A method and device for watermark extraction of a three-dimensional model

Abstract

The present application discloses a method and apparatus for watermark extraction of a three-dimensional model. The method includes: obtaining multiple images of the three-dimensional model captured from multiple perspectives; respectively for each image, screening out potential watermark embedding regions according to the gray-scale gradient change characteristics of the image in the model texture coordinate space; performing multi-scale analysis on the potential watermark embedding regions to obtain the high-frequency texture characteristics of the potential watermark embedding regions; performing frequency-domain transformation on the potential watermark embedding regions to analyze the high-frequency frequency-domain characteristics of the potential watermark embedding regions based on the frequency-domain data and the high-frequency texture characteristics; fusing the high-frequency texture characteristics and the high-frequency frequency-domain characteristics of the potential watermark embedding regions of each image; performing statistical analysis on the fused two characteristics to determine the final watermark embedding region; mapping the fused frequency-domain characteristics of the final watermark embedding region back to the model texture coordinate space to obtain the watermark extraction result.

Description

Technical Field

[0001] The present application relates to the technical field of watermark extraction, and particularly relates to a method and device for extracting watermarks from 3D models. Background Art

[0002] With the development of 3D digital technology, 3D models have been widely used in the fields of industrial design, cultural heritage protection, and virtual reality. In order to protect the intellectual property rights of 3D digital assets, digital watermark technology has gradually become an effective protection means. Digital watermarks store copyright information by embedding hidden information in the texture or geometric characteristics of 3D models without significantly affecting the appearance of the models. Subsequently, when verification is required, the watermarks in the 3D models need to be extracted and then verified.

[0003] Currently, the methods for extracting watermarks from 3D models mainly rely on two-dimensional image processing and frequency domain analysis methods. Specifically, mainly through texture mapping, the surface of the 3D model is unfolded into a two-dimensional texture image. Then, the data in the two-dimensional texture image is subjected to frequency domain conversion, and the high-frequency vectors in the two-dimensional texture image are extracted, so that the watermark can be determined through the distribution of the high-frequency vectors.

[0004] However, due to the complexity and non-linearity of texture mapping of 3D models, there are also problems such as shooting angles, texture occlusion, and mapping distortion, which are likely to cause loss of texture information. Therefore, the single mapped image cannot guarantee the accuracy of the global texture of the model. Moreover, it only simply analyzes the characteristics of the frequency domain and does not perform effective verification. Therefore, the existing methods cannot effectively guarantee the accuracy of the extracted watermarks. Summary of the Invention

[0005] Based on the above deficiencies of the prior art, the present application provides a method and device for extracting watermarks from 3D models to solve the problem that the prior art cannot guarantee the accurate extraction of watermarks from 3D models.

[0006] To achieve the above object, the present application provides the following technical solutions:

[0007] The first aspect of the present application provides a method for extracting watermarks from 3D models, including:

[0008] Obtaining multiple target images of a target 3D model taken from multiple perspectives;

[0009] For each of the target images, screening out potential watermark embedding regions of the target image according to the gray gradient change characteristics of the pixel points of the target image in the model texture coordinate space;

[0010] Perform multi-scale high-frequency texture feature analysis on the potential watermark embedding region to obtain the high-frequency texture features of the potential watermark embedding region;

[0011] Perform frequency domain transformation on the potential watermark embedding region, and analyze the high-frequency frequency domain features of the potential watermark embedding region based on the frequency domain data of the potential watermark embedding region and the high-frequency texture features;

[0012] Fuse the high-frequency texture features and the high-frequency frequency domain features of each potential watermark embedding region of each target image respectively to obtain fused texture features and fused frequency domain features;

[0013] Determine the final watermark embedding region through statistical analysis of the fused texture features and the fused frequency domain features;

[0014] Map the fused frequency domain features of the final watermark embedding region back to the model texture coordinate space to obtain the watermark extraction result of the target 3D model.

[0015] Optionally, in the above watermark extraction method for 3D models, after obtaining multiple target images of the target 3D model taken from multiple perspectives, it further includes:

[0016] Respectively use histogram equalization technology to enhance the contrast of each target image data to obtain each enhanced target image;

[0017] Perform gamma correction processing on each enhanced target image to adjust the gray dynamic range of the target image;

[0018] Correct the geometric distortion of each adjusted target image through homography transformation;

[0019] Use the adaptive median filtering algorithm to filter the noise of each corrected target image.

[0020] Optionally, in the above watermark extraction method for 3D models, before respectively screening out the potential watermark embedding regions of the target images according to the gray gradient change characteristics of the pixel points of the target images in the model texture coordinate space for each target image, it further includes:

[0021] Respectively extract the feature points in each target image for each target image;

[0022] Calculate the characteristic description information of each feature point; wherein, the characteristic description information of the feature point is the information characterizing the local texture of the feature point;

[0023] Based on the texture mapping parameters of the target 3D model and the camera calibration parameters when the target image is captured, establish a mapping model between each of the feature points and the model texture coordinate space;

[0024] Optimize the mapping model by minimizing the total error function;

[0025] Calculate the local consistency degree of each of the feature points in the optimized mapping model;

[0026] Perform optimization processing on each of the feature points whose local consistency degree does not meet the preset conditions.

[0027] Optionally, in the above watermark extraction method for the 3D model, the step of respectively screening out the potential watermark embedding regions of each target image according to the gray gradient change characteristics of the pixel points of the target image in the model texture coordinate space includes:

[0028] For each target image respectively, determine the target texture region of the target image in the model texture coordinate space;

[0029] Extract the gray gradient change characteristics of each pixel point in the target texture region of the target image;

[0030] Based on the gray gradient change characteristics of each pixel point in the target texture region of the target image, screen out the potential watermark embedding regions of the target image through the region growing criterion.

[0031] Optionally, in the above watermark extraction method for the 3D model, the step of performing multi-scale high-frequency texture characteristic analysis on the potential watermark embedding regions to obtain the high-frequency texture characteristics of the potential watermark embedding regions includes:

[0032] Perform multi-scale characteristic analysis on the potential watermark embedding regions to obtain characteristics at multiple scales;

[0033] Calculate the difference between the characteristics of every two adjacent layers of scales to obtain multi-layer initial high-frequency texture characteristics;

[0034] Combine the gray gradient change characteristics of the potential watermark embedding regions with the initial texture high-frequency information of each layer to obtain an optimized high-frequency texture characteristic enhancement map.

[0035] Optionally, in the above watermark extraction method for the 3D model, after combining the gray gradient change characteristics of the potential watermark embedding regions with the initial texture high-frequency information of each layer to obtain an optimized high-frequency texture characteristic enhancement map, it further includes:

[0036] Perform boundary extraction and confirmation on the enhanced high-frequency texture characteristic map, and perform regional consistency verification.

[0037] Optionally, in the above watermark extraction method for a three-dimensional model, analyzing the high-frequency frequency domain characteristics of the potential watermark embedding region based on the frequency domain data of the potential watermark embedding region and the high-frequency texture characteristics includes:

[0038] Extract high-frequency frequency domain data from the frequency domain data of the potential watermark embedding region;

[0039] Weight the high-frequency frequency domain data with the enhanced high-frequency texture characteristic map to obtain the high-frequency energy of the potential watermark embedding region;

[0040] Perform discrete wavelet transform on the frequency domain data of the potential watermark embedding region to decompose the frequency domain data of the potential watermark embedding region into multi-level frequency domain components; wherein, the frequency domain components include low-frequency components, horizontal high-frequency components, vertical high-frequency components, and diagonal high-frequency components;

[0041] Calculate multi-level energy ratios using the enhanced high-frequency texture characteristic map and each level of the frequency domain components;

[0042] Combine the enhanced high-frequency texture characteristic map, the high-frequency energy, and the energy ratios of each level to obtain a total characteristic component.

[0043] Optionally, in the above watermark extraction method for a three-dimensional model, fusing the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each potential watermark embedding region of each target image respectively to obtain fused texture characteristics and fused frequency domain characteristics includes:

[0044] Fuse the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each potential watermark embedding region of each target image respectively according to the weights corresponding to the target attributes of each target image to obtain fused texture characteristics and fused frequency domain characteristics; wherein, the target attributes include at least shooting angle, resolution, and confidence parameter.

[0045] Optionally, in the above watermark extraction method for a three-dimensional model, mapping the fused frequency domain characteristics of the final watermark embedding region back to the model texture coordinate space to obtain the watermark extraction result of the target three-dimensional model includes:

[0046] Map the fused frequency domain characteristics of the final watermark embedding region back to the model texture coordinate space to obtain the distribution map of the final watermark embedding region;

[0047] Obtain the position information of each feature point in the distribution map of the final watermark embedding area; wherein, the position information includes texture coordinates and physical coordinates;

[0048] Calculate the confidence level of each feature point based on the total characteristic component and the maximum characteristic component of each feature point in the distribution map of the final watermark embedding area;

[0049] Generate and output a watermark extraction report by using the position information, confidence level of each feature point and the distribution map of the final watermark embedding area.

[0050] The second aspect of the present application provides a watermark extraction device for a 3D model, including:

[0051] An image acquisition unit, configured to acquire multiple target images of the target 3D model taken from multiple perspectives;

[0052] A preliminary screening unit, configured to respectively screen out potential watermark embedding areas of each target image according to the gray gradient change characteristics of the points of the target image in the model texture coordinate space;

[0053] A texture characteristic analysis unit, configured to perform multi-scale high-frequency texture characteristic analysis on the potential watermark embedding area to obtain the high-frequency texture characteristics of the potential watermark embedding area;

[0054] A frequency domain transformation unit, configured to perform frequency domain transformation on the potential watermark embedding area;

[0055] A frequency domain characteristic analysis unit, configured to analyze the high-frequency frequency domain characteristics of the potential watermark embedding area based on the frequency domain data and the high-frequency texture characteristics of the potential watermark embedding area;

[0056] An image characteristic fusion unit, configured to fuse the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each potential watermark embedding area of each target image respectively to obtain fused texture characteristics and fused frequency domain characteristics;

[0057] A characteristic verification unit, configured to determine the final watermark embedding area by performing statistical analysis on the fused texture characteristics and the fused frequency domain characteristics;

[0058] A result generation unit, configured to map the fused frequency domain characteristics back to the model texture coordinate space to obtain the watermark extraction result of the target 3D model.

[0059] Optionally, in the above watermark extraction device for a 3D model, it further includes:

[0060] An enhancement unit for enhancing the contrast of each piece of the target image data by using the histogram equalization technique respectively to obtain each enhanced target image;

[0061] A dynamic adjustment unit for performing gamma correction processing on each enhanced target image to adjust the gray-scale dynamic range of the target image;

[0062] A correction unit for correcting the geometric distortion of each adjusted target image through homography transformation;

[0063] A filtering unit for filtering the noise of each corrected target image by using an adaptive median filtering algorithm.

[0064] Optionally, in the above watermark extraction device for a three-dimensional model, it further includes:

[0065] A feature point extraction unit for extracting feature points in each target image respectively for each piece of the target image;

[0066] A description information calculation unit for calculating the characteristic description information of each feature point; wherein, the characteristic description information of the feature point is information characterizing the local texture of the feature point;

[0067] A mapping model establishment unit for establishing a mapping model between each feature point and the model texture coordinate space based on the texture mapping parameters of the target three-dimensional model and the camera calibration parameters when shooting the target image;

[0068] An optimization unit for optimizing the mapping model by minimizing the total error function;

[0069] A consistency degree calculation unit for calculating the local consistency degree of each feature point in the optimized mapping model;

[0070] An optimization unit for performing optimization processing on each feature point whose local consistency degree does not meet the preset conditions.

[0071] Optionally, in the above watermark extraction device for a three-dimensional model, the preliminary screening unit includes:

[0072] A region determination unit for determining the target texture region of the target image in the model texture coordinate space respectively for each piece of the target image;

[0073] A gradient characteristic extraction unit for extracting the gray-scale gradient change characteristics of each pixel point in the target texture region of the target image;

[0074] A feature screening unit, configured to screen out potential watermark embedding regions of the target image based on the gray gradient change features of each pixel point within the target texture region of the target image through a region growing criterion.

[0075] Optionally, in the above watermark extraction device for a 3D model, the texture feature analysis unit includes:

[0076] A multi-scale analysis unit, configured to perform multi-scale feature analysis on the potential watermark embedding regions to obtain features at multiple scales;

[0077] A texture feature calculation unit, configured to calculate the difference between the features at adjacent two scales to obtain multi-layer initial high-frequency texture features;

[0078] A texture feature combination unit, configured to combine the gray gradient change feature of the potential watermark embedding region with the initial texture high-frequency information at each layer to obtain an optimized high-frequency texture feature enhancement map.

[0079] Optionally, in the above watermark extraction device for a 3D model, it further includes:

[0080] A region verification unit, configured to perform boundary extraction and confirmation on the high-frequency texture feature enhancement map and perform region consistency verification.

[0081] Optionally, in the above watermark extraction device for a 3D model, the frequency domain feature analysis unit includes:

[0082] A frequency domain data extraction unit, configured to extract high-frequency frequency domain data from the frequency domain data of the potential watermark embedding region;

[0083] A frequency domain energy calculation unit, configured to weight the high-frequency frequency domain data by using the high-frequency texture feature enhancement map to obtain the high-frequency energy of the potential watermark embedding region;

[0084] A decomposition unit, configured to perform discrete wavelet transform on the frequency domain data of the potential watermark embedding region to decompose the frequency domain data of the potential watermark embedding region into multi-level frequency domain components; wherein, the frequency domain components include low-frequency components, horizontal high-frequency components, vertical high-frequency components, and diagonal high-frequency components;

[0085] An energy ratio calculation unit, configured to calculate multi-level energy ratios by using the high-frequency texture feature enhancement map and the frequency domain components at each level;

[0086] A total feature component calculation unit, configured to combine the high-frequency texture feature enhancement map, the high-frequency energy, and the energy ratios at each level to obtain a total feature component.

[0087] Optionally, in the above watermark extraction device for a three-dimensional model, the image feature fusion unit includes:

[0088] An image feature fusion subunit, configured to fuse the high-frequency texture features and the high-frequency frequency domain features of each potential watermark embedding region of each target image according to the weights corresponding to the target attributes of each target image, to obtain fused texture features and fused frequency domain features; wherein, the target attributes at least include shooting angle, resolution, and confidence parameter.

[0089] Optionally, in the above watermark extraction device for a three-dimensional model, the result generation unit includes:

[0090] An inverse mapping unit, configured to map the fused frequency domain features of the final watermark embedding region back to the model texture coordinate space to obtain a distribution map of the final watermark embedding region;

[0091] A coordinate acquisition unit, configured to acquire the position information of each feature point in the distribution map of the final watermark embedding region; wherein, the position information includes texture coordinates and physical coordinates;

[0092] A confidence calculation unit, configured to calculate the confidence of each feature point according to the total feature component and the maximum feature component of each feature point in the distribution map of the final watermark embedding region;

[0093] A report generation unit, configured to generate and output a watermark extraction report by using the position information, confidence, and the distribution map of the final watermark embedding region of each feature point.

[0094] A method for extracting watermarks from a 3D model provided by the present application obtains multiple target images of the target 3D model taken from multiple perspectives. Thus, through the images from multiple perspectives, the texture of the 3D model can be accurately reflected, and more comprehensive and accurate features can be provided. Moreover, the features of each image can be verified to ensure the accuracy of the features, thereby improving the accuracy of watermark extraction. Then, for each target image respectively, according to the gray gradient change characteristics of the pixel points of the target image in the model texture coordinate space, the potential watermark embedding regions of the target image are screened out, and multi-scale high-frequency texture feature analysis is performed on the potential watermark embedding regions to obtain the high-frequency texture features of the potential watermark embedding regions. Thus, through multi-scale analysis, accurate texture features are extracted. Then, frequency domain transformation is performed on the potential watermark embedding regions, and based on the frequency domain data and high-frequency texture features of the potential watermark embedding regions, the high-frequency frequency domain features of the potential watermark embedding regions are analyzed. Thus, not only the texture feature information and the frequency domain feature information are combined and analyzed, but also the watermark can be more accurately located. The high-frequency texture features and high-frequency frequency domain features of each potential watermark embedding region of each target image are respectively fused to obtain fused texture features and fused frequency domain features, so as to comprehensively consider the features of each image. Then, through statistical analysis of the fused texture features and fused frequency domain features, the final watermark embedding region is determined, so that the accuracy of the analyzed features can be verified, and an accurate final watermark embedding region can be obtained. Finally, the fused frequency domain features are mapped back to the model texture coordinate space to obtain the watermark extraction result of the target 3D model, thus realizing a method that can accurately extract watermarks from 3D models. Description of the Drawings

[0095] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0096] Figure 1 It is a flowchart of a method for extracting watermarks from a 3D model provided by an embodiment of the present application;

[0097] Figure 2 It is a flowchart of a preprocessing method for a target image provided by an embodiment of the present application;

[0098] Figure 3 It is a flowchart of a method for mapping a target image to a model texture coordinate space provided by an embodiment of the present application;

[0099] Figure 4 It is a flowchart of a method for screening out potential watermark embedding regions of a target image provided by an embodiment of the present application;

[0100] Figure 5 It is a flowchart of a method for analyzing texture characteristics provided by an embodiment of the present application;

[0101] Figure 6 It is a flowchart of a method for analyzing frequency domain characteristics provided by an embodiment of the present application;

[0102] Figure 7 It is a flowchart of a method for generating a watermark extraction result of a target three-dimensional model provided by an embodiment of the present application;

[0103] Figure 8 It is a schematic diagram of the architecture of a watermark extraction device for a three-dimensional model provided by an embodiment of the present application. Detailed implementation manners

[0104] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0105] In the present application, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover a non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device. Without further limitations, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article or device including the element.

[0106] An embodiment of the present application provides a watermark extraction method for a three-dimensional model, as Figure 1 shown, including the following steps:

[0107] S101. Obtain multiple target images of the target three-dimensional model taken from multiple perspectives.

[0108] Among them, the target three-dimensional model refers to the three-dimensional model for which watermark extraction needs to be performed currently.

[0109] In order to accurately restore the texture of the 3D model and obtain an accurate special shape, in the embodiments of the present application, the target 3D model will be photographed from multiple perspectives to obtain images of the target 3D model from multiple perspectives. Thus, through the images of the target 3D model from multiple perspectives, the full amount of texture can be obtained, and then the texture of the 3D model can be accurately restored, improving the accuracy of watermark extraction. Moreover, the texture characteristics of each image can be verified and supplemented with each other, thereby improving the accuracy of the finally extracted watermark.

[0110] Optionally, the shooting angle communication satisfies the coverage of the area where the watermark is embedded. Specifically, the area can be photographed at 0 to 180 degrees, so as to avoid the loss of key texture details due to occlusion or projection distortion.

[0111] Optionally, in order to improve the quality of the target image and thus improve the accuracy of watermark extraction, in another embodiment of the present application, after performing step S101, the target image can also be further preprocessed first. As Figure 2 shown, a method for preprocessing a target image provided by an embodiment of the present application includes:

[0112] S201. Respectively use histogram equalization technology to enhance the contrast of each piece of target image data to obtain each enhanced target image.

[0113] Specifically, when using histogram equalization technology to enhance the contrast of the target image data, the pixel value of the enhanced image data can be obtained as:

[0114]

[0115] where I(x,y) is the gray value of the original target image; I min and I max are the minimum gray value and the maximum gray value on the target image.

[0116] S202. Perform gamma correction processing on each enhanced target image to adjust the gray dynamic range of the target image.

[0117] Specifically, through the correction coefficient, the data of the target image is adjusted, and the pixel value of the adjusted image data is obtained as:

[0118]

[0119] where is the correction coefficient, which can be adjusted according to the characteristics of the image acquisition device. When it is used to enhance the dark area of the image, and when

[0120] S203. Correct the geometric distortion of each adjusted target image through a homography transformation.

[0121] Changes in the shooting angle may cause geometric distortion in the image. Especially when the texture and shape of the target object are projected onto the image plane at different perspectives, straight lines or planes in the image may appear curved or irregular. Therefore, through geometric correction of the image based on the homography transformation, these distorted images are adjusted back to a unified perspective, so that the texture mapping of the image is consistent with that of the 3D model.

[0122] S204. Use an adaptive median filtering algorithm to filter out noise from each corrected target image.

[0123] Specifically, the pixel values of the target image can be filtered according to a filtering window of a certain size. If the pixel value is within the gray level range of the filtering window, it remains unchanged; if it exceeds the gray level range, the median of the pixel values within the window can be used to replace it.

[0124] It should also be noted that in order to accurately analyze the texture characteristics of each target image and facilitate the subsequent fusion of the characteristics of each target image, after performing step S101 and before performing step S102, each target image needs to be mapped to the model texture coordinate space first.

[0125] Specifically, a mapping relationship between the captured target image and the 3D model texture coordinates is established according to the texture mapping parameters when the target 3D model is generated. Specifically, a projection transformation matrix is used to uniformly map each target image to a model texture coordinate space.

[0126] Optionally, in another embodiment of the present application, a specific implementation manner of mapping each target image to the model texture coordinate space is as follows Figure 3 shown, including the following steps:

[0127] S301. For each target image, extract the feature points in the target image.

[0128] Optionally, multiple feature points p i =[x i , y i , 1] T can be extracted from the target image for subsequent processing; where x i , y i are the abscissa value and ordinate value of the feature point. Specifically, a feature point detection algorithm can be used to identify the key feature points in the texture area of the target image.

[0129] S302. Calculate the characteristic description information of each feature point.

[0130] Among them, the characteristic description information of the feature point is the information characterizing the local texture of the feature point, which can be composed of information in multiple dimensions.

[0131] S303. Based on the texture mapping parameters of the target three-dimensional model and the camera calibration parameters when shooting the target image, establish a mapping model between each feature point and the model texture coordinate space.

[0132] Specifically, based on the texture mapping parameters of the target three-dimensional model and the camera calibration parameters when shooting the target image, the projection transformation matrix can be obtained. Then, through the projection transformation matrix, the data of the target image can be mapped to the model texture coordinate space, that is, a mapping model between each feature point and the model texture coordinate space is established. Therefore, for any point, the mapping relationship can be expressed as:

[0133]

[0134] Among them, H enh is the projection transformation matrix, and D(p i ) is the coordinate of the feature point.

[0135] Optionally, the coordinates of the feature point after distortion correction can be specifically used for mapping, that is, D(p i ) is the coordinate of the feature point after distortion correction. Therefore, if the distortion coefficients are k 1 , k 2 , p 1 and p 2 , the coordinates of the feature point after distortion correction can be expressed as:

[0136]

[0137] Among them, x i and y i are the coordinates of the feature point i before correction; x c and y c are the coordinates of the camera principal point; r = (x i - x c ) 2 + (y i - y c ) 2 .

[0138] S304. Optimize the mapping model by minimizing the total error function.

[0139] In order to ensure the quality after mapping, in the embodiments of the present application, further optimization is required. First, the mapping relationship can be optimized by minimizing the following total error function E:

[0140]

[0141] Wherein, N is the number of feature points; α is the weight coefficient for balancing the reprojection error and geometric consistency, and "G(.)" is the geometric feature function. The geometric feature function extracts the geometric attributes such as curvature or edge direction at the corresponding texture coordinates. The optimization process is performed for the parameters of the enhanced projection transformation matrix H enh .

[0142] S305. Calculate the local consistency degree of each feature point in the optimized mapping model.

[0143] In order to verify using the geometric topological consistency of the texture of the target three-dimensional model, the local consistency of each feature point is calculated:

[0144]

[0145] Wherein, N(i) is the neighborhood set of the feature point p i , w ij is the weight, and u i and u j represent the coordinates after the i-th and j-th feature points are mapped into the model texture coordinate space.

[0146] S306. Perform optimization processing on each feature point whose local consistency degree does not meet the preset conditions.

[0147] Specifically, it can be to extract each feature point whose local consistency degree does not meet the preset conditions, so as to avoid the interference of these abnormal points on the mapping accuracy noise, and thus obtain an accurate mapping result, accurately reflecting the texture of the target three-dimensional model.

[0148] S102. For each target image, respectively, according to the gray gradient change characteristics of the pixel points of the target image in the model texture coordinate space, screen out the potential watermark embedding areas of the target image.

[0149] Since there is a position where the embedded watermark exists, the degree of change of the gray value of the image will be relatively large. Therefore, in the embodiments of the present application, the gray gradient change characteristics are calculated by using the gray values of the pixel points of the target image mapped into the model texture coordinate space, and then based on the gray gradient change characteristics, through an image segmentation algorithm, the areas where watermark embedding may exist in the target image can be screened out.

[0150] Optionally, in another embodiment of the present application, a specific implementation manner of step S102 is as Figure 4 shown, and includes the following steps:

[0151] S401. For each target image, determine the target texture region of the target image in the model texture coordinate space.

[0152] In order to speed up the processing speed, in the embodiments of the present application, the region most likely to have a watermark is first determined as the target texture region, or if it is the region where the consistent watermark is located, this region can be directly determined as the target texture region.

[0153] S402. Extract the gray gradient change characteristics of each pixel point in the target texture region of the target image.

[0154] Specifically, calculate the gray gradient change based on the gray values of each point on the target image to obtain the gradient vector , and then approximately calculate the gradient assignment through the Sobel operator , so as to obtain the gray gradient change characteristics of each point.

[0155] S403. Based on the gray gradient change characteristics of each pixel point in the target texture region of the target image, screen out the potential watermark embedding region of the target image through the region growing criterion.

[0156] Specifically, screen out the pixel points that meet the region growing criterion, and the region where these points are located is the potential watermark embedding region. Among them, the region growing criterion is defined as:

[0157]

[0158] Among them, I seed is the gray value of a selected seed point; I(u i ) is the gray value of the i-th pixel point; T and are the gray difference and gradient assignment threshold respectively.

[0159] S103. Perform multi-scale high-frequency texture characteristic analysis on the potential watermark embedding region to obtain the high-frequency texture characteristics of the potential watermark embedding region.

[0160] It should be noted that since the watermark region is mainly manifested in high-frequency features, and in order to obtain accurate texture characteristics and improve the accuracy of the final watermark extraction, in the implementation of the present application, multi-scale high-frequency texture characteristics of the potential watermark embedding region are analyzed, so as to separate the background changes and embedded information in the texture of the target image, obtain the high-frequency texture characteristics of the potential watermark embedding region, and the region where these characteristics are located is a more accurate watermark embedding region, that is, through texture characteristic analysis, the potential watermark embedding region can be limited to a more accurate range.

[0161] Optionally, in another embodiment of the present application, a specific implementation of step S103 is as follows Figure 5 shown, including the following steps:

[0162] S501. Perform multi-scale characteristic analysis on the potential watermark embedding region to obtain characteristics at multiple scales.

[0163] Optionally, performing multi-scale characteristic analysis on the potential watermark embedding region can be specifically represented by a multi-scale Gaussian pyramid:

[0164]

[0165] where "*" represents a convolution operation, and is the scale parameter of the k-th layer of the pyramid.

[0166] S502. Calculate the difference between the characteristics of adjacent two layers of scales to obtain multi-layer initial high-frequency texture characteristics.

[0167] Specifically, by calculating the high-frequency features H k (u i ), the high-frequency texture characteristics related to watermark embedding are initially extracted:

[0168]

[0169] S503. Combine the gray-scale gradient change characteristics of the potential watermark embedding region with the initial texture high-frequency information of each layer to obtain an optimized high-frequency texture characteristic enhancement map.

[0170] Specifically, after initially obtaining the high-frequency texture characteristics, the gray-scale gradient change and the high-frequency texture characteristics are also combined to further optimize and screen the potential watermark embedding region, making it further restricted, which can be specifically expressed as:

[0171]

[0172] where w k is the weight of the high-frequency texture characteristics at different scales.

[0173] It should be noted that the optimization process is only to further remove regions with low confidence, that is, to restrict the potential watermark embedding region to make it the final watermark embedding region.

[0174] Optionally, in another embodiment of the present application, after performing step S503, it can further include:

[0175] Perform boundary extraction and confirmation on the high-frequency texture characteristic enhancement map, and perform regional consistency verification.

[0176] Specifically, the boundary of the high-frequency texture feature enhancement map is extracted and confirmed, and the boundary points are detected, and then the regional consistency verification is performed.

[0177] S104: Perform frequency domain transformation on the potential watermark embedding area.

[0178] In order to perform watermark analysis from the frequency, in the embodiment of the present application, the frequency domain transformation is performed on the potential watermark embedding area, so that the frequency domain data of the potential watermark embedding area can be obtained.

[0179] Optionally, the grayscale signal of the potential watermark embedding area in the texture coordinate space may be converted from the space domain to the frequency domain by discrete preselection transformation, thereby obtaining the frequency domain data of the potential watermark embedding area.

[0180] Specifically, for each grayscale signal in the potential watermark embedding area in the texture coordinate space, its two-dimensional frequency domain transform, namely, DCT transform is:

[0181]

[0182] Among them, C (u, v) is the frequency coefficient matrix, that is, the frequency domain data after transformation; (u, v) is the frequency domain coordinate; and is the scaling factor; M and N represent the number of rows and columns of pixels in the region, respectively.

[0183] S105: Analyze the high-frequency frequency domain characteristics of the potential watermark embedding area based on the frequency domain data and high-frequency texture characteristics of the potential watermark embedding area.

[0184] It should be noted that the area where the watermark is located also shows high frequency in the frequency domain, so it is necessary to classify the high-frequency frequency domain characteristics. Since the high-frequency texture characteristics reflect its high-frequency characteristics in the texture, it can be used as a reference to analyze the high-frequency frequency domain characteristics of the potential watermark embedding area. In addition, the high-frequency texture characteristics can be combined to make the analyzed high-frequency frequency domain characteristics more accurate.

[0185] Optionally, in another embodiment of the present application, a specific implementation of step S105 is as follows: Figure 6 As shown, the following steps are included:

[0186] S601. Extract high-frequency frequency domain data from the frequency domain data in the potential watermark embedding area.

[0187] Specifically, the high-frequency components of the frequency domain data can be analyzed to filter out the high-frequency frequency domain data, thereby obtaining the area where the high-frequency frequency domain data is located, and then the specific location where the watermark is embedded can be further determined.

[0188] S602. Use the high-frequency texture feature enhancement map to weight the frequency-domain data of the high-frequency part of the graph to obtain the high-frequency energy of the potential watermark embedding area.

[0189] Specifically, in the embodiment of the present application, the energy in the frequency domain is used as the feature in the frequency domain to reflect the embedding position of the watermark through the distribution of the energy. The high-frequency texture feature enhancement map also reflects the position of the watermark embedding, so it is necessary to consider comprehensively. Therefore, the high-frequency energy of the potential watermark embedding area is calculated by combining the high-frequency texture feature enhancement map with the frequency-domain data of the high-frequency part.

[0190] Specifically, define the high-frequency ability as:

[0191]

[0192] where, and u t and v t are frequency-domain thresholds, which are determined by weighting the high-frequency ability in combination with the high-frequency texture feature enhancement map E(u i ).

[0193] S603. Perform discrete wavelet transform on the frequency-domain data of the potential watermark embedding area to decompose the frequency-domain data of the potential watermark embedding area into multi-level frequency-domain components.

[0194] Among them, the frequency-domain components include low-frequency components, horizontal high-frequency components, vertical high-frequency components, and diagonal high-frequency components.

[0195] Specifically, decompose the frequency-domain data of the watermark embedding area into low-frequency component LL k , horizontal high-frequency component LH k , vertical high-frequency component HL k and diagonal high-frequency component HH k , which can be specifically expressed as:

[0196]

[0197] S604. Use the high-frequency texture feature enhancement map and each level of frequency-domain components to calculate multi-level energy ratios.

[0198] In order to represent the distribution of the embedded signal in the high-frequency components, based on the high-frequency texture feature enhancement map, multi-level energy ratios are calculated, specifically:

[0199]

[0200] S605. Combine the high-frequency texture feature enhancement map, the high-frequency energy, and the energy ratios of each level to obtain the total characteristic component.

[0201] In order to comprehensively consider the characteristics of each part analyzed and finally extract the watermark, it is necessary to combine the enhanced high-frequency texture characteristics map, high-frequency energy, and frequency-domain components at each level analyzed, all of which are combined to obtain a total characteristic component. Specifically, it can be expressed as:

[0202]

[0203] Among them, 、 and are weight coefficients, which can be optimized and adjusted based on the gray scale, frequency domain, and high-frequency gradient characteristics of the embedded signal.

[0204] S106. Respectively fuse the high-frequency texture characteristics and high-frequency frequency-domain characteristics of each potential watermark embedding area of each target image to obtain the fused texture characteristics and fused frequency-domain characteristics.

[0205] After the above analysis, the characteristics of each potential watermark embedding area in each target image are obtained. Therefore, in order to comprehensively consider the characteristics of each potential watermark embedding area of each target image and accurately analyze the position of the final watermark, the high-frequency texture characteristics and high-frequency frequency-domain characteristics of each potential watermark embedding area of each target image are respectively fused to obtain the fused texture characteristics and fused frequency-domain characteristics.

[0206] Optionally, for the features belonging to the same area, they can be weighted according to their confidence levels, and the features of different areas can be combined to obtain the fused texture characteristics and fused frequency-domain characteristics, that is, the distribution maps of the fused texture characteristics and the fused frequency-domain characteristics are obtained.

[0207] Optionally, in another embodiment of the present application, the specific implementation manner of step S106 includes:

[0208] Respectively fuse the high-frequency texture characteristics and high-frequency frequency-domain characteristics of each potential watermark embedding area of each target image according to the weights corresponding to the target attributes of each target image to obtain the fused texture characteristics and fused frequency-domain characteristics.

[0209] Among them, the target attributes at least include the shooting angle, resolution, and confidence parameter.

[0210] S107. Determine the final watermark embedding area by statistically analyzing the fused texture characteristics and fused frequency-domain characteristics.

[0211] It should be noted that both the fused texture feature and the fused frequency domain feature can reflect the location of the watermark. Therefore, if both the fused texture feature and the fused frequency domain feature indicate a certain location as the watermark location, that is, the distributions of the fused texture feature and the fused frequency domain feature at a certain location are consistent, it means that it is the watermark location. Therefore, by using the fused texture feature, verify whether the distribution area of the fused frequency domain feature accurately reflects the watermark embedding area. Specifically, the fused texture feature and the fused frequency domain feature can be statistically analyzed, and then verified according to the statistical results, so as to determine the final watermark embedding area, that is, the area where the analyzed features are verified is the watermark area, and thus the next step can be carried out.

[0212] Specifically, the frequency domain energy and texture change characteristics of the fused high-frequency texture feature enhanced image and the fused total feature component watermark embedding area can be used to verify whether they are consistent. The area where the two are consistent is the watermark embedding area.

[0213] S108. Map the fused frequency domain feature back to the model texture coordinate space to obtain the watermark extraction result of the target 3D model.

[0214] The fused frequency domain feature combines the texture feature and the frequency domain feature, so it can accurately reflect the watermark location, and through the previous step, it is verified that its distribution location is the watermark location. However, the fused frequency domain feature is a frequency domain feature and cannot intuitively reflect the watermark location. Therefore, it needs to be mapped back to the model texture coordinate space to obtain the distribution map of the characteristics of the final watermark embedding area, and the distribution map of the characteristics obviously presents the watermark in the target 3D model. Optionally, different colors can be used for rendering according to the different eigenvalue sizes, so as to more intuitively present the watermark in the target 3D model, so as to verify the watermark. Therefore, based on the distribution map of the final watermark embedding area, the watermark extraction result of the target 3D model containing detailed information can be generated and output.

[0215] Optionally, in another embodiment of the present application, a specific implementation manner of step S108 is as Figure 7 shown, including:

[0216] S701. Map the fused frequency domain feature of the final watermark embedding area back to the model texture coordinate space to obtain the distribution map of the final watermark embedding area.

[0217] S702. Obtain the position information of each feature point in the distribution map of the final watermark embedding area.

[0218] Among them, the position information includes texture coordinates and physical coordinates.

[0219] S703. Calculate the confidence level of each feature point based on the total feature component and the maximum feature component of each feature point in the distribution map of the final watermark embedding area.

[0220] Among them, the maximum feature component refers to the maximum value in the total feature components of each feature point. Therefore, the confidence level of each feature point can be expressed as:

[0221]

[0222] Among them, T c (u i ) is the total feature component of the feature point. max(T c ) is the maximum feature component.

[0223] S704. Generate and output a watermark extraction report using the position information, confidence level of each feature point, and the distribution map of the final watermark embedding area.

[0224] The embodiment of the present application provides a method for extracting watermarks from a 3D model. Multiple target images of the target 3D model taken from multiple perspectives are obtained, so that the texture of the 3D model can be accurately reflected through the images from multiple perspectives, and more comprehensive and accurate features can be provided. And the features of each image can be verified to ensure the accuracy of the features, thereby improving the accuracy of watermark extraction. Then, for each target image respectively, according to the gray gradient change characteristics of the points of the target image in the model texture coordinate space, the potential watermark embedding areas of the target image are screened out, and multi-scale high-frequency texture feature analysis is performed on the potential watermark embedding areas to obtain the high-frequency texture features of the potential watermark embedding areas, so that accurate texture features are extracted through multi-scale analysis. Then, frequency domain transformation is performed on the potential watermark embedding areas, and based on the frequency domain data and high-frequency texture features of the potential watermark embedding areas, the high-frequency frequency domain features of the potential watermark embedding areas are analyzed, so that not only the texture features and frequency domain features are combined for analysis, but also the watermark can be more accurately located. The high-frequency texture features and high-frequency frequency domain features of each potential watermark embedding area of each target image are respectively fused to obtain fused texture features and fused frequency domain features, so as to comprehensively consider the features of each image. Then, through statistical analysis of the fused texture features and fused frequency domain features, the final watermark embedding area is determined, so that the accuracy of the analyzed features can be verified, and an accurate final watermark embedding area can be obtained. Finally, the fused frequency domain features are mapped back to the model texture coordinate space to obtain the watermark extraction result of the target 3D model, thus realizing a method that can accurately extract watermarks from a 3D model.

[0225] Another embodiment of the present application provides a watermark extraction device for a 3D model, as Figure 8 shown, including:

[0226] An image acquisition unit 801, configured to obtain multiple target images of a target three-dimensional model captured from multiple perspectives.

[0227] A preliminary screening unit 802, configured to respectively screen out potential watermark embedding regions of each target image according to the gray gradient change characteristics of the points of the target image in the model texture coordinate space.

[0228] A texture characteristic analysis unit 803, configured to perform multi-scale high-frequency texture characteristic analysis on the potential watermark embedding regions to obtain the high-frequency texture characteristics of the potential watermark embedding regions.

[0229] A frequency domain transformation unit 804, configured to perform frequency domain transformation on the potential watermark embedding regions.

[0230] A frequency domain characteristic analysis unit 805, configured to analyze the high-frequency frequency domain characteristics of the potential watermark embedding regions based on the frequency domain data and high-frequency texture characteristics of the potential watermark embedding regions.

[0231] An image characteristic fusion unit 806, configured to respectively fuse the high-frequency texture characteristics and high-frequency frequency domain characteristics of each potential watermark embedding region of each target image to obtain fused texture characteristics and fused frequency domain characteristics.

[0232] A characteristic verification unit 807, configured to determine the final watermark embedding region by performing statistical analysis on the fused texture characteristics and fused frequency domain characteristics.

[0233] A result generation unit 808, configured to map the fused frequency domain characteristics back to the model texture coordinate space to obtain the watermark extraction result of the target three-dimensional model.

[0234] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, it further includes:

[0235] An enhancement unit, configured to respectively enhance the contrast of each target image data by using histogram equalization technology to obtain each enhanced target image.

[0236] A dynamic adjustment unit, configured to perform gamma correction processing on each enhanced target image to adjust the gray dynamic range of the target image.

[0237] A correction unit, configured to correct the geometric distortion of each adjusted target image through homography transformation.

[0238] A filtering unit, configured to filter out noise from each corrected target image by using an adaptive median filtering algorithm.

[0239] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, it further includes:

[0240] A feature point extraction unit, which is used to extract feature points in each target image respectively.

[0241] A description information calculation unit, which is used to calculate the characteristic description information of each feature point. Wherein, the characteristic description information of the feature point is the information characterizing the local texture of the feature point.

[0242] A mapping model establishment unit, which is used to establish a mapping model between each feature point and the model texture coordinate space based on the texture mapping parameters of the target 3D model and the camera calibration parameters when shooting the target image.

[0243] An optimization unit, which is used to optimize the mapping model by minimizing the total error function.

[0244] A consistency degree calculation unit, which is used to calculate the local consistency degree of each feature point in the optimized mapping model.

[0245] An optimization unit, which is used to perform optimization processing on each feature point whose local consistency degree does not meet the preset conditions.

[0246] Optionally, in the watermark extraction device for a 3D model provided in another embodiment of the present application, the preliminary screening unit includes:

[0247] A region determination unit, which is used to determine the target texture region of the target image in the model texture coordinate space for each target image respectively.

[0248] A gradient characteristic extraction unit, which is used to extract the gray gradient change characteristics of each pixel point in the target texture region of the target image.

[0249] A characteristic screening unit, which is used to screen out the potential watermark embedding region of the target image based on the gray gradient change characteristics of each pixel point in the target texture region of the target image through the region growing criterion.

[0250] Optionally, in the watermark extraction device for a 3D model provided in another embodiment of the present application, the texture characteristic analysis unit includes:

[0251] A multi-scale analysis unit, which is used to perform multi-scale characteristic analysis on the potential watermark embedding region to obtain characteristics of multiple scales.

[0252] A texture characteristic calculation unit, which is used to calculate the difference between the characteristics of every two adjacent layers of scales to obtain multi-layer initial high-frequency texture characteristics.

[0253] A texture characteristic combination unit, which is used to combine the gray gradient change characteristics of the potential watermark embedding region with the initial texture high-frequency information of each layer to obtain an optimized high-frequency texture characteristic enhancement map.

[0254] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, it further includes:

[0255] A region verification unit, configured to perform boundary extraction and confirmation on the high-frequency texture feature enhancement map, and perform region consistency verification.

[0256] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, the frequency domain feature analysis unit includes:

[0257] A frequency domain data extraction unit, configured to extract high-frequency domain data from the frequency domain data of the potential watermark embedding region.

[0258] A frequency domain energy calculation unit, configured to weight the high-frequency domain data by using the high-frequency texture feature enhancement map to obtain the high-frequency energy of the potential watermark embedding region.

[0259] A decomposition unit, configured to perform discrete wavelet transform on the frequency domain data of the potential watermark embedding region, and decompose the frequency domain data of the potential watermark embedding region into multi-level frequency domain components. Among them, the frequency domain components include low-frequency components, horizontal high-frequency components, vertical high-frequency components, and diagonal high-frequency components.

[0260] An energy ratio calculation unit, configured to calculate multi-level energy ratios by using the high-frequency texture feature enhancement map and each level of frequency domain components.

[0261] A total feature component calculation unit, configured to combine the high-frequency texture feature enhancement map, the high-frequency energy, and each level of energy ratio to obtain a total feature component.

[0262] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, the image feature fusion unit includes:

[0263] An image feature fusion subunit, configured to fuse the high-frequency texture features and high-frequency frequency domain features of each potential watermark embedding region of each target image respectively according to the weights corresponding to the target attributes of each target image, to obtain fused texture features and fused frequency domain features. Among them, the target attributes at least include shooting angle, resolution, and confidence parameter.

[0264] Optionally, in the watermark extraction device for a three-dimensional model provided in another embodiment of the present application, the result generation unit includes:

[0265] An inverse mapping unit, configured to map the fused frequency domain features of the final watermark embedding region back to the model texture coordinate space to obtain a distribution map of the final watermark embedding region.

[0266] A coordinate acquisition unit, configured to acquire the position information of each feature point in the distribution map of the final watermark embedding region.

[0267] Among them, the position information includes texture coordinates and physical coordinates.

[0268] A confidence calculation unit, configured to calculate the confidence of each feature point according to the total characteristic component and the maximum characteristic component of each feature point in the distribution map of the final watermark embedding area.

[0269] A report generation unit, configured to generate and output a watermark extraction report by using the position information, confidence, and the distribution map of the final watermark embedding area of each feature point.

[0270] It should be noted that for the specific working processes of the above-mentioned units provided in the embodiments of the present application, reference can be made to the implementation processes of the corresponding steps in the above-mentioned method embodiments, which will not be elaborated here.

[0271] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0272] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A watermark extraction method for a three-dimensional model, characterized in that: include: Acquire multiple target images of a target three-dimensional model shot from multiple perspectives; For each of the target images, according to the grayscale gradient variation characteristics of the pixel points of the target image in the model texture coordinate space, a potential watermark embedding area of ​​the target image is screened out; The method comprises: performing a multi-scale high-frequency texture characteristic analysis on the latent watermark embedding area to obtain the high-frequency texture characteristic of the latent watermark embedding area, comprising: performing a multi-scale characteristic analysis on the latent watermark embedding area to obtain characteristics of multiple scales; calculating the difference between the characteristics of scales of each two adjacent layers to obtain the initial high-frequency texture characteristics of multiple layers; combining the grayscale gradient change characteristics of the latent watermark embedding area with the initial high-frequency texture characteristics of each layer to obtain an optimized high-frequency texture characteristic enhancement map; Performing frequency domain transformation on the potential watermark embedding area, and analyzing the high-frequency frequency domain characteristics of the potential watermark embedding area based on the frequency domain data of the potential watermark embedding area and the high-frequency texture characteristics; analyzing the high-frequency frequency domain characteristics of the potential watermark embedding area based on the frequency domain data of the potential watermark embedding area and the high-frequency texture characteristics, including: extracting high-frequency frequency domain data from the frequency domain data of the potential watermark embedding area; weighting the high-frequency frequency domain data using the high-frequency texture characteristic enhancement map to obtain the high-frequency energy of the potential watermark embedding area; performing discrete wavelet transformation on the frequency domain data of the potential watermark embedding area, and decomposing the frequency domain data of the potential watermark embedding area into frequency domain components at multiple levels; wherein the frequency domain components include low-frequency components, horizontal high-frequency components, vertical high-frequency components and diagonal high-frequency components; calculating the energy ratio of multiple levels using the high-frequency texture characteristic enhancement map and the frequency domain components at each level; combining the high-frequency texture characteristic enhancement map, the high-frequency energy and the energy ratios at each level to obtain the total characteristic component; The high-frequency texture characteristics and the high-frequency frequency domain characteristics of each of the latent watermark embedding regions of each of the target images are respectively fused to obtain fused texture characteristics and fused frequency domain characteristics; Determining a final watermark embedding area by statistically analyzing the fused texture characteristics and the fused frequency domain characteristics; The fused frequency domain characteristics of the final watermark embedding area are mapped back to the model texture coordinate space to obtain a watermark extraction result of the target three-dimensional model.

2. The method according to claim 1, characterized in that After acquiring a plurality of target images of the target three-dimensional model shot from multiple perspectives, the method further includes: Using histogram equalization technology to perform contrast enhancement on each of the target image data to obtain enhanced target images; Performing gamma correction processing on each of the enhanced target images to adjust the grayscale dynamic range of the target image; Correcting the geometric distortion of each of the adjusted target images by homography transformation; Adopting an adaptive median filtering algorithm to filter out noise from each of the corrected target images.

3. The method according to claim 1, characterized in that Before filtering out the potential watermark embedding area of ​​each target image according to the grayscale gradient variation characteristics of the pixel points of the target image in the model texture coordinate space, the method further includes: For each of the target images, extract feature points in the target image; Calculating characteristic description information of each of the feature points; wherein the characteristic description information of the feature point is information characterizing the local texture of the feature point; Based on the texture mapping parameters of the target three-dimensional model and the camera calibration parameters when shooting the target image, a mapping model between each of the feature points and the model texture coordinate space is established; Optimizing the mapping model by minimizing a total error function; Calculating the local consistency of each of the feature points in the optimized mapping model; Optimizing the feature points whose local consistency does not meet the preset conditions.

4. The method according to claim 1, characterized in that: The method of selecting a potential watermark embedding area of ​​each target image according to the grayscale gradient variation characteristics of the pixel points of the target image in the model texture coordinate space comprises: For each of the target images, determining a target texture region of the target image in the model texture coordinate space; Extracting the grayscale gradient variation characteristics of each pixel point in the target texture area of ​​the target image; Based on the grayscale gradient variation characteristics of each pixel point in the target texture region of the target image, the potential watermark embedding region of the target image is screened out through a region growing criterion.

5. The method according to claim 1, characterized in that: After combining the grayscale gradient change characteristics of the potential watermark embedding area with the initial high-frequency texture characteristics of each layer to obtain an optimized high-frequency texture characteristic enhancement map, the method further includes: Boundary extraction and confirmation are performed on the high-frequency texture characteristic enhancement map, and regional consistency verification is performed.

6. The method according to claim 1, characterized in that The step of fusing the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each of the latent watermark embedding regions of each of the target images to obtain fused texture characteristics and fused frequency domain characteristics comprises: According to the weights corresponding to the target attributes of each of the target images, the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each of the potential watermark embedding areas of each of the target images are fused respectively to obtain fused texture characteristics and fused frequency domain characteristics; wherein the target attributes include at least shooting angle, resolution and confidence parameters.

7. The method according to claim 1, characterized in that Mapping the fused frequency domain characteristics of the final watermark embedding area back to the model texture coordinate space to obtain the watermark extraction result of the target three-dimensional model includes: Mapping the fused frequency domain characteristics of the final watermark embedding area back to the model texture coordinate space to obtain a distribution map of the final watermark embedding area; Acquire the position information of each feature point in the distribution map of the final watermark embedding area; wherein the position information includes texture coordinates and physical coordinates; Calculate the confidence of each feature point according to the total feature component and the maximum feature component of each feature point in the distribution diagram of the final watermark embedding area; Using the location information and confidence level of each feature point and the distribution map of the final watermark embedding area, a watermark extraction report is generated and output.

8. A watermark extraction device for a three-dimensional model, characterized in that: include: An image acquisition unit, used to acquire multiple target images of a target three-dimensional model shot from multiple perspectives; A primary screening unit, for screening out a potential watermark embedding region of each target image according to the grayscale gradient variation characteristics of the pixel points of the target image in the model texture coordinate space; A texture characteristic analysis unit, used for performing multi-scale high-frequency texture characteristic analysis on the latent watermark embedding area to obtain the high-frequency texture characteristics of the latent watermark embedding area; The texture characteristic analysis unit comprises: a multi-scale analysis unit, a texture characteristic calculation unit and a texture characteristic combination unit; The multi-scale analysis unit is used to perform multi-scale characteristic analysis on the potential watermark embedding area to obtain characteristics of multiple scales; The texture characteristic calculation unit is used to calculate the difference between the scale characteristics of each two adjacent layers to obtain the initial high-frequency texture characteristics of multiple layers; The texture characteristic combining unit is used to combine the grayscale gradient change characteristics of the potential watermark embedding area with the initial high-frequency texture characteristics of each layer to obtain an optimized high-frequency texture characteristic enhancement map; A frequency domain transform unit, used for performing frequency domain transform on the latent watermark embedding area; A frequency domain characteristic analysis unit, configured to analyze the high frequency domain characteristics of the latent watermark embedding region based on the frequency domain data of the latent watermark embedding region and the high frequency texture characteristics; The frequency domain characteristic analysis unit includes: a frequency domain data extraction unit, a frequency domain energy calculation unit, a decomposition unit, an energy ratio calculation unit and a total characteristic component calculation unit; The frequency domain data extraction unit is used to extract high-frequency frequency domain data from the frequency domain data of the potential watermark embedding area; The frequency domain energy calculation unit is used to weight the high-frequency frequency domain data using the high-frequency texture characteristic enhancement map to obtain the high-frequency energy of the potential watermark embedding area; The decomposition unit is used to perform discrete wavelet transform on the frequency domain data of the potential watermark embedding area, and decompose the frequency domain data of the potential watermark embedding area into multi-level frequency domain components; wherein the frequency domain components include low frequency components, horizontal high frequency components, vertical high frequency components and diagonal high frequency components; The energy ratio calculation unit is used to calculate the energy ratio of multiple levels by using the high-frequency texture characteristic enhancement map and the frequency domain components of each level; The total characteristic component calculation unit is used to combine the high-frequency texture characteristic enhancement map, the high-frequency energy and the energy ratio of each level to obtain a total characteristic component; An image characteristic fusion unit, used to fuse the high-frequency texture characteristics and the high-frequency frequency domain characteristics of each of the latent watermark embedding regions of each of the target images, respectively, to obtain fused texture characteristics and fused frequency domain characteristics; A characteristic verification unit, used for determining a final watermark embedding area by statistically analyzing the fused texture characteristic and the fused frequency domain characteristic; The result generating unit is used to map the fused frequency domain characteristics of the final watermark embedding area back to the model texture coordinate space to obtain the watermark extraction result of the target three-dimensional model.

Images