Three-dimensional model mapping watermark detection method and system based on multi-scale scaling

Abstract

The invention discloses a three-dimensional model chartlet watermark detection method and system based on multi-scale zooming. The method comprises the following specific steps: 1) reading a chartlet image, dividing image blocks, embedding a watermark and storing the chartlet image embedded with the watermark; 2) rotating the screenshot of a certain part of the three-dimensional model; 3) generating a multi-scale image; 4) performing noise processing to obtain a preprocessed image; 5) constructing a detection area, and extracting effective watermarks; and 6) performing statistics and verification to obtain a final watermark. The method has the advantages of multi-scale detection, high robustness, high efficiency, accuracy and high adaptability. The detection comprehensiveness is improved by generating a multi-scale image; the robustness is enhanced by combining multiple preprocessing means; statistical analysis is adopted to quickly screen real watermarks, and accuracy is improved. The method is suitable for mapping of various complex three-dimensional models and has a wide application prospect.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of digital image processing and watermark detection, and particularly relates to a three-dimensional model map watermark detection method and system based on multi-scale scaling. BACKGROUND

[0002] With the rapid development of three-dimensional modeling and digital content creation, the copyright protection of three-dimensional models has become an important issue. Watermark technology, as an effective means of copyright protection, is widely used in digital images and videos. However, traditional watermark detection methods mainly target two-dimensional images, and there are certain limitations for watermark detection of three-dimensional model maps. The complexity and diversity of three-dimensional model maps make watermark detection face many challenges, such as the watermark may be hidden in complex textures, or be destroyed in the transformation and rendering process of the model. Therefore, developing a detection method specifically for three-dimensional model map watermarking has important practical significance, which can effectively protect the copyright of three-dimensional models and prevent unauthorized copying and dissemination. SUMMARY

[0003] The present application aims to solve the problems existing in the prior art and provide a three-dimensional model map watermark detection method and system based on multi-scale scaling. This method can efficiently and accurately detect watermark information in three-dimensional model maps through multi-scale image processing, preprocessing and watermark extraction steps.

[0004] In order to achieve the above-mentioned application purposes, the present application specifically adopts the following technical solutions:

[0005] In the first aspect, the present application provides a three-dimensional model map watermark detection method based on multi-scale scaling, which includes the following steps:

[0006] S1. Obtain the map image of the three-dimensional model, divide the map image into a plurality of image blocks of the same size, and embed the same watermark information in each image block, to form an embedded watermark image and save it for three-dimensional model rendering and copyright protection;

[0007] S2. Use the embedded watermark map image to perform texture mapping on the three-dimensional model, take a screenshot of a part of the three-dimensional model with embedded watermark and obtain an image as a screenshot image, and perform rotation processing on the screenshot image to generate a plurality of rotated image versions;

[0008] S3. Perform multi-scale transformation on each rotated image version, and generate a plurality of images of different scales corresponding to each rotated image version;

[0009] S4. taking a rotated image version as a current image version, performing noise processing on each scale of the current image version to obtain a preprocessed image;

[0010] S5. performing multi-position detection on the preprocessed image, processing the detected watermark information, extracting valid watermarks conforming to a specific format, performing statistical analysis on all the extracted valid watermarks, recording the number of occurrences of each valid watermark using a hash table, and taking the valid watermark with the largest number of occurrences as the final detection result of the current image version;

[0011] S6. when the final detection results of all the rotated image versions are obtained, performing statistics on all the final detection results, taking the watermark with the largest number of occurrences as the final watermark, or otherwise, returning an error message.

[0012] On the basis of the above scheme, each step can be implemented in the following preferred specific manner.

[0013] As a preferred embodiment of the first aspect, in step S1, when embedding the watermark information in the image block, each character of the 4-bit hexadecimal watermark is first repeated once and connected, and then expanded into an 8-bit hexadecimal watermark, and then the watermark is embedded in each image block by a watermark embedding algorithm.

[0014] As a preferred embodiment of the first aspect, in step S2, when rotating the screenshot image, first, maintain the same angle interval within a preset angle range, and then rotate the screenshot image in the clockwise and counterclockwise directions in an angle-increment manner to generate a plurality of rotated image versions.

[0015] As a preferred embodiment of the first aspect, in step S3, the specific method for performing multi-scale transformation on each rotated image version is to scale the rotated image version according to a preset scale factor to obtain a plurality of images of different scales.

[0016] As a preferred embodiment of the first aspect, in step S4, the specific process of performing noise processing on each scale of the image is as follows: taking each scale of the image as an original image, calculating the median filtering result of the original image to obtain a noise image, then subtracting the noise image from the original image to obtain a noise difference image, then performing scaling processing on each pixel value of the noise difference image to enhance the noise feature to obtain a feature-enhanced image, and finally performing an inversion operation on the feature-enhanced image to obtain a preprocessed image.

[0017] As a preferred embodiment of the first aspect, the specific process of step S5 is as follows: a detection window with the same size as the image block is set, and the detection window is moved on the preprocessed image with a preset step size, so that a detection region is formed each time the detection window is moved, and a watermark extraction algorithm is used to perform watermark detection on each detection region to obtain watermark information.

[0018] As a preferred embodiment of the first aspect, the specific process of step S5 is as follows: when the characters in the detected watermark information satisfy the condition that every two characters are the same, the corresponding character combination is extracted to form the final effective watermark.

[0019] In a second aspect, the present application provides a three-dimensional model map watermark detection system based on multi-scale zooming, which comprises:

[0020] A watermark embedding module is configured to obtain a map image of a three-dimensional model, divide the map image into a plurality of image blocks with the same size, and embed the same watermark information in each image block, so that the image blocks after embedding the watermark form a map image after embedding the watermark and are saved for three-dimensional model rendering and copyright protection.

[0021] A rotation processing module is configured to use the map image after embedding the watermark to perform texture mapping on the three-dimensional model, take a screenshot of a part of the three-dimensional model after embedding the watermark and obtain an image after the screenshot, and perform rotation processing on the image after the screenshot to generate a plurality of rotated image versions.

[0022] A multi-scale transformation module is configured to perform multi-scale transformation on each rotated image version, so that each rotated image version generates a plurality of images with different scales.

[0023] A noise processing module is configured to use one rotated image version as a current image version, perform noise processing on the images with different scales of the current image version, and obtain a preprocessed image.

[0024] An effective watermark extraction module is configured to perform multi-position detection on the preprocessed image, process the detected watermark information, extract an effective watermark that meets a specific format, statistically analyze all the extracted effective watermarks, use a hash table to record the number of occurrences of each effective watermark, and use the effective watermark with the highest number of occurrences as the final detection result of the current image version.

[0025] A statistical and verification module is configured to, after all the rotated image versions obtain the final detection results, statistically analyze all the final detection results, use the watermark with the highest number of occurrences as the final watermark, or return an error message if the watermark detection fails.

[0026] In a third aspect, the present application provides a computer readable storage medium, wherein the storage medium stores a computer program, and when the computer program is executed by a processor, the method for detecting a multi-scale zoom-based three-dimensional model map watermarking according to any one of the first aspect is implemented.

[0027] In a fourth aspect, the present application provides a computer electronic device, comprising a memory and a processor;

[0028] The memory is configured to store a computer program;

[0029] The processor is configured to implement the method for detecting a multi-scale zoom-based three-dimensional model map watermarking according to any one of the first aspect when executing the computer program.

[0030] Compared with the prior art, the method for detecting a three-dimensional model map watermarking has the following advantages:

[0031] 1. Multi-scale detection: the present application can effectively cover different size ranges where the watermark may exist by generating multi-scale images, thereby improving the comprehensiveness and accuracy of detection.

[0032] 2. Strong robustness: the present application combines various preprocessing methods such as noise processing, color replacement and rotation processing, which can effectively deal with noise interference, color change and rotation deformation in the map image, and enhance the robustness of detection.

[0033] 3. High efficiency and accuracy: the present application uses statistical analysis method to verify the detected watermark, which can quickly filter out the real existing watermark, and improve the accuracy and reliability of detection.

[0034] 4. Strong adaptability: the method of the present application is suitable for various complex three-dimensional model maps, and can meet the watermark detection needs in different scenes, and has wide application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0035] Figure 1 The flowchart of the steps of the method of the present application;

[0036] Figure 2 The display diagram of a three-dimensional model;

[0037] Figure 3 The display diagram of the image after the part of the three-dimensional model is captured;

[0038] Figure 4 The system block diagram of the present application;

[0039] Figure 5 The display diagram of the first map image of the embodiment of the present application;

[0040] Figure 6 a display diagram of the first kind of map image after embedding the watermark;

[0041] Figure 7 a display diagram of a part of the first kind of map image after embedding the watermark;

[0042] Figure 8 a display diagram of the watermark extraction result of the first kind of map image;

[0043] Figure 9 a display diagram of the second kind of map image according to the embodiment of the present application;

[0044] Figure 10 a display diagram of the second kind of map image after embedding the watermark;

[0045] Figure 11 a display diagram of the second kind of map image after embedding the watermark added to the three-dimensional model;

[0046] Figure 12 a display diagram of a part of the second kind of map image after embedding the watermark;

[0047] Figure 13 a display diagram of the watermark extraction result of the second kind of map image;

[0048] Figure 14 a display diagram of the third kind of map image according to the embodiment of the present application;

[0049] Figure 15 a display diagram of the third kind of map image after embedding the watermark;

[0050] Figure 16 a display diagram of the third kind of map image after embedding the watermark added to the three-dimensional model;

[0051] Figure 17 a display diagram of a part of the third kind of map image after embedding the watermark;

[0052] Figure 18 a display diagram of the watermark extraction result of the third kind of map image. DETAILED DESCRIPTION

[0053] In order to make the above objectives, features and advantages of the present application more apparent, specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. It will be apparent, however, to one skilled in the art that the present application can be practiced without some or all of these specific details. In other instances, well known process steps have not been described in detail in order not to unnecessarily obscure the present application. The various embodiments of the present application can be combined with each other without conflict.

[0054] As Figure 1 shown in the figure, in a preferred implementation of the present application, the above-mentioned watermark detection method based on multi-scale zooming of three-dimensional model texture mapping includes the following steps S1-S6. The specific implementation process will be described below.

[0055] Step one: watermark embedding

[0056] S1. Obtain the texture image of the three-dimensional model, divide the texture image into multiple image blocks of the same size, and embed the same watermark information in each image block. The image blocks after embedding the watermark constitute the texture image after embedding the watermark and are saved for three-dimensional model rendering and copyright protection.

[0057] In step S1 of the present embodiment, first, the texture image of the three-dimensional model is read from the specified path, and then the texture image is divided into multiple image blocks with a size of 256x256 pixels. If the size of the texture image is not an integer multiple of 256, the texture image is appropriately cropped or padded to ensure that the size of each image block is 256x256 pixels. Then input a 4-bit hexadecimal watermark, and when embedding, each character is repeated once and connected to expand it into an 8-bit hexadecimal watermark. For example, if the input 4-bit hexadecimal watermark is 1234, each character in it is expanded to two identical characters in sequence and connected to form the string "11223344". After watermark expansion, the same watermark information is embedded in each 256x256 pixel image block by a watermark embedding algorithm. Finally, the texture image after embedding the watermark is saved to the specified path for subsequent three-dimensional model rendering and copyright protection.

[0058] It should be noted that in the present application, the above-mentioned watermark embedding algorithm can be selected by those skilled in the art according to actual needs, that is, the present application is not limited to a certain specific watermark embedding algorithm, which is not limited here.

[0059] Step two: rotation processing of the screenshot

[0060] S2. Use the texture image after embedding the watermark to texture map the three-dimensional model, take a screenshot of a part of the three-dimensional model after embedding the watermark and take the obtained image as the screenshot image, and perform rotation processing on the screenshot image: for different rotation angles θ, the rotation matrix takes the center (c x ,c y ) of the screenshot image as the origin, constructs an affine matrix M, performs affine transformation on the screenshot image to obtain a rotated image, and generates multiple rotated image versions under different rotation angles.

[0061]

[0062] Where c x and c y They are the two-dimensional coordinates of the center of the screenshot image.

[0063] In step S2 of this embodiment, after the watermark is embedded in the texture image of the 3D model, the texture image with embedded watermark is obtained, and the texture image with embedded watermark is pasted on the original 3D model to form the 3D model with embedded watermark. Then, a certain part of the 3D model with embedded watermark is screenshotted to obtain the screenshot image. Figure 2 As shown in the figure, it is a display diagram of a three-dimensional model. After taking a screenshot, the screenshot image is obtained, as shown in Figure 3 As shown. Furthermore, considering that the 3D model may rotate during use, causing the watermark's orientation to change, the screenshot image is rotated. Specifically, within a preset angle range, the screenshot image is rotated in both the clockwise and counterclockwise directions at the same angle interval, generating multiple rotated image versions. For example, the angle range is 2 degrees to 8 degrees, and each rotation increases by 2 degrees (that is, the angle interval is 2 degrees). During the first rotation, the screenshot image is first rotated 2 degrees clockwise to obtain the first rotated image, and then the screenshot image is rotated 2 degrees counterclockwise to obtain the second rotated image. During the second rotation, the screenshot image is first rotated 4 degrees clockwise to obtain the third rotated image, and then the screenshot image is rotated 4 degrees counterclockwise to obtain the fourth rotated image. During the third rotation, the screenshot image is first rotated 6 degrees clockwise to obtain the fifth rotated image, and then the screenshot image is rotated 6 degrees counterclockwise to obtain the sixth rotated image. During the fourth rotation, the screenshot image is first rotated 8 degrees clockwise to obtain the seventh rotated image, and then the screenshot image is rotated 8 degrees counterclockwise to obtain the eighth rotated image. Through this process, a total of 8 rotated images were generated, and each angle (2 degrees, 4 degrees, 6 degrees, 8 degrees) had both clockwise and counterclockwise rotation results to further increase the robustness of the detection.

[0064] Step 3: Multi-scale image generation

[0065] S3. Perform multi-scale transformation on each rotated image version, and each rotated image version generates multiple images of different scales.

[0066] In step S3 of this embodiment, when performing a multi-scale transformation on each rotated image version, the specific method is to scale the rotated image version according to a preset scale factor to obtain multiple image versions with different resolutions (i.e., images of different scales). This is done to cover the different scales at which the watermark may exist and increase the comprehensiveness of the detection. Finally, the generated multi-scale images are stored in a list to facilitate subsequent processing one by one.

[0067] Step 4: Noise Processing

[0068] S4. Taking a rotated image version as the current image version, performing noise processing on each scale image of the current image version to obtain a preprocessed image.

[0069] In step S4 of this embodiment, the specific process of performing noise processing on the image of each scale is as follows: taking the image of each scale as the original image I, calculating the median filtering result of the original image to obtain a noise image, and then subtracting the noise image I from the original image. noise , get the noise difference image I diff , and then scale each pixel value of the noise difference image p scaled , enhance the noise features and obtain the feature-enhanced image I enhance Finally, the image after feature enhancement is inverted to better highlight the watermark information and the preprocessed image I is obtained. output .

[0070] I noise =median(I,5)(2)

[0071] I diff =II noise (3)

[0072]

[0073] I output =255-I enhance (5)

[0074] Where median represents the median filter operation; Calcul represents the scaling process; scale represents the scaling factor; and p represents a pixel value in the original image.

[0075] Step 5: Watermark extraction

[0076] S5. Multi-position detection is performed on the preprocessed image, the detected watermark information is processed, valid watermarks conforming to a specific format are extracted, all the extracted valid watermarks are statistically analyzed, a hash table is used to record the number of occurrences of each valid watermark, and the valid watermark with the largest number of occurrences is taken as the final detection result of the current image version.

[0077] In step S5 of the embodiment, the specific process of performing multi-position detection on the preprocessed image is as follows: a detection window with the same size as the image block is set, the detection window is moved on the preprocessed image at a preset step, a detection region is formed each time the detection window is moved, a watermark extraction algorithm is used to perform watermark detection on each detection region to obtain watermark information.

[0078] Specifically, in order to improve the robustness of detection, the embodiment sets a 256*256 detection window, moves the detection window at a certain step (the horizontal step is 4 and the vertical step is 3), and calls a pre-written watermark extraction algorithm for each detection region to obtain the detected watermark information. It should be noted that in the present application, the watermark extraction algorithm described above is matched with the watermark embedding algorithm described above, and a person skilled in the art can select a watermark extraction algorithm corresponding to the watermark embedding algorithm according to the need, which is not limited herein.

[0079] In step S5 of the embodiment, the specific process of processing the detected watermark information to extract valid watermarks is as follows: when the characters in the detected watermark information satisfy the condition that every two characters are the same, the corresponding character combination is extracted to form the final valid watermark.

[0080] Step six: watermark verification

[0081] S6. After the final detection results are obtained for all the rotated image versions, the final detection results are statistically analyzed, the watermark with the largest number of occurrences is taken as the final watermark, otherwise the watermark detection fails, and an error message is returned.

[0082] It should be noted that in step S6 of the present application, when there is no watermark with the largest number of occurrences or the watermarks with the largest number of occurrences under different rotation angles are different, the watermark detection is considered to fail.

[0083] The present application will now be described through a specific example to show the application effect of the multi-scale zoom-based three-dimensional model texture watermark detection method described in steps S1-S6 of the above embodiment on an example, so as to facilitate understanding of the essence of the present application.

[0084] Embodiment

[0085] The specific implementation process of the multi-scale zoom-based three-dimensional model texture watermark detection method used in the embodiment is as described above and will not be repeated here.

[0086] As shown in Figure 5 , it is an example of a map image, and the map image obtained after embedding the watermark 1234 is shown in Figure 6 , which is pasted on the three-dimensional model of the bamboo chair using blender. It can be seen that there is no obvious difference between the two images after embedding the watermark. Then, the handle part of the bamboo chair (as shown in Figure 7 ) is intercepted for watermark detection, and it can be seen that the watermark can be successfully detected, as shown in Figure 8 . As shown in Figure 9 , it is a second example of a map image, and the map image obtained after embedding the watermark 2345 is shown in Figure 10 . It can be seen that there is no obvious difference between the two images after embedding the watermark. Then, the map image after embedding the watermark is pasted on the three-dimensional model of the rice washing basin using blender, as shown in Figure 11 . The bottom part of the rice washing basin (as shown in Figure 12 ) is intercepted for watermark detection, and the watermark detection result is shown in Figure 13 . As shown in Figure 14 , it is a third example of a map image, and the map image obtained after embedding the watermark 3344 is shown in Figure 15 . It can be seen that there is no obvious difference between the two images after embedding the watermark. Then, the map image after embedding the watermark is pasted on the three-dimensional model of the table using blender, as shown in Figure 16 . Part of the tabletop (as shown in Figure 17 ) is intercepted for watermark detection, and the watermark detection result is shown in Figure 18 .

[0087] In addition, it should be noted that the three-dimensional model map watermark detection method based on multi-scale zooming in the above embodiment can be executed by a computer program or module. Therefore, based on the same inventive concept, another preferred embodiment of the present application also provides a three-dimensional model map watermark detection system based on multi-scale zooming corresponding to the three-dimensional model map watermark detection method based on multi-scale zooming provided by the above embodiment, as shown in Figure 4 , which comprises:

[0088] a watermark embedding module for obtaining a map image of a three-dimensional model, dividing the map image into a plurality of image blocks of the same size, and embedding the same watermark information in each image block, the image blocks after embedding the watermark being used to form a map image after embedding the watermark and being saved for three-dimensional model rendering and copyright protection;

[0089] a rotation processing module for using the map image after embedding the watermark to perform texture mapping on the three-dimensional model, intercepting a part of the three-dimensional model after embedding the watermark and obtaining an image as a screenshot image, and performing rotation processing on the screenshot image to generate a plurality of rotated image versions;

[0090] a multi-scale transformation module, configured to perform a multi-scale transformation on each rotated image version, each rotated image version corresponding to generate a plurality of images of different scales;

[0091] a noise processing module, configured to take one rotated image version as a current image version, perform noise processing on each scale of image of the current image version, and obtain a preprocessed image;

[0092] an effective watermark extraction module, configured to perform multi-position detection on the preprocessed image, process the detected watermark information, extract an effective watermark conforming to a specific format, and perform statistical analysis on all the extracted effective watermarks, use a hash table to record the number of occurrences of each effective watermark, and take the effective watermark with the largest number of occurrences as the final detection result of the current image version;

[0093] a statistical and verification module, configured to, after the final detection result is obtained for all the rotated image versions, perform statistics on all the final detection results, take the watermark with the largest number of occurrences as the final watermark, or return an error message if the watermark detection fails.

[0094] Similarly, based on the same inventive concept, another preferred embodiment of the present application also provides a computer electronic device corresponding to the multi-scale zoom-based three-dimensional model map watermark detection method provided by the above-mentioned embodiment, which comprises a memory and a processor;

[0095] The memory is configured to store a computer program.

[0096] The processor is configured to, when executing the computer program, implement the multi-scale zoom-based three-dimensional model map watermark detection method in the above-mentioned embodiment.

[0097] In addition, the logical instructions in the above-mentioned memory can be implemented in the form of a software functional unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium, includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in the embodiments of the present application.

[0098] Therefore, based on the same inventive concept, another preferred embodiment of the present application also provides a computer readable storage medium corresponding to the multi-scale zooming based three-dimensional model map watermark detection method provided by the above-mentioned embodiment, and the storage medium stores a computer program. When the computer program is executed by a processor, the multi-scale zooming based three-dimensional model map watermark detection method in the above-mentioned embodiment can be realized.

[0099] It can be understood that the above-mentioned storage medium can include a random access memory (RAM) and can also include a non-volatile memory (NVM), such as at least one disk memory. Meanwhile, the storage medium can also be a U disk, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0100] It can be understood that the above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; and can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.

[0101] In addition, it should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-mentioned system can refer to the corresponding process in the above-mentioned method embodiment, and will not be described here. In the embodiments provided in the present application, the division of steps or modules in the system and method is only a logical functional division, and actual implementation can have another division mode, for example, multiple modules or steps can be combined or integrated together, or a module or step can be split.

[0102] The above-mentioned embodiments are only a preferred scheme of the present application, and are not used to limit the present application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, any technical scheme obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present application.

Claims

1. A 3D model texture watermark detection method based on multi-scale scaling, characterized in that: The following steps are involved: S1. Obtain a texture image of a 3D model, divide the texture image into multiple image blocks of equal size, and embed the same watermark information in each image block. The watermarked image blocks are used to form a watermarked texture image and save it for 3D model rendering and copyright protection. S2. Using the watermarked image to texture map the 3D model, taking a screenshot of a portion of the 3D model after the watermark is embedded and using the resulting image as the screenshot image, rotating the screenshot image to generate multiple rotated image versions; S3. Perform a multi-scale transformation on each rotated image version, where each rotated image version generates multiple images of different scales. S4. Using a rotated image version as the current image version, performing noise processing on each scale of the current image version to obtain a preprocessed image; S5. Perform multi-position detection on the preprocessed image, process the detected watermark information, extract valid watermarks that conform to a specific format, perform statistical analysis on all extracted valid watermarks, use a hash table to record the number of occurrences of each valid watermark, and use the valid watermark with the largest number of occurrences as the final detection result for the current image version; S6. After all the rotated image versions have obtained the final detection results, all the final detection results are counted and the watermark with the most occurrences is taken as the final watermark. Otherwise, the watermark detection fails and an error message is returned.

2. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S1, when embedding watermark information into an image block, each character of a 4-bit hexadecimal watermark is first repeated once and connected to expand it into an 8-bit hexadecimal watermark, and then the watermark is embedded into each image block using a watermark embedding algorithm.

3. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S2, when rotating the screenshot image, first, within a preset angle range, the same angle interval is maintained, and then the screenshot image is rotated in both clockwise and counterclockwise directions in an increasing angle manner to generate multiple rotated image versions.

4. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S3, the specific method of performing multi-scale transformation on each rotated image version is: scaling the rotated image version according to a preset scale factor to obtain multiple images of different scales.

5. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S4, the specific process of performing noise processing on the image of each scale is as follows: taking the image of each scale as the original image, calculating the median filtering result of the original image to obtain a noise image, then subtracting the noise image from the original image to obtain a noise difference image, then scaling each pixel value of the noise difference image to enhance the noise feature to obtain a feature-enhanced image, and finally performing an inversion operation on the feature-enhanced image to obtain a preprocessed image.

6. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S5, the specific process of performing multi-position detection on the preprocessed image is as follows: setting a detection window of the same size as the image block, moving the detection window on the preprocessed image with a preset step size, forming a detection area each time it moves, and using a watermark extraction algorithm to perform watermark detection on each detection area to obtain watermark information.

7. The method for detecting watermarks on a 3D model based on multi-scale scaling according to claim 1, wherein: In step S5, the specific process of processing the detected watermark information to extract the effective watermark is: when the characters in the detected watermark information satisfy that every two characters are the same, the corresponding character combination is extracted to form the final effective watermark.

8. A 3D model watermark detection system based on multi-scale scaling, characterized in that: include: A watermark embedding module is used to obtain a texture image of a 3D model, divide the texture image into multiple image blocks of the same size, and embed the same watermark information in each image block. The watermarked texture image is formed from the watermarked image blocks and saved for 3D model rendering and copyright protection. a rotation processing module, configured to use the watermarked image to perform texture mapping on the three-dimensional model, screenshot a portion of the watermarked three-dimensional model and use the resulting image as the screenshot image, and rotate the screenshot image to generate multiple rotated image versions; A multi-scale transformation module is used to perform multi-scale transformation on each rotated image version, and each rotated image version generates multiple images of different scales; A noise processing module is used to use a rotated image version as the current image version, perform noise processing on each scale of the current image version, and obtain a preprocessed image; The effective watermark extraction module is used to perform multi-position detection on the pre-processed image, process the detected watermark information, extract the effective watermarks that conform to the specific format, perform statistical analysis on all the extracted effective watermarks, use a hash table to record the number of times each effective watermark appears, and use the effective watermark with the largest number of occurrences as the final detection result of the current image version; The statistics and verification module is used to count all the final detection results after all the rotated image versions have obtained the final detection results, and take the watermark with the most occurrences as the final watermark, otherwise the watermark detection fails and an error message is returned.

9. A computer-readable storage medium, characterized in that The storage medium stores a computer program, and when the computer program is executed by the processor, the method for detecting watermarks of a three-dimensional model based on multi-scale scaling according to any one of claims 1 to 7 is implemented.

10. A computer electronic device, characterized in that: including memory and processor; The memory is used to store computer programs; The processor is configured to implement the three-dimensional model texture watermark detection method based on multi-scale scaling as claimed in any one of claims 1 to 7 when executing the computer program.

Images