Geographic information data digital watermark detection method and system based on different scales
Patent Information
- Application Number
- CN202410215353.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-02-27
- Publication Date
- 2024-12-31
Smart Images
Figure CN119229158A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of data processing technology, and in particular relates to a method and system for detecting digital watermarks of geographic information data based on different scales. Background Art
[0002] Geographic information data refers to geographic spatial information represented in the form of electronic maps, satellite images or terrain models, which has important application value in many fields. However, with the development of digital technology, geographic information data also faces security risks such as illegal copying, piracy, and tampering. In order to protect the copyright and integrity of geographic information data, an effective technical means is needed to identify and verify the data.
[0003] Digital watermarking is an information security technology that can embed some hidden information into digital media, thereby authenticating, tracking and protecting digital content. Digital watermarking technology has been widely used in digital media such as images and videos, but for geographic information data, digital watermarking technology still has some challenges and shortcomings.
[0004] Geographic information data has the characteristics of different scales, that is, the resolution and details of the data change with the change of scale. This requires digital watermarking technology to be able to adapt to data of different scales while ensuring the concealment and robustness of the watermark. Geographic information data has high accuracy requirements, that is, the location, shape, topology and other attributes of the data cannot be destroyed or changed. This requires digital watermarking technology to be able to embed and detect watermarks without affecting the quality of the data.
[0005] Most of the existing digital watermarking technologies are designed for general image or video data, and do not take into account the characteristics and requirements of geographic information data. Therefore, it is necessary to develop a new method that specifically targets geographic information data of different scales to achieve effective digital watermark embedding and detection. Summary of the invention
[0006] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a digital watermark detection method and system for geographic information data of different scales, which solves the problem that the watermark detection technology in the prior art cannot effectively detect geographic information data of different scales.
[0007] The purpose of this disclosure can be achieved through the following technical solutions:
[0008] A method for detecting digital watermarks of geographic information data based on different scales comprises the following steps:
[0009] Data input and preprocessing: input geographic information data of different scales and standardize the input data;
[0010] Wavelet transform application: Use discrete wavelet transform to perform multi-scale decomposition on the pre-processed image and extract sub-images with different frequencies and spatial details;
[0011] Watermark detection: Apply discrete Fourier transform to each sub-image, analyze its frequency characteristics, extract possible watermark features from the sub-image, and determine the presence of watermark by comparing the output of the detection function with the threshold.
[0012] Furthermore, the method also includes using local binary patterns to extract basic texture and color features of the image, applying random forest methods to classify and identify key features and patterns in geographic information data, combining image segmentation technology to identify homogeneous areas in the image, marking these areas, and performing watermark detection again.
[0013] Furthermore, the method also includes detecting watermarks embedded at different scales by using a pyramid decomposition method.
[0014] Furthermore, the square pyramid decomposition method specifically comprises the following steps:
[0015] Apply multiple levels of downsampling to the image, each level of downsampling reduces the resolution by half, generating a series of down-resolution images;
[0016] Each level of downsampling can be achieved by subsampling after Gaussian smoothing to avoid aliasing effects;
[0017] At each level of the pyramid, the watermark detection algorithm is executed separately;
[0018] Images at each level may contain watermark information of different scales, and feature extraction technology is applied to images at each level.
[0019] Furthermore, after the watermark is detected, the decoded watermark content is compared with the original information to verify its accuracy.
[0020] Furthermore, the method locally amplifies the marked area.
[0021] Furthermore, the method applies discrete Fourier transform or discrete cosine transform to the marked area to analyze the frequency characteristics of the watermark.
[0022] Furthermore, the method applies a high-pass filter or a band-pass filter to the marked area to remove low-frequency background noise that can mask the watermark signal.
[0023] A digital watermark detection system for geographic information data based on different scales, the system comprising:
[0024] Data input and preprocessing module: used to input geographic information data of different scales and standardize the input data;
[0025] Wavelet transform application module: used to perform multi-scale decomposition of the pre-processed image using discrete wavelet transform and extract sub-images with different frequencies and spatial details;
[0026] Watermark detection module: It is used to apply discrete Fourier transform to each sub-image, analyze its frequency characteristics, extract possible watermark features from the sub-image, and determine the presence of watermark by comparing the output of the detection function with the threshold.
[0027] Beneficial effects of the present disclosure:
[0028] The present invention combines wavelet transform and frequency analysis to improve the recognition rate of various watermark types, especially those watermarks that are not easily perceived visually. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0030] Figure 1 It is a schematic diagram of the overall process of the embodiment of the present disclosure;
[0031] Figure 2 It is a flowchart of the corresponding watermark embedding method of the embodiment of the present disclosure. DETAILED DESCRIPTION
[0032] The following will be combined with the drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.
[0033] like Figure 1 As shown, a digital watermark detection method for geographic information data based on different scales includes the following steps:
[0034] Data input and preprocessing: input geographic information data of different scales, such as satellite images or terrain models.
[0035] The input data is standardized, including image scaling, format conversion, and basic image enhancement.
[0036] Image scaling: for original data of different resolutions, it is uniformly scaled to a standard size to maintain the integrity of image information and the correctness of proportion.
[0037] Format conversion: converting raw data into a unified image format, such as TIFF or JPEG, for subsequent processing.
[0038] Image enhancement, including contrast enhancement, sharpening, denoising, etc., to improve the recognizability of details in the image.
[0039] Apply wavelet transform to the standardized data;
[0040] Use discrete wavelet transform to perform multi-scale decomposition on the preprocessed image and extract sub-images with different frequencies and spatial details;
[0041] Discrete wavelet transform is used to decompose images and extract frequency and spatial information. For a given image I, its discrete wavelet transform can be expressed as
[0042] I DWT =DWT(I)
[0043] Among them I DWT is the wavelet coefficient matrix obtained after applying the discrete wavelet transform.
[0044] In discrete wavelet transform application, the image is decomposed into multiple sub-bands: near (LL), horizontal (LH), vertical (HL) and diagonal (HH). The watermark is usually embedded in one or more of these sub-bands.
[0045] Each decomposed sub-image is processed independently to retrieve the watermark that may be embedded in each frequency layer.
[0046] Discrete Fourier transform is applied to each sub-image to analyze its frequency characteristics.
[0047] Apply discrete Fourier transform to each subband after discrete wavelet transform to analyze its frequency characteristics. For each subband S, its DFT is defined as:
[0048] S DFT =DFT(S)
[0049] Where S DFT is the frequency representation of the subband.
[0050] Using specific algorithms, such as spectrum analysis or pattern recognition techniques, possible watermark features are extracted from the sub-images.
[0051] In the frequency domain, the presence of a watermark may affect the intensity or phase of a specific frequency. Define a detection function D(S DFT ) to identify these changes:
[0052] D(S DFT )=∑μ,υ|S DFT (μ,υ)-S DFTref (u,υ)|
[0053] Where S DFTref is the frequency representation of the reference subband without watermark, u, v, are the coordinates in the frequency domain; S DFT (u, v,) represents the amplitude of the subband S at the frequency point (u, v,) after discrete Fourier transform, S DFTref (u, v,) represents the amplitude of the corresponding reference subband without watermark embedded at the same frequency point.
[0054] The presence of a watermark is determined by comparing the output of the detection function with a value T. DFT )>T, the watermark exists.
[0055] If the above method does not detect the watermark;
[0056] Furthermore, local binary patterns are used to extract basic texture and color features of images, and random forest methods are applied to classify and identify key features and patterns in geographic information data.
[0057] Combined with image segmentation technology, homogeneous areas in the image are identified, such as large monochrome areas or areas with consistent textures, and these areas are marked.
[0058] After determining the most likely watermark insertion area, the marked area is processed as follows;
[0059] Zoom in locally to observe and process details more clearly.
[0060] Adjust contrast and brightness of selected areas to emphasize watermark features.
[0061] Apply discrete Fourier transform or discrete cosine transform on the selected area to analyze the frequency characteristics of the watermark.
[0062] Use a high-pass filter or a band-pass filter to remove low-frequency background noise that may mask the watermark signal.
[0063] Use edge detection techniques to detect subtle image changes caused by watermark embedding.
[0064] Furthermore, the initial threshold is set based on experience, and the watermark detection algorithm is tested using the collected samples. The detection results under different thresholds are recorded, and the test results are analyzed to determine the impact of the threshold on the detection accuracy. The threshold is adjusted according to the test results. If the false positive rate is too high, increase the threshold; if the false negative rate is too high, lower the threshold, and iterate the test: multiple iterations of the test, each time fine-tuning the threshold until the best balance is found.
[0065] The above-mentioned area is subjected to the pyramid decomposition method to detect the watermark embedded at different scales. The specific steps are as follows:
[0066] Apply multiple levels of downsampling to the image, each level of downsampling reduces the resolution by half, generating a series of downsampled images.
[0067] Each level of downsampling can be achieved by subsampling after Gaussian smoothing to avoid aliasing effects.
[0068] At each level of the pyramid, a watermark detection algorithm is performed separately. This may include the discrete wavelet transform and discrete Fourier transform mentioned earlier.
[0069] Each level of image may contain watermark information of different scales.
[0070] Feature extraction techniques, such as edge detection or texture analysis, are applied to the images at each level to determine possible locations for watermark embedding.
[0071] Analyze the features of images at different levels and identify the patterns and changes of watermarks.
[0072] After confirming the existence of the watermark, try to decode the watermark at each level.
[0073] Select an appropriate decoding strategy based on the watermark encoding method and embedding technology.
[0074] Furthermore, some disclosures combine known watermark embedding methods to analyze image features at different scales and identify regional watermark settings that may hide watermarks.
[0075] If a watermark is detected, further attempts are made to decode the watermark content, and the decoded watermark content is compared with the original information to verify its accuracy.
[0076] Furthermore, a separate database of detected watermarks is established and updated in real time for comparison and verification.
[0077] Embodiment 2, the detection method is applicable to the watermark embedding method, such as Figure 1 As shown, a method for embedding digital watermarks of geographic information data based on different scales includes the following steps:
[0078] S1: Data acquisition and analysis;
[0079] Obtain geographic information data, including satellite images, maps, terrain models, etc. Divide the data into high-resolution detailed data and low-resolution summary data according to the scale of the data. Perform feature analysis on these data, including color distribution, texture complexity, edge clarity, etc., to determine the most suitable area and method for embedding watermarks.
[0080] Data preprocessing,Before scale classification, the collected geographic information data must first be preprocessed.,This includes format unification, resolution adjustment, and basic image processing, such as,noise removal, contrast adjustment, etc., to ensure data quality and consistency.
[0081] Scale definition, High-resolution data: defined as maps with rich details, showing street level and even building details. Maps with a resolution of 1:5000 or higher fall into this category. Low-resolution data: defined as macro views, maps showing city overviews or larger areas of terrain. Maps with a resolution of 1:50,000 or lower fall into this category.
[0082] Determine the most suitable area and method for embedding watermarks; Area selection; High-resolution data: select areas that are rich in details but not too noticeable, such as the roofs of buildings, the edges of roads, etc. These areas are usually not the main focus of the observer and are therefore more suitable for hiding watermarks. Low-resolution data: select areas that are not noticeable in a macro view, such as open areas, the edges of water bodies, etc.
[0083] High-resolution data uses multi-layer watermarking technology, and the depth of each layer of watermarking does not exceed 5% of the original image depth. In the selected area, the pixel density of the watermark is kept at no more than 2,000 pixels per square centimeter.
[0084] For low-resolution data, a single-layer watermarking technology is used, and the watermark depth does not exceed 10% of the original image depth. In the selected area, the pixel density of the watermark is kept at no more than 1,000 pixels per square centimeter.
[0085] S2: Multi-layer watermark design
[0086] High-resolution data watermark design: Design complex watermark content for high-resolution data, such as copyright information, creation date, and geographic coordinates. Low-resolution data watermark design: Design simplified watermark content for low-resolution data, such as identifiers and short copyright statements.
[0087] High-resolution data watermark design, watermark content: For high-resolution data, the designed watermark content includes copyright information, creation date, geographic coordinates, copyright holder identification, and data source identification. The combination of this information not only provides copyright protection for the data, but also increases the possibility of its traceability.
[0088] Data matrix coding can store a large amount of information in a limited space and is easy to automatically identify and decode.
[0089] Watermarking levels and complexity: In high-resolution data, multiple watermarking levels can be set, each corresponding to a different type of information. For example, the first level contains basic copyright information, and the second level may contain more detailed data source and creation date information. This layered design increases the complexity of the watermark, thereby improving security.
[0090] Low-resolution data watermark design,Watermark content: The watermark content designed for low-resolution data is relatively simple, mainly including identifiers and copyright statements. Although the amount of information is small, it is enough to provide basic copyright protection. Coding method: Simple digital coding is used. This coding method occupies less space and is suitable for use in low-resolution data. At the same time, it can ensure the effectiveness and security of the watermark.
[0091] Through the above multi-layer watermark design, this embodiment can effectively protect geographic information data of different scales. The complex watermark content and advanced coding technology of high-resolution data provide comprehensive copyright protection and data traceability, while the simplified watermark content and compact coding method of low-resolution data ensure the concealment of the watermark and the visual quality of the data.
[0092] S3: Watermark Embedding Strategy
[0093] High-resolution data embedding: Combine spatial and frequency domain techniques for watermark embedding, such as DCT and DWT. Use machine learning algorithms to identify the best embedding area.
[0094] Low-resolution data embedding: A cautious embedding strategy is adopted to select non-critical visual areas of the data for embedding.
[0095] S3: Watermark embedding method
[0096] The watermark embedding method is optimized for geographic information data of different resolutions to ensure the effectiveness and concealment of the watermark.
[0097] High-resolution data embedding combines spatial and frequency domain techniques, discrete cosine transform and discrete wavelet transform for watermark embedding. DCT is suitable for capturing high-frequency information in images, while DWT is suitable for capturing multi-scale features of images. These techniques can be used to embed watermarks at different levels and regions, increasing the robustness and concealment of watermarks.
[0098] Region selection and embedding algorithm: Use machine learning algorithms and convolutional neural networks to analyze the characteristics of geographic information data and automatically identify the areas most suitable for embedding watermarks.
[0099] Regions with rich textures but not visually prominent are preferentially selected for embedding, such as building shadows, areas with dense trees, and areas without protrusions.
[0100] Building Shadows:
[0101] Characteristics: Shadow areas cast by buildings usually have more consistent textures, darker colors, and less visual attention. Differentiation method: Use image analysis technology to identify shadow areas, such as color histogram analysis and edge detection technology to find areas with darker colors and less obvious edge changes.
[0102] Densely treed areas: Features: Treed areas usually have complex textures, but they are usually not the main visual focus due to the repetitiveness of color and shape. Differentiation method: Use texture analysis methods, such as gray-level co-occurrence matrix or wavelet transform, to identify areas with high texture complexity.
[0103] Non-raised areas: Features: Non-raised areas, such as flat roofs or open spaces, are usually not visually prominent and have relatively uniform colors and textures. Differentiation method: Use height analysis and 3D modeling technology to identify flat or low-relief areas in geographic information data. At the same time, combined with illumination model analysis, the visual effects of these areas under different lighting conditions can be further confirmed.
[0104] Watermark strength and transparency adjustment, adjust the strength and transparency of the watermark according to the characteristics of the selected area to ensure the concealment of the watermark while ensuring that the watermark can still be detected under standard image processing and attacks.
[0105] Low-resolution data embedding adopts a more cautious and simplified embedding strategy. This paper discloses a direct embedding method based on the spatial domain.
[0106] Select non-critical visual areas of the data for embedding, such as large areas of natural terrain, water bodies, etc.
[0107] The watermark content should be more simplified and the size should be relatively small to reduce the impact on the overall image quality.
[0108] While maintaining the detectability of the watermark, its transparency is maximized to ensure that the original visual effect of the data is not affected.
[0109] S4: Watermark testing and optimization
[0110] The embedded watermark is tested for robustness such as compression, cropping, and rotation to ensure that the watermark is still recognizable after different processing. The embedding strategy is adjusted based on the test results to achieve the best watermark effect.
[0111] The specific steps of watermark robustness testing are as follows:
[0112] The watermarked images are compressed at different levels, such as JPEG and PNG, to simulate data compression situations that may be encountered in actual applications.
[0113] Evaluate the detectability and integrity of watermarks under different compression ratios.
[0114] The image is cropped at different scales and positions to test the stability of the watermark when part of the data is lost.
[0115] Apply different rotation angles and different scaling ratios to the image to test the watermark's ability to survive geometric transformations of the image. Evaluate the watermark's persistence under common geometric transformation operations.
[0116] In some disclosures, the image may be subjected to additional processing during the robustness test, such as adding noise, changing color balance, applying filter effects, etc., to simulate various possible image editing operations.
[0117] In some disclosures, the weaknesses of watermarks are analyzed based on test results, such as whether the watermark is vulnerable to damage under a specific compression ratio or cropping ratio, etc. The types of processing that have the greatest impact on the watermark are identified to prioritize optimization.
[0118] Adjust the embedding parameters of the watermark based on the analysis results, such as increasing the strength of the watermark or changing the embedding area.
[0119] For areas that are easily affected, consider using more complex watermarking techniques or adding redundant watermark information.
[0120] Multiple rounds of testing and tuning are performed to gradually improve the robustness and invisibility of the watermark.
[0121] After each round of testing, the watermark strategy is adjusted based on the feedback information, and redundant information is introduced into the watermark design, so that even if part of the watermark information is lost, the original information can still be restored until the best effect is achieved.
[0122] In some disclosures, the watermark is tested and optimized as follows;
[0123] During the compression process of formats such as JPEG or PNG, the watermark may become difficult to recognize due to data loss. By adjusting the embedding depth and strength of the watermark, more robust coding techniques such as enhanced error correction capabilities can be used.
[0124] Cropping of an image may cause the area containing the watermark to be removed, making the watermark undetectable. Distributed watermark embedding is used to distribute the watermark information in multiple areas of the image. In this way, even if some areas are cropped, the watermark can still be recovered from other areas.
[0125] Rotation and scaling of the image may distort or change the size of the watermark, affecting its detection. Use geometrically invariant watermarking techniques, such as those based on invariant moments or adaptive watermarking techniques, to maintain the stability of the watermark after geometric transformation.
[0126] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.
[0127] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.
Claims
1. A digital watermark detection method for geographic information data based on different scales, characterized in that: The following steps are involved: Data input and preprocessing: input geographic information data of different scales and standardize the input data; Wavelet transform: Use discrete wavelet transform to perform multi-scale decomposition on the preprocessed image and extract sub-images with different frequencies and spatial details; Watermark detection: Apply discrete Fourier transform to each sub-image, analyze its frequency characteristics, extract possible watermark features from the sub-image, and determine the presence of watermark by comparing the output of the detection function with the threshold.
2. According to claim 1, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The method also includes using local binary patterns to extract basic texture and color features of the image, applying random forest methods to classify and identify key features and patterns in geographic information data, combining image segmentation technology to identify homogeneous areas in the image, marking these areas, and performing watermark detection again.
3. According to claim 1, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The method further comprises, Watermarks embedded at different scales are detected by pyramid decomposition method.
4. According to claim 3, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The square pyramid decomposition method specifically comprises the following steps: Apply multiple levels of downsampling to the image, each level of downsampling reduces the resolution by half, generating a series of down-resolution images; Each level of downsampling can be achieved by subsampling after Gaussian smoothing to avoid aliasing effects; At each level of the pyramid, the watermark detection algorithm is executed separately; Images at each level may contain watermark information of different scales, and feature extraction technology is applied to images at each level.
5. According to claim 1, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: After the watermark is detected, the decoded watermark content is compared with the original information to verify its accuracy.
6. According to claim 2, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The method locally magnifies the marked area.
7. According to claim 2, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The method applies a discrete Fourier transform or a discrete cosine transform to the marked area to analyze the frequency characteristics of the watermark.
8. According to claim 2, a method for detecting digital watermarks of geographic information data based on different scales is characterized in that: The method applies a high-pass filter or a band-pass filter to the marked area to remove low-frequency background noise that can mask the watermark signal.
9. A digital watermark detection system for geographic information data based on different scales, characterized in that: The system comprises: Data input and preprocessing module: used to input geographic information data of different scales and standardize the input data; Wavelet transform application module: used to perform multi-scale decomposition of the pre-processed image using discrete wavelet transform and extract sub-images with different frequencies and spatial details; Watermark detection module: It is used to apply discrete Fourier transform to each sub-image, analyze its frequency characteristics, extract possible watermark features from the sub-image, and determine the presence of watermark by comparing the output of the detection function with the threshold.