Processing method for paleontoid cuticle graphic data

Abstract

The invention discloses a processing method for paleontoid epidermis graphic data, which is characterized in that in a data acquisition link, three-dimensional scanning equipment is applied to grasp original three-dimensional graphic data for subsequent processing, an analysis foundation is laid, a preprocessing stage is of great significance, high-frequency noise caused by a scanning environment and equipment precision is eliminated through de-noising processing, and the processing efficiency is improved. And when the obtained smooth three-dimensional data is converted into the two-dimensional texture mapping graph, errors and distortion are reduced, and the data availability is improved. In the textural feature extraction step, a gray-level co-occurrence matrix algorithm is applied to the field, textural feature parameters such as contrast, correlation, energy and entropy are calculated, quantitative description of paleontoid epidermis textures is realized, abstract textures are converted into comparable and analyzable numerical values, and key characteristics for identifying species are mined. In the texture classification stage, based on the extracted feature parameters, a model is constructed by means of machine learning algorithms such as a support vector machine and a random forest, and paleontological epidermis textures are accurately classified.

Description

Technical Field

[0001] The present application relates to the technical field of paleontological epidermal graphic data processing, and in particular to a method for processing paleontological epidermal graphic data. Background Art

[0002] Ancient organisms have gone through a long time, and their fossil forms are often incomplete, but their skin texture may retain key identification information. Accurately processing graphic data, extracting features such as texture direction and density, and comparing them with the skin texture database of known species can help determine the order of ancient organisms, clarify species boundaries, and reduce classification disputes.

[0003] By analyzing the evolution of epidermal textures of species from different geological periods and from the same or similar lineages, researchers were able to track the path of biological evolution. Data processing revealed the gradual change of textures from simple to complex, from regular to specific, which supplemented the key visual evidence chain for understanding how organisms adapt to environmental changes and gradually evolve.

[0004] Complete restoration of ancient biological ecology is inseparable from the accurate grasp of its body surface appearance. The processed epidermal graphic data can allow paleontologists to know the protective structure and camouflage texture of the animal's body surface, and thus infer its survival strategy and habitat in the ancient ecosystem, making the disappearing prehistoric ecological picture more vivid and detailed.

[0005] However, existing processing methods have many shortcomings. For example, texture analysis is not accurate enough and key features cannot be effectively extracted, which leads to insufficient depth in the study of paleontological epidermal graphic data and limits the progress of paleontological research. Summary of the invention

[0006] In an exemplary embodiment of the present application, a method for processing paleontological epidermal graphic data is provided to solve the problems in the prior art of inaccurate texture analysis and inability to effectively extract key features, thereby improving the research value of paleontological epidermal graphic data.

[0007] The present application provides a method for processing paleontological epidermal graphic data, which comprises:

[0008] Data collection step S1: using a 3D scanning device to scan the epidermis of a paleontological fossil or a paleontological specimen to obtain original 3D graphic data;

[0009] Preprocessing step S2: performing denoising on the original three-dimensional graphic data to remove high-frequency noise interference caused by the scanning environment and equipment accuracy, so as to obtain smooth three-dimensional graphic data and convert the three-dimensional graphic data into a two-dimensional texture map;

[0010] Texture feature extraction step S3: using a gray level co-occurrence matrix algorithm to calculate a plurality of texture feature parameters for the two-dimensional texture map, wherein the texture feature parameters include contrast, correlation, energy and entropy;

[0011] Texture classification step S4: Based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm to classify the ancient biological epidermal texture into a preset category system. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm.

[0012] Furthermore, the three-dimensional scanning equipment has an accuracy level of micrometer level, a scanning resolution of not less than 1000 sampling points per millimeter, and uses multi-view stitching technology to scan the epidermis of paleontological fossils or paleontological specimens from at least three different angles.

[0013] Furthermore, the preprocessing step S2: denoising the original three-dimensional graphic data by combining a Gaussian filtering algorithm and a median filtering algorithm, first removing salt and pepper noise by using a median filter, and then removing Gaussian noise by using a Gaussian filter, removing high-frequency noise interference caused by the scanning environment and equipment accuracy, and obtaining smooth three-dimensional graphic data; at the same time, converting the three-dimensional graphic data into a two-dimensional texture mapping image, and the conversion process follows the principle of conformal mapping to maintain the geometric characteristics of the texture.

[0014] Furthermore, in the Gaussian filtering algorithm, the filter kernel size is adaptively adjusted according to the noise intensity, and its value range is 3×3 to 7×7. The larger the standard deviation of the noise signal, the larger the filter kernel.

[0015] Furthermore, the texture feature extraction step S3: adopts the grayscale co-occurrence matrix algorithm, sets the grayscale quantization level to 16 levels, calculates multiple texture feature parameters for the two-dimensional texture map, and the texture feature parameters include contrast, correlation, energy and entropy. The calculation step size is set according to the fineness of the texture. The step size of the fine texture is 1 pixel, and the step size of the rough texture is 2-3 pixels.

[0016] Furthermore, in the texture feature extraction step S3, a wavelet transform algorithm is combined to perform secondary extraction on high-frequency texture details.

[0017] Furthermore, the texture classification step S4: based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm, and the data set is first divided into a training set and a test set in a ratio of 7:3. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm. A cross-validation method is used during training to optimize model hyperparameters.

[0018] Furthermore, it also includes a result verification step S5: cross-validating the classification model using paleontological epidermal sample data of known texture types, using a K-fold cross-validation method with a K value of 5-10. If the accuracy rate is lower than 90%, adjusting the classification model parameters and reclassifying, and the adjusted parameters include the kernel function type and the decision tree depth.

[0019] The embodiments of the present application have the following beneficial effects: the processing method for paleontological epidermal graphic data of the present invention uses a three-dimensional scanning device in the data collection link to capture the original three-dimensional graphic data for subsequent processing and lay the foundation for analysis. High-precision scanning can capture subtle textures that are difficult to see with the naked eye and avoid missing key information. The preprocessing stage is of great significance. The denoising process removes the high-frequency noise caused by the scanning environment and the accuracy of the equipment. When the obtained smooth three-dimensional data is converted into a two-dimensional texture map, the error and distortion are reduced, and the data availability is improved. In the texture feature extraction step, the grayscale co-occurrence matrix algorithm is used to calculate the texture feature parameters such as contrast, correlation, energy, and entropy to achieve a quantitative description of the epidermal texture of paleontological organisms, convert abstract textures into comparable and analyzable values, and dig out the key characteristics of species identification. In the texture classification stage, based on the extracted feature parameters, a model is constructed with the help of machine learning algorithms such as support vector machines and random forests to accurately classify the epidermal texture of paleontological organisms. This approach not only improves the classification accuracy, but also helps paleontologists quickly identify species and sort out the evolutionary context, making the entire paleontological epidermal graphic data processing more scientific and efficient, and providing solid support for scientific research. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0021] Figure 1 A flowchart of a method for processing ancient biological epidermal graphic data provided in an embodiment of the present application is exemplified. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0023] To further illustrate the technical solution provided by the embodiment of the present application, this is described in detail below in conjunction with the accompanying drawings and specific implementation methods. Although the embodiment of the present application provides the method operation steps shown in the following embodiments or drawings, more or fewer operation steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided in the embodiment of the present application.

[0024] refer to Figure 1 As shown, the present application provides a method for processing paleontological epidermal graphic data, which comprises:

[0025] Data collection step S1: Scan the epidermis of paleontological fossils or paleontological specimens using a three-dimensional scanning device to obtain original three-dimensional graphic data.

[0026] Furthermore, the accuracy level of the three-dimensional scanning equipment is at the micron level, and the scanning resolution is not less than 1000 sampling points per millimeter. It uses multi-view stitching technology to scan the epidermis of paleontological fossils or paleontological specimens from at least three different angles.

[0027] In the data collection step S1, a 3D scanning device with micron-level accuracy is used to scan the epidermis of fossils or specimens of ancient organisms. The scanning resolution of the device is not less than 1,000 sampling points per millimeter. This high-resolution setting ensures that extremely subtle texture information can be captured, without missing any potential key identification features on the epidermis of ancient organisms.

[0028] At the same time, multi-view stitching technology is used to scan from at least three different angles. Since the morphology of paleontological fossils or specimens is often irregular, scanning from a single perspective will inevitably miss the texture data of some areas. Multi-view scanning can cover the sample in all directions, make up for the blind spots of the perspective, and fully record the epidermal texture at all angles to obtain complete original three-dimensional graphic data.

[0029] Micron-level precision and high scanning resolution are combined to restore the true, microscopic texture of the ancient organism's epidermis to the greatest extent possible. The epidermis texture of ancient organisms contains rich species information, and high-precision acquisition can avoid data blurring and loss of details, so that subsequent processing and analysis have accurate basic materials.

[0030] The multi-angle stitching technology is used to scan from multiple angles in order to overcome the difficulty of scanning irregular samples. Paleontological samples have different shapes due to the erosion of time. It is impossible to see all the texture features by scanning from only one direction. Multi-angle scanning can fully record them, providing paleontological researchers with comprehensive and all-round epidermal texture data to support subsequent accurate identification and research work.

[0031] Preprocessing step S2: De-noising the original three-dimensional graphic data to remove high-frequency noise interference caused by the scanning environment and equipment accuracy, so as to obtain smooth three-dimensional graphic data and convert the three-dimensional graphic data into a two-dimensional texture map.

[0032] Further, the pre-processing step S2: the original three-dimensional graphic data is denoised by combining the Gaussian filtering algorithm and the median filtering algorithm, firstly the salt and pepper noise is removed by the median filtering, and then the Gaussian noise is removed by the Gaussian filtering, and the high-frequency noise interference caused by the scanning environment and the equipment accuracy is removed to obtain smooth three-dimensional graphic data. At the same time, the three-dimensional graphic data is converted into a two-dimensional texture mapping map, and the conversion process follows the principle of angle-preserving mapping to maintain the geometric characteristics of the texture.

[0033] Furthermore, in the Gaussian filtering algorithm, the filter kernel size is adaptively adjusted according to the noise intensity, and its value range is 3×3 to 7×7. The larger the standard deviation of the noise signal, the larger the filter kernel.

[0034] In the preprocessing step S2, the original three-dimensional graphic data collected in step S1 is firstly subjected to denoising. Here, a composite denoising method combining the Gaussian filter algorithm and the median filter algorithm is adopted:

[0035] First, the median filtering algorithm is used. This algorithm has a good removal effect on salt and pepper noise. It can effectively eliminate the salt and pepper noise that may appear in the data collection process, making the data relatively clean.

[0036] Then, the Gaussian filter algorithm is used to further process the median filtered data to remove Gaussian noise. In addition, the filter kernel size of the Gaussian filter will be adaptively adjusted according to the noise intensity, and its value range is 3×3 to 7×7, which is based on the standard deviation of the noise signal. The larger the standard deviation, the larger the filter kernel.

[0037] After the denoising operation is completed, the 3D graphic data is converted into a 2D texture map. In this conversion process, the angle-preserving mapping principle is followed, which ensures that during the conversion, the 2D texture map can maintain the geometric features of the texture in the original 3D graphic data to the greatest extent, so that the texture will not be seriously distorted in terms of geometric shape and relative position relationship of the texture during the conversion from 3D to 2D.

[0038] Since the original 3D graphics data is obtained in a scanning environment, it will inevitably be affected by the scanning environment and equipment accuracy and produce various noises, such as salt and pepper noise and Gaussian noise. These noises will interfere with the subsequent extraction and classification of texture features and affect the final analysis results. Therefore, they need to be removed to improve the quality and reliability of the data.

[0039] It is difficult for a single filtering algorithm to remove all types of noise. The combination of median filtering and Gaussian filtering can give full play to their respective advantages, carry out targeted processing of different types of noise, make the denoising effect more ideal, and allow subsequent processing processes to be based on purer data.

[0040] The conversion of 3D graphic data into 2D texture maps is to facilitate subsequent texture feature extraction and analysis operations. The application of the conformal mapping principle is to ensure that the geometric features of the texture are maintained during the conversion process, because these geometric features may contain important paleontological information, avoid the loss of key information due to conversion, and provide more accurate basic data for subsequent operations such as classification based on texture features.

[0041] Texture feature extraction step S3: using a gray level co-occurrence matrix algorithm to calculate multiple texture feature parameters for the two-dimensional texture map, the texture feature parameters including contrast, correlation, energy and entropy.

[0042] Furthermore, in the texture feature extraction step S3: a grayscale co-occurrence matrix algorithm is adopted, the grayscale quantization level is set to 16 levels, and multiple texture feature parameters are calculated for the two-dimensional texture mapping image. The texture feature parameters include contrast, correlation, energy and entropy. The calculation step size is set according to the fineness of the texture. The step size of the fine texture is 1 pixel, and the step size of the rough texture is 2-3 pixels.

[0043] Furthermore, in the texture feature extraction step S3, the high-frequency texture details are secondary extracted in combination with the wavelet transform algorithm.

[0044] In the texture feature extraction step S3, a variety of technical means are used to extract the texture feature parameters of the two-dimensional texture map of the ancient biological epidermis:

[0045] Grayscale co-occurrence matrix algorithm: First, the two-dimensional texture map is processed using the grayscale co-occurrence matrix algorithm. The grayscale quantization level is set to 16 levels, so that the grayscale information of the image can be divided into 16 levels, which is convenient for analyzing texture features of different grayscale levels. When calculating texture feature parameters, different calculation steps are set according to the fineness of the texture. For fine textures, the calculation step size is 1 pixel, while for rough textures, the step size is 2-3 pixels. This algorithm can calculate multiple texture feature parameters, including contrast, correlation, energy, and entropy.

[0046] Contrast reflects the contrast between different gray levels in an image and can be used to measure the clarity of a texture; correlation characterizes the linearity and directionality of a texture; energy reflects the consistency of a texture; the larger the energy value, the more regular the texture; entropy indicates the complexity of a texture; the higher the entropy value, the more complex and diverse the texture.

[0047] Combined with wavelet transform algorithm: In addition to the gray-level co-occurrence matrix algorithm, the wavelet transform algorithm is also combined to perform secondary extraction of high-frequency texture details. Wavelet transform is a multi-scale analysis method that can decompose an image into sub-bands of different frequencies. By analyzing the high-frequency sub-bands, more subtle texture details can be extracted, further supplementing and improving the texture feature information extracted by the gray-level co-occurrence matrix algorithm, and enhancing the comprehensive understanding and grasp of texture features.

[0048] Wavelet transform is a signal processing technology that decomposes the signal into sub-bands of different frequencies. For the two-dimensional texture map of the epidermis of ancient organisms, the wavelet transform can decompose the texture information into components of different scales and frequencies. Specifically, it uses a set of wavelet basis functions to transform the two-dimensional texture map, decomposing it into a low-frequency approximate part and high-frequency detail parts of different levels. The low-frequency part represents the general outline and trend of the texture, while the high-frequency part contains the detailed information of the texture, such as subtle changes in the texture, local mutations and texture edges.

[0049] In this scheme, after the gray-level co-occurrence matrix algorithm is used to extract the texture feature parameters (such as contrast, correlation, energy and entropy), the wavelet transform algorithm is used to extract the high-frequency texture details again, which means that the local detail information in the epidermal texture of the ancient organism can be captured more meticulously. For the high-frequency part of the information, it may be further quantified and analyzed, such as calculating the energy distribution and variance of the high-frequency part. These features can supplement the texture features obtained by the gray-level co-occurrence matrix algorithm, making the feature description of the epidermal texture of the ancient organism more comprehensive and rich.

[0050] The epidermal texture of ancient organisms often has a complex microstructure. Using only the gray-level co-occurrence matrix algorithm may not be able to fully capture all the texture details, especially the details of the high-frequency part. These high-frequency details may contain important identification information, such as fine lines, spots, cracks, etc. on the epidermis of ancient organisms. For different types of ancient organisms, their high-frequency texture details may be unique. By combining the wavelet transform algorithm to perform secondary extraction of high-frequency texture details, it is possible to avoid missing important texture features and make feature extraction more complete.

[0051] The main purpose of using the gray-level co-occurrence matrix algorithm is to quantify the texture features of the epidermis of ancient organisms. For paleontological research, the epidermal texture often contains important information. By quantifying these characteristic parameters, researchers can more intuitively compare the texture differences of different paleontological specimens, providing an objective basis for subsequent classification and identification. Different texture feature parameters describe the characteristics of the texture from different angles, providing multi-dimensional data support for accurately depicting the epidermal texture of ancient organisms.

[0052] Adjusting the calculation step size according to the fineness of the texture can better adapt to the characteristics of different textures, avoid losing information or introducing errors due to improper step size setting when extracting features, and improve the accuracy of feature extraction.

[0053] The introduction of wavelet transform algorithm is to make up for the shortcomings of a single algorithm. Since the epidermal texture of ancient organisms may contain rich high-frequency details, the gray-level co-occurrence matrix algorithm alone may not be able to fully extract these details, while wavelet transform can effectively capture high-frequency texture details, thereby more comprehensively extracting the texture features of the epidermis of ancient organisms, improving the integrity and accuracy of texture feature extraction, and providing a richer information basis for the subsequent construction of a more accurate classification model, thereby improving the processing effect and analysis ability of the entire processing method for the epidermal graphic data of ancient organisms.

[0054] Texture classification step S4: Based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm to classify the ancient biological epidermal texture into a preset category system. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm.

[0055] Furthermore, texture classification step S4: based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm. The data set is first divided into a training set and a test set in a ratio of 7:3. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm. Cross-validation method is used during training to optimize model hyperparameters.

[0056] In the texture classification step S4, a machine learning classification algorithm is used to classify the epidermal texture of the ancient organism. The specific operation is as follows:

[0057] Dataset division: First, the collected data set containing texture feature parameters is divided into a training set and a test set in a ratio of 7:3. The training set is used to train the classification model, and the test set is used to evaluate the performance of the model. This division method can ensure that there is enough data for model training, while leaving out a portion of independent data for testing, avoiding overfitting of the model to the training data and ensuring that the model has good generalization ability.

[0058] Classification algorithm selection: The machine learning classification algorithm used includes at least one of support vector machine (SVM) and random forest algorithm.

[0059] Support vector machine: It is a supervised learning method based on statistical learning theory. It separates data points of different categories by finding an optimal hyperplane. When dealing with the problem of ancient biological epidermal texture classification, according to the characteristics of the training set, the support vector machine can find a decision boundary that maximizes the interval between different texture categories, thereby classifying new texture data. Its kernel function can be selected from linear kernel, polynomial kernel, Gaussian kernel, etc. according to the specific situation to adapt to different data distribution and feature complexity.

[0060] Random forest: is an ensemble learning algorithm that classifies samples by building multiple decision trees and combining them. When building a random forest, the number of decision trees is dynamically set based on the sample size and feature dimension, ranging from a certain number (e.g., 50-500 trees). Each tree is trained based on random samples and random features of the training set, and the final classification result is determined by voting or averaging. It has good noise resistance and high accuracy.

[0061] Hyperparameter optimization: During the training process, the cross-validation method is used to optimize the model's hyperparameters. By further dividing the training set into multiple subsets, the hyperparameters are continuously adjusted and training and validation are performed on different subsets to find the optimal hyperparameter combination to improve the performance and accuracy of the classification model.

[0062] The purpose is to accurately classify the epidermal texture of ancient organisms into a preset category system based on the extracted texture feature parameters. The epidermal texture of ancient organisms contains rich information. Different textures may represent different species, evolutionary stages or living environments. By building a classification model, these texture features can be linked to known categories, providing a basis for the identification and research of ancient organisms.

[0063] The purpose of dividing the dataset into training and test sets and using cross-validation to optimize hyperparameters is to avoid overfitting and improve the generalization ability of the model. Overfitting can cause the model to perform well on the training data but poorly on new, unseen data. These operations ensure that the classification model can not only fit the training data well, but also accurately classify new paleontological epidermal texture samples in practical applications.

[0064] Both support vector machine and random forest algorithms are provided because different algorithms have different advantages and applicable scenarios. Support vector machines perform well when processing linear and nonlinear separable data, while random forests are more advantageous when processing high-dimensional data and noisy data. Providing multiple algorithm options allows you to flexibly select the most appropriate classification method based on the characteristics of different paleontological epidermal texture datasets, improve the accuracy and reliability of classification, provide more accurate classification tools for paleontological research, and help scientists better understand the diversity and evolution of paleontology.

[0065] Furthermore, it also includes a result verification step S5: cross-validating the classification model using paleontological epidermal sample data of known texture types, using a K-fold cross-validation method with a K value of 5-10. If the accuracy rate is lower than 90%, adjusting the classification model parameters and reclassifying, and the adjusted parameters include the kernel function type and the decision tree depth.

[0066] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0067] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0068] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0070] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for processing paleontological epidermal graphic data, characterized in that: It includes: Data collection step S1: using a 3D scanning device to scan the epidermis of a paleontological fossil or a paleontological specimen to obtain original 3D graphic data; Preprocessing step S2: performing denoising on the original three-dimensional graphic data to remove high-frequency noise interference caused by the scanning environment and equipment accuracy, so as to obtain smooth three-dimensional graphic data and convert the three-dimensional graphic data into a two-dimensional texture map; Texture feature extraction step S3: using a gray level co-occurrence matrix algorithm to calculate a plurality of texture feature parameters for the two-dimensional texture map, wherein the texture feature parameters include contrast, correlation, energy and entropy; Texture classification step S4: Based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm to classify the ancient biological epidermal texture into a preset category system. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm.

2. The method for processing ancient biological epidermal graphic data according to claim 1, characterized in that: The three-dimensional scanning device has an accuracy level of micrometers, a scanning resolution of not less than 1000 sampling points per millimeter, and uses multi-view stitching technology to scan the epidermis of paleontological fossils or paleontological specimens from at least three different angles.

3. The method for processing paleontological epidermal graphic data according to claim 2, characterized in that: The preprocessing step S2: denoising the original three-dimensional graphic data by combining a Gaussian filtering algorithm and a median filtering algorithm, first removing salt and pepper noise by using a median filter, and then removing Gaussian noise by using a Gaussian filter, removing high-frequency noise interference caused by the scanning environment and equipment accuracy, and obtaining smooth three-dimensional graphic data; at the same time, converting the three-dimensional graphic data into a two-dimensional texture mapping image, and the conversion process follows the principle of conformal mapping to maintain the geometric characteristics of the texture.

4. The method for processing paleontological epidermal graphic data according to claim 3, characterized in that: In the Gaussian filtering algorithm, the filter kernel size is adaptively adjusted according to the noise intensity, and its value range is 3×3 to 7×7. The larger the standard deviation of the noise signal, the larger the filter kernel.

5. The method for processing paleontological epidermal graphic data according to claim 4, characterized in that: The texture feature extraction step S3: adopts the grayscale co-occurrence matrix algorithm, sets the grayscale quantization level to 16 levels, calculates multiple texture feature parameters for the two-dimensional texture map, and the texture feature parameters include contrast, correlation, energy and entropy. The calculation step is set according to the fineness of the texture. The step size of the fine texture is 1 pixel, and the step size of the rough texture is 2-3 pixels.

6. The method for processing paleontological epidermal graphic data according to claim 5, characterized in that: In the texture feature extraction step S3, a wavelet transform algorithm is combined to perform secondary extraction on high-frequency texture details.

7. The method for processing paleontological epidermal graphic data according to claim 6, characterized in that: The texture classification step S4: based on the extracted texture feature parameters, a classification model is constructed using a machine learning classification algorithm. The data set is first divided into a training set and a test set in a ratio of 7:

3. The machine learning classification algorithm includes at least one of a support vector machine and a random forest algorithm. Cross-validation method is used during training to optimize model hyperparameters.

8. The method for processing paleontological epidermal graphic data according to claim 7, characterized in that: The method also includes a result verification step S5: using the ancient biological epidermis sample data with known texture types to cross-validate the classification model, using the K-fold cross-validation method, with K ranging from 5 to 10. If the accuracy is lower than 90%, the classification model parameters are adjusted and reclassified, and the adjustment parameters include the kernel function type and the decision tree depth.

Images