinate value of each triangular patch under the PCA coordinate system, and sequencing the corresponding triangular patches from large to small according to the coordinate value to obtain a triangular patch sequence F;

2.1 In a cartesian coordinate system), the model is translated such that the center of gravity of the model coincides with the origin of the coordinate system;

2.2 Carrying out PCA processing on the coordinate values of the triangular surface in the model to obtain three characteristic vectors;

2.3 Three eigenvectors are used as coordinate axes of a PCA space, and the model is mapped to a PCA coordinate system from a Cartesian coordinate system;

2.4 A known triangle patch sequence F is obtained by comparing the central coordinate values of the triangle patches in the model, wherein the comparison method comprises comparing the central coordinate values of the triangle patches, comparing the x coordinate values if the central coordinate values are the same, comparing the y coordinate values if the x coordinate values are the same, and comparing the z coordinate values if the y coordinate values are the same.

Images