Method for carrying out statistical analysis on concentric oolitic limestone strata based on brightness sensitivity

Abstract

The invention discloses a method for statistical analysis of concentric oolitic limestone strata based on brightness sensitivity, relates to the field of geology and sedimentary data analysis, and solves the problems of low efficiency and poor analysis precision of an existing oolitic analysis method. The method comprises the following steps: S1, collecting concentric oolitic image data; carrying out data preprocessing on the collected concentric oolitic image data; s2, regulating and analyzing the preprocessed data through a big data model, identifying light and dark fringe layer features in the concentric oolitic image, and performing fringe layer division on concentric oolites according to brightness change and thickness change features of fringe layers; and S3, according to a statistical analysis result, generating a concentric oolitic lamina graph. According to the method for performing statistical analysis on the concentric oolitic limestone lamina based on the brightness sensitivity, the precision of lamina identification can be improved, the analysis efficiency can be improved, the oolitic lamina can be efficiently and accurately subjected to comprehensive analysis, and automation and intelligence of oolitic lamina analysis are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of geological and sedimentological data analysis, and in particular to a method for statistically analyzing concentric oolite limestone laminae based on brightness sensitivity. Background Art

[0002] Ooliths are special particles in sedimentary rocks, mainly formed by carbonate materials. They usually have concentric light and dark alternating laminae. These laminae not only reflect the formation process of ooliths, but also record geological information such as sedimentary environment, climate change, chemical composition, etc. Therefore, the laminae analysis of ooliths is of great significance in the study of sedimentology, paleoenvironmental reconstruction and stratigraphic correlation.

[0003] Traditional oolite analysis methods rely on the experience of geologists and sedimentologists, who observe oolite samples under a microscope and manually divide and measure laminae. This method has significant limitations and low accuracy, which are mainly reflected in the following aspects:

[0004] High time cost: The manual observation and analysis process is very tedious, especially when a large number of oolite samples need to be processed, which is inefficient.

[0005] High subjectivity: Differences in experience and judgment standards among analysts can easily lead to differences in analysis results, affecting the accuracy and consistency of research results.

[0006] Lack of precision: Due to the possibility of fatigue and error in manual observation, traditional methods lack analytical accuracy and reliability.

[0007] With the rapid development of computer vision, big data analysis and artificial intelligence technology, the introduction of these emerging technologies into the field of oolite analysis has become the key to improving analysis efficiency and accuracy. Some existing studies have attempted to use automated image recognition and machine learning algorithms to analyze concentric oolite samples, but most methods are still in the preliminary stage and have not yet formed a mature system and method to conduct a comprehensive analysis of oolite laminae efficiently and accurately. Summary of the invention

[0008] The present invention aims to solve the problems of low efficiency and poor analysis accuracy of the existing oolite analysis methods in the background technology, and provide a method for statistical analysis of concentric oolite limestone laminae based on brightness sensitivity. The method for statistical analysis of concentric oolite limestone laminae based on brightness sensitivity can improve the accuracy of laminae identification, improve analysis efficiency, achieve efficient and accurate comprehensive analysis of oolite laminae, and realize automation and intelligence of oolite laminae analysis.

[0009] The present invention solves the problem through the following technical solution: the method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity comprises the following steps:

[0010] S1, collecting concentric oolite image data; performing data preprocessing on the collected concentric oolite image data;

[0011] S2. The preprocessed data is regulated and analyzed through a big data model to identify the light and dark laminae characteristics in the concentric oolite images, and the concentric oolite is divided into laminae according to the brightness and thickness variation characteristics of the laminae; and statistical analysis results are generated according to the distribution of the laminae;

[0012] S3. Generate a lamination map of the concentric oolite based on the statistical analysis results; the lamination map displays the light and dark lamination structure inside the concentric oolite and the characteristics of each lamination.

[0013] Furthermore, the method of collecting concentric ooid image data in step S1 includes:

[0014] Use a microscope to acquire high-resolution images of concentric oolitic grains; in the process of acquiring images of concentric oolitic grains, accurately adjust the microscope parameters and perform necessary image processing; according to the specific functions of the microscope, adjust the color balance and contrast parameters to optimize the image quality;

[0015] The image processing process includes adjusting the magnification of the microscope and carefully adjusting the focus to ensure that the image reaches the best clarity according to the size, characteristics and detail level of the concentric oolite image; adjusting the intensity of the reflected light to highlight the light and dark layers and details of the image until the image reaches the best clarity and detail display.

[0016] Furthermore, the step S1 performs data preprocessing on the collected concentric oolite image data, including: denoising, contrast enhancement, color correction and data conversion.

[0017] Furthermore, the concentric oolite image data was denoised using a filter algorithm to eliminate image interference caused by instrument noise or environmental factors.

[0018] Furthermore, the data conversion process uses labelimg and python to convert image information into a data set, and its specific process includes the following steps:

[0019] S21, reading concentric ooid image data;

[0020] S22, successfully loading and reading the concentric oolite image;

[0021] S23, color space conversion: after the image is loaded successfully, the image is converted from RGB color space to HSV color space;

[0022] S24, define a mouse callback function: after the color space conversion, define a function for extracting HSV values ​​in a directional manner, which will be called when the mouse clicks the image; use this function to obtain the HSV value of the click position and print it out;

[0023] S25. Create a window and set a mouse callback function to obtain the HSV value of the user's click position; click on different positions of the concentric oolite layer one by one in this way to obtain the unique color information of the concentric oolite, that is, the HSV value range of the concentric oolite; complete the data conversion process.

[0024] Furthermore, when the concentric ooid image is read in step S21, the image is read using the cv2.imread function;

[0025] The process of successfully loading the read concentric oolite image in step S22 is as follows: checking whether the concentric oolite image is successfully loaded, if the image fails to be loaded, printing an error message and exiting the program, and re-entering the program until the image is successfully loaded;

[0026] In the color space conversion process of step S23, after the concentric ooid image is successfully loaded, the concentric ooid image is converted from the RGB color space to the HSV color space using the cv2.cvtColor function;

[0027] The step S24 defines the mouse callback function process as follows: after the color space conversion, define a function for extracting the HSV value in a directional manner, and the function will be called when the mouse clicks the image; the function parameters include the event type event, the x and y coordinates of the click position, the flags flags and the passed parameter param; when using this function, inside the function, check whether the operation type is a left-click (cv2.EVENT_LBUTTONDOWN), if it is a left-click, the program understands that the HSV value of this position is to be tested, and the HSV value of the click position is obtained and printed out;

[0028] The process of creating a window and setting a mouse callback function in step S25 is as follows:

[0029] Use cv2.namedWindow to create a window, and use cv2.setMouseCallback to set the mouse callback function, and then you can get the HSV value of the position where the user clicks. By clicking different positions of the concentric oolite layer one by one in this way, you can get the HSV value range of the concentric oolite.

[0030] Furthermore, step S2 inputs the preprocessed data into the Ultralytics big data model for regulation and analysis; the Ultralytics model automatically identifies the light and dark lamination features in the concentric oolite image based on deep learning and big data algorithms, and divides the concentric oolite into laminae according to the brightness and thickness change characteristics of the laminae.

[0031] Furthermore, the recognition of the light and dark laminae features in the concentric oolite image can obtain the shape conditions of the concentric oolite based on the custom labels; the range of the color of the concentric oolite laminae can be obtained based on the color feature extraction and color analysis; the two conditions can be screened by the YOLO (You Only Look Once) large model to determine that the concentric oolite that meets both the shape condition and the color S value is the concentric oolite to be found; the specific steps include:

[0032] S31. Prepare annotated dataset

[0033] Custom label: objects with circular concentric layers and HSV color information within a certain range are defined as concentric oolitic particles;

[0034] The specific operation method of custom labeling is as follows: create a new data set, select circular, elliptical and nearly circular objects in labelimg; the selected circular, elliptical and nearly circular objects are shape information; in this new data set, not only the position of the object is marked, but also its corresponding color category is recorded;

[0035] S32. Color analysis

[0036] Target detection: First, the YOLO model is used to find circular, elliptical and nearly circular objects based on the shape information, and the location information of each detected object and the corresponding confidence score and label are obtained, aiming to preliminarily determine the location of objects that may be concentric oolitic particles based on simple shape information;

[0037] Color feature extraction and target screening: Using the location information obtained from the target detection in the previous step, the sub-images where the objects to be tested are located are intercepted on the original image; these sub-images are converted from the RGB color space to the HSV color space, where HSV is a color space more suitable for describing color characteristics; in the HSV color space, the saturation S component has an outstanding advantage over the hue H and brightness V in expressing image features. Therefore, after these sub-images are converted to the HSV space, the S value is assigned to the YOLO model; the S value range of this HSV is the range obtained after measuring a large number of sample images, that is, the range of oolite color; the HSV values ​​of each point in these sub-images are traversed, and objects whose HSV exceeds the above-assigned HSV range are eliminated, that is, objects whose shapes meet the requirements but whose colors do not meet the requirements are eliminated;

[0038] Perform color analysis: calculate the distribution of channel values ​​within the position information range obtained in the previous target detection step, the channel values ​​include the average value and the mode, so as to characterize the overall color tone tendency; or use a clustering algorithm (such as K-means) to analyze the color feature vector of each pixel inside the detection frame to obtain the most likely color category; the color inside the laminae and the color of the laminae transition boundary are very different, usually the boundary color is darker and the inside of the laminae is lighter; then set an upper limit based on the extreme difference between the color inside the laminae and the color characteristics of the laminae transition boundary. If this upper limit is exceeded, it means that the laminae boundary is included in the range, and the model automatically narrows the range until the boundary is not included; then the laminae can be identified.

[0039] Furthermore, the process of automatically dividing the concentric ooids into laminae according to the brightness change and thickness change characteristics of the laminae is as follows:

[0040] After completing the identification of concentric oolites, the model automatically divides the concentric oolite laminae according to the identified laminae characteristics, marks the boundaries of each layer, and generates analysis results based on the distribution of laminae;

[0041] After the concentric oolites were divided into laminae, the analysis results were verified and optimized through relevant samples in the big data to correct them and ensure the accuracy and consistency of the analysis.

[0042] Furthermore, according to the statistical analysis results, YOLO is used to automatically generate a lamination map of concentric oolites, which clearly shows the light and dark lamination structure inside the concentric oolites and the distribution of each lamination.

[0043] Compared with the above background technology, the present invention has the following beneficial effects:

[0044] The present invention provides a method for statistically analyzing concentric oolite limestone laminae based on brightness sensitivity, which has the following beneficial effects:

[0045] 1. Automation and intelligence

[0046] The present invention realizes the automation of concentric oolitic limestone laminae analysis through the Ultralytics big data model, greatly reducing the complexity and subjective interference of manual operation and improving the analysis efficiency. The entire process from data collection to laminae division can be completed automatically, saving a lot of time and labor costs.

[0047] 2. High-precision analysis

[0048] With the support of big data models, the present invention can accurately identify subtle differences in oolite laminae, avoiding errors caused by factors such as fatigue and experience differences during manual analysis. The deep learning algorithm of the model can fully explore the characteristics of laminae and improve the accuracy of analysis results.

[0049] 3. Innovative data processing methods

[0050] This paper proposes an innovative image data processing method, the core of which is to use the LabelImg tool to efficiently convert images into data sets that can be used for analysis. In the HSV color space, the three components of hue (H), saturation (S) and brightness (V) are deeply compared and analyzed, and it is found that the saturation S component has a particularly prominent advantage in expressing image characteristics.

[0051] Based on the outstanding advantages of the saturation S component in expressing image features, the present invention mainly focuses on in-depth mining and utilization of the saturation S feature in the subsequent data processing stage. This method not only optimizes the process of image feature extraction, but also significantly improves the efficiency and accuracy of image processing, providing a new and more efficient analysis method for image processing, target detection and image segmentation technologies.

[0052] 4. Data Reliability

[0053] The Ultralytics big data model can comprehensively consider the commonalities and differences of multiple samples by comparing and optimizing a large number of oolite sample data, thereby providing more reliable analysis results. The model's verification mechanism ensures the consistency and reliability of the analysis results.

[0054] 5. Visualize the analysis results

[0055] The generated laminae map intuitively displays the internal structure of oolites, providing researchers with a powerful analysis tool. Researchers can quickly understand the laminae structure of oolites through laminae maps, thus providing a basis for further geological research. The color information after data processing is attached to the original detection items, which can form a new data set with color annotations for subsequent operations.

[0056] In addition to using color, in order to improve the accuracy of prediction, we also provide processing that uses the shape of the object as an auxiliary. The specific operation is to first use labelimg to select the target object in the existing image to obtain a data set suitable for yolo; then divide the data set into a test set (test) and a training set (train). The specific training and adjustment process is roughly the same as above. Combining shape and color can accurately separate the ooids in the image. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flowchart of a method for statistically analyzing concentric oolite limestone laminae based on brightness sensitivity of the present invention;

[0058] Figure 2This is the HSV diagram of the concentric oolitic limestone laminae according to the embodiment of the present invention, wherein: a is the hue (H) diagram; b is the brightness (V) diagram; c is the saturation (S) diagram;

[0059] Figure 3 This is a comparison diagram of the oolite layer division of the Mantou Formation in an embodiment of the present invention; wherein: a is a microscopic diagram of the oolite of the Mantou Formation; b is a diagram of the oolite layer division of the Mantou Formation;

[0060] Figure 4 This is an oolitic laminae diagram of the Mantou Formation according to an embodiment of the present invention;

[0061] Figure 5 This is an oolitic laminae diagram of the Zhangxia Formation according to an embodiment of the present invention;

[0062] Figure 6 This is a yolo label data diagram of oolitic grains in the steamed bread group according to an embodiment of the present invention;

[0063] Figure 7 This is a partial HSV data diagram of oolitic grains after pretreatment in the steamed bread group according to an embodiment of the present invention. DETAILED DESCRIPTION

[0064] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0065] Figure 1 This is a flow chart of a method for statistical analysis of concentric oolite limestone laminae based on brightness sensitivity. It includes: S1, collecting concentric oolite image data; preprocessing the collected concentric oolite image data; S2, regulating and analyzing the preprocessed data through a big data model, identifying the characteristics of light and dark laminae in the concentric oolite images, and dividing the concentric oolite laminae according to the brightness change and thickness change characteristics of the laminae; generating statistical analysis results according to the distribution of laminae; S3, generating a laminae map of concentric oolite according to the statistical analysis results; the laminae map displays the light and dark laminae structure inside the concentric oolite and the characteristics of each laminae.

[0066] like Figure 1 As shown, a method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity comprises the following steps:

[0067] S1, collecting concentric oolite image data; performing data preprocessing on the collected concentric oolite image data;

[0068] S11. Collect concentric oolitic limestone sample data;

[0069] First, high-resolution images of the concentric oolitic limestone samples are collected using a microscope, preferably a scanning electron microscope (SEM). In the process of collecting images of samples such as concentric oolitic limestone, the key is to accurately adjust the microscope parameters and perform necessary image processing. This includes adjusting the magnification of the microscope according to the size, characteristics and degree of detail required for observation of the sample, carefully adjusting the focus to ensure the image is as clear as possible, adjusting the intensity of the reflected light to highlight the light and dark layers and details of the sample until the image achieves optimal clarity and detail. According to the specific functions of the microscope, other parameters such as exposure time, shutter speed, color balance and contrast need to be adjusted to further optimize the image quality. The collected data should ensure the clarity and details of the image for subsequent processing. The obtained light and dark layer image data will be input into the Ultralytics big data model for further analysis.

[0070] S12. Preprocessing the collected ooid image data to ensure the quality and accuracy of the data;

[0071] The collected ooid sample data is preprocessed to ensure the quality and accuracy of the data. That is, the image is enhanced and sharpened to highlight key features and ensure image clarity and information richness. This series of steps together ensures the high quality of the collected images and provides a reliable basis for subsequent analysis and research. The preprocessing includes the following steps:

[0072] De-noising: De-noising is performed on image data, using a filter algorithm to eliminate image interference caused by instrument noise or environmental factors.

[0073] Enhance contrast: Enhance the contrast of the image to make the brightness difference between oolite layers more significant, which is helpful for subsequent feature extraction and analysis.

[0074] Color correction: Since different devices may cause color deviation, the color of the image must be corrected to ensure that it is consistent with the actual situation so that the model can better identify the stratum features.

[0075] Data conversion: mainly use labelimg and python to convert image information into data sets. The specific process is as follows:

[0076] (1) Read the image: Use the cv2.imread function to read the image.

[0077] (2) Check whether the image is loaded successfully: If the image fails to load, print an error message and exit the program, then re-enter the program until the image is loaded successfully.

[0078] (3) Color space conversion: Use cv2.cvtColor function to convert the image from RGB color space to HSV color space.

[0079] (4) Mouse callback function: defines a function for extracting HSV values ​​in a directional manner, which will be called when the mouse clicks the image. The function parameters include the event type event, the x and y coordinates of the click position, the flags flags, and the passed parameter param. When using this function, inside the function, we check whether the operation type is a left-click (cv2.EVENT_LBUTTONDOWN). If it is, the program understands that the HSV value of this position is to be tested, and obtains the HSV value of the click position and prints it out. If this part of the function is not added, it is necessary to enter the coordinate position in the code to get the HSV value of a certain point, which is more convenient and quick. It should be noted that the function for determining the HSV value is built into the OpenCV library.

[0080] (5) Create a window and set the mouse callback function: Use cv2.namedWindow to create a window and use cv2.setMouseCallback to set the mouse callback function, and then you can get the HSV value of the position where the user clicks. In this way, by clicking on different positions of the ooid layer one by one, you can get the unique color information of the ooid, that is, the unique HSV value range of the ooid.

[0081] (6) Waiting for the user to press a key: Use cv2.waitKey to wait for the user to press a key, and close the window when the user presses any key.

[0082] S2. The pre-processed data is regulated and analyzed through a big data model to identify the characteristics of light and dark laminae in oolite samples, and the oolites are divided into laminae according to the brightness and thickness variation characteristics of the laminae; statistical analysis results are generated based on the distribution of laminae;

[0083] The preprocessed data is input into the Ultralytics big data model for further control and analysis. Based on deep learning and big data algorithms, the Ultralytics model can automatically identify the characteristics of light and dark laminae in oolite samples, and divide oolites into laminae according to the characteristics of laminae brightness changes, thickness changes, etc.

[0084] S21. Texture feature recognition: Using the YOLO (You Only Look Once) model, the process of color analysis and framing the target object mainly involves target detection related technologies. The specific steps include:

[0085] (1) Prepare annotated dataset

[0086] Custom label: Objects with circular concentric layers and HSV color information within a certain range are defined as oolitic particles.

[0087] The specific operation method is: create a new data set, select circular, elliptical and nearly circular objects in labelimg. In this new data set, not only the location of the object is marked, but also its corresponding color category is recorded. The previously selected circular, elliptical and nearly circular concentric layers are the shape information.

[0088] (2) Color analysis

[0089] Object detection: First, use the YOLO model to find circular, elliptical, and nearly circular objects based on shape information, and obtain the location information of each detected object and the corresponding confidence score and label. The purpose is to roughly determine the location of objects that may be ooids based on simple shape information.

[0090] Color feature extraction and target screening: Using the location information obtained from the previous target detection step, cut out the sub-images where each object to be tested is located on the original image. Convert these sub-images from the RGB color space to HSV, which is a color space more suitable for describing color characteristics.

[0091] In the HSV color space, Figure 2 This is the HSV graph of the oolite of the steamed bread group in the embodiment of the present invention, where a is the hue (H) graph; b is the brightness (V) graph; c is the saturation (S) graph; in-depth comparative analysis of the three components of hue (H), saturation (S) and brightness (V), it is found that the saturation S component has a particularly outstanding advantage in expressing image features. Therefore, after these sub-graphs are converted to the HSV space, the S value of the YOLO model is assigned. In the embodiment of the present invention, the oolite of the steamed bread group is taken as an example, and its S value is in the range of 89-247. The S value range of this HSV is the range we obtained after measuring a large number of sample images, that is, the range of oolite color. The computer can traverse the HSV value of each point in these sub-graphs and eliminate objects whose HSV exceeds the above range, that is, eliminate objects whose shapes meet the requirements but whose colors do not meet the requirements.

[0092] Perform color analysis: calculate the distribution of the values ​​of each channel within the range of the position information obtained previously, such as the mean and mode, to characterize the overall color tone tendency; or use a clustering algorithm (such as K-means) to analyze the color feature vector of each pixel inside the detection box to obtain the most likely color category. The color inside the laminae and the color of the laminae transition boundary are very different. Generally speaking, the boundary color is darker and the color inside the laminae is lighter. Then set an upper limit based on the extreme difference. In the embodiment of the present invention, taking the oolitic grains of the steamed bread group as an example, the upper limit of the difference between the color inside the laminae and the color of the laminae transition boundary is set to 20. If this upper limit is exceeded, it means that the laminae boundary is included in the range, and the model automatically narrows the range until the boundary is not included. The laminae can then be identified.

[0093] In summary, the process of lamination and oolite recognition is as follows: based on custom labels, the shape conditions of oolites can be obtained; based on color feature extraction and color analysis, the color range of oolite laminations can be obtained. By screening these two conditions with YOLO, it can be determined that those that simultaneously meet the shapes of circular, elliptical, and approximately circular and the color S value of 89-247 are the oolites we need to find.

[0094] S22. Automatic lamination division: After the oolitic grains are identified, the model automatically divides the laminations according to the identified lamination features, marks the boundaries of each layer, and generates analysis results based on the distribution of the laminations.

[0095] S23. Optimization and verification: Verify and optimize through relevant samples in big data, correct the analysis results, and ensure the accuracy and consistency of the analysis.

[0096] S3. Generate a laminogram of oolites based on the statistical analysis results; display the light and dark laminae structure inside the oolites and the characteristics of each laminae;

[0097] Based on the analysis results, YOLO was used to automatically generate the lamination map of oolites. The lamination map clearly shows the light and dark lamination structure inside the oolites, as well as the distribution of each lamination. The generated lamination map visualizes the microstructure of oolites, which is convenient for researchers to intuitively understand and further analyze.

[0098] Example 1

[0099] In order to make the purpose, technical scheme and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings, taking the oolitic limestone of Zhangxia Formation and Mantou Formation as an example.

[0100] The present invention provides a method for automatically counting the thickness of concentric oolitic limestone laminae based on brightness sensitivity. The conventional sample analysis process is as follows:

[0101] Collect concentric oolite sample data. The samples come from various sources, covering oolites in different geological backgrounds. This time, the concentric oolite limestone from the Zhangxia Formation and the Mantou Formation was mainly used.

[0102] After collecting the sample and making the thin section, a high-precision microscope is used to perform preliminary image acquisition of the concentric oolite images. During image acquisition, the magnification and focal length of the microscope are adjusted according to the specific conditions of the sample to ensure that the light and dark laminae of the oolite can be clearly captured. The light and dark laminae are important structural features inside the oolite. They record the environmental changes during the sedimentation process and are crucial for subsequent analysis.

[0103] The characteristics of oolitic grains observed under the microscope at this time are as follows: the oolitic structure inside the concentric layers of the Zhangxia Formation oolitic grains is micrite, the color gradually deepens from the inside to the outside, and the junction between the outer concentric layers and the inner micrite structure is a dark black organic layer; the oolitic grains of the Mantou Formation are mostly bright crystal cores, with small micrite or sand cores and concentric radial laminae, and occasionally bioclast concentric layers between oolitic grains. If the traditional manual method is used, the workload will be extremely huge.

[0104] After the image acquisition is completed, the image data needs to be preprocessed for subsequent statistical analysis. The main purpose of preprocessing is to remove noise from the image, enhance the contrast of the image, and perform color correction and data conversion. Denoising can eliminate image interference caused by instrument noise or environmental factors; enhancing contrast can make the light and dark layers in the image more distinct, which is convenient for subsequent feature recognition; color correction is to ensure the authenticity of the image color and avoid misleading during the analysis process. Data conversion: labelimg and python are mainly used to convert images into data sets to obtain the color information and shape position information of most oolitic layers for the next step of layer identification.

[0105] Preprocessing ensures that the collected image data is accurate and reliable, making it easier for the Ultralytics model to automatically identify the light and dark lamination features in oolite samples and divide the oolites into laminae based on the brightness and thickness changes of the laminae.

[0106] In this embodiment, the steamed bread group ooids are taken as an example. Figure 6 This is the label data graph of the oolo in the steamed bread group. Figure 7 This is the HSV data diagram of oolitic grains in the mantou group.

[0107] Using the model that selects circular, elliptical, and nearly circular objects in labelimg, we obtained a data set that represents the shape characteristics of ooids and can be applied to YOLO. Then the data set is divided into a test set and a training set to separate the ooids in the image, and the ooid image data that can be applied to the YOLO large model analysis is obtained. With the above data set, we can know that the color characteristics of the ooids in the Mantou Formation have a representative S value range of 89-247. The processing method of the ooids in the Zhangxia Formation is the same as above.

[0108] After preprocessing, the processed image data is input into the Ultralytics big data model to generate a laminogram. The Ultralytics model is built based on big data technology and machine learning algorithms. It can automatically identify laminar features in images, perform laminar division, and continuously optimize the results based on the model to ensure the accuracy and reliability of the analysis.

[0109] Figure 3 This is a comparison chart of the oolitic strata of the Mantou Formation. Figure 3 a is the oolitic image of the Mantou Formation, Figure 3 b is the laminae division diagram of the Mantou Formation oolites; by comparing the two, it can be found that the generated laminae diagram clearly shows the light and dark laminae structure inside the oolites, as well as the distribution of each laminae. The visual display of the microstructure of the oolites facilitates researchers' intuitive understanding and further analysis, making the conclusions more accurate and detailed.

[0110] Figure 4 This is the laminae of the Mantou Formation oolites. As shown in the figure, the laminae are generated. Through observation and analysis, it can be found that the depositional environment of the Mantou Formation oolites is sometimes a low-energy shallow marine shelf environment and sometimes a medium-energy shallow marine shelf environment. The oolite structure inside the concentric layer is micrite, and the color changes evenly. Some oolite concentric layers are affected by recrystallization and partially recrystallize to form spar calcite, and the original structure of the oolites is damaged to a certain extent.

[0111] Figure 5 This is the laminogram of oolitic grains of the Zhangxia Formation. As shown in the figure, the same operation was performed on the oolitic grains of the Zhangxia Formation, and it was found that the sedimentary environment when the oolitic grains of the Zhangxia Formation were formed was mainly a high-energy shallow marine shelf environment, with a small amount of medium-energy shallow marine shelf environment. The oolitic grain structure inside the concentric layer is micrite, and the color gradually deepens from the inside to the outside. The junction between the outer concentric layer and the internal micrite structure is a dark black organic layer. In addition, some oolitic grains are affected by recrystallization and are partially recrystallized into spar calcite, but a part of the original structure of the oolitic grains is still retained.

[0112] After actual operation, this method is more convenient than the traditional method. The efficient automatic processing can greatly improve the analysis efficiency of oolite laminae. The intelligent recognition method can accurately collect microscopic features that are difficult to identify with the naked eye, greatly improving the accuracy of oolite laminae analysis. In addition, the automated statistical analysis and generated laminae maps can be used to demonstrate and support subsequent conclusions.

[0113] Through the description of the above embodiments, the present invention solves the deficiencies in the existing ooid analysis methods through the following innovations:

[0114] Automated analysis: Automates the entire process of data collection, preprocessing, analysis and stratification, and automatic reading of valid data, reducing manual intervention and improving analysis efficiency.

[0115] High-precision recognition: Utilize big data models and machine learning algorithms to improve the accuracy of texture recognition and avoid errors caused by human factors.

[0116] Data reliability: Ensure the accuracy and consistency of analysis results through comparison and optimization of big data models.

[0117] Lamina map generation: Automatically generate clear lamina maps to intuitively display the internal structure of oolites, helping relevant researchers to interpret oolite information more accurately.

[0118] Through the above examples, the statistical analysis method of the present invention makes full use of modern artificial intelligence and big data technology to solve the problems of low efficiency and poor accuracy in traditional analysis methods, and provides more efficient and reliable technical support for the analysis of concentric oolitic limestone laminae. Through automated analysis and the generation of laminae maps, the present invention greatly improves the efficiency and accuracy of geological and sedimentological research, and helps to further promote scientific research and technological progress in experimental sedimentology.

[0119] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the implementation methods of the present invention, and should be understood that the protection scope of the present invention is not limited to such special statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A method for statistical analysis of concentric oolitic limestone laminae based on brightness sensitivity, characterized in that: The following steps are involved: S1, collecting concentric oolite image data; performing data preprocessing on the collected concentric oolite image data; S2. The preprocessed data is regulated and analyzed through a big data model to identify the light and dark laminae characteristics in the concentric oolite images, and the concentric oolite is divided into laminae according to the brightness and thickness variation characteristics of the laminae; and statistical analysis results are generated according to the distribution of the laminae; S3. Generate a lamination map of the concentric oolite based on the statistical analysis results; the lamination map displays the light and dark lamination structure inside the concentric oolite and the characteristics of each lamination.

2. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 1, characterized in that: Step S1: The method for collecting concentric ooid image data comprises: Use a microscope to acquire high-resolution images of concentric oolitic grains; in the process of acquiring images of concentric oolitic grains, accurately adjust the microscope parameters and perform necessary image processing; according to the specific functions of the microscope, adjust the color balance and contrast parameters to optimize the image quality; The image processing process includes adjusting the magnification of the microscope and carefully adjusting the focus to ensure that the image reaches the best clarity according to the size, characteristics and detail level of the concentric oolite image; adjusting the intensity of the reflected light to highlight the light and dark layers and details of the image until the image reaches the best clarity and detail display.

3. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 1, characterized in that: Step S1 performs data preprocessing on the collected concentric oolite image data, including: denoising, contrast enhancement, color correction and data conversion.

4. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 3, characterized in that: The concentric oolite image data were denoised using a filter algorithm to eliminate image interference caused by instrument noise or environmental factors.

5. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 3, characterized in that: The data conversion process uses labelimg and python to convert image information into a data set, and its specific process includes the following steps: S21, reading concentric ooid image data; S22, successfully loading and reading the concentric oolite image; S23, color space conversion: after the image is loaded successfully, the image is converted from RGB color space to HSV color space; S24, define a mouse callback function: after the color space conversion, define a function for extracting HSV values ​​in a directional manner, which will be called when the mouse clicks the image; use this function to obtain the HSV value of the click position and print it out; S25. Create a window and set a mouse callback function to obtain the HSV value of the user's click position; click on different positions of the concentric oolite layer one by one in this way to obtain the unique color information of the concentric oolite, that is, the HSV value range of the concentric oolite; complete the data conversion process.

6. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 5, characterized in that: When the step S21 reads the concentric ooid image, the cv2.imread function is used to read the image; The process of successfully loading the read concentric oolite image in step S22 is as follows: checking whether the concentric oolite image is successfully loaded, if the image fails to be loaded, printing an error message and exiting the program, and re-entering the program until the image is successfully loaded; In the color space conversion process of step S23, after the concentric ooid image is successfully loaded, the concentric ooid image is converted from the RGB color space to the HSV color space using the cv2.cvtColor function; The step S24 defines the mouse callback function process as follows: after the color space conversion, a function for extracting the HSV value in a directional manner is defined, and the function is called when the mouse clicks the image; the function parameters include the event type event, the x and y coordinates of the click position, the flag flags, and the passed parameter param; when the function is used, inside the function, it is checked whether the operation type is a left-click, and if it is a left-click, the program understands that the HSV value of this position is to be tested, and the HSV value of the click position is obtained and printed out; The process of creating a window and setting the mouse callback function in step S25 is as follows: using cv2.namedWindow to create a window, and using cv2.setMouseCallback to set the mouse callback function, and then the HSV value of the user's clicked position can be obtained; according to this method, the HSV value range of the concentric oolite layers can be obtained by clicking on different positions of the concentric oolite layers one by one.

7. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 1, characterized in that: Step S2 inputs the preprocessed data into the Ultralytics big data model for regulation and analysis; the Ultralytics model automatically identifies the light and dark lamination features in the concentric oolite image based on deep learning and big data algorithms, and divides the concentric oolite into laminae according to the brightness and thickness change characteristics of the laminae.

8. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 7, characterized in that: The recognition of light and dark laminae features in concentric oolite images can obtain the shape conditions of concentric oolites based on custom labels; the color range of concentric oolite laminae can be obtained based on color feature extraction and color analysis; through the screening of these two conditions by the YOLO large model, it can be determined that those that meet both the shape conditions and the color S value are the concentric oolites to be found; the specific steps include: S31. Prepare annotated dataset Custom label: objects with circular concentric layers and HSV color information within a certain range are defined as concentric oolites; The specific operation method of custom labeling is as follows: create a new data set, select circular, elliptical and nearly circular objects in labelimg; the selected circular, elliptical and nearly circular objects are shape information; in this new data set, not only the position of the object is marked, but also its corresponding color category is recorded; S32. Color analysis Target detection: First, the YOLO model is used to find circular, elliptical and nearly circular objects based on the shape information, and the location information of each detected object and the corresponding confidence score and label are obtained, aiming to preliminarily determine the location of objects that may be concentric oolitic particles based on simple shape information; Color feature extraction and target screening: Using the location information obtained from the target detection in the previous step, the sub-images where the objects to be tested are located are intercepted on the original image; these sub-images are converted from the RGB color space to the HSV color space. In the HSV color space, the saturation S component has an outstanding advantage over the hue H and brightness V in expressing image features. Therefore, after these sub-images are converted to the HSV space, the S value is assigned to the YOLO model; the S value range of this HSV is the range of the oolitic color; the HSV value of each point in these sub-images is traversed, and the objects whose HSV exceeds the above-assigned HSV range are eliminated, that is, the objects whose shapes meet the requirements but whose colors do not meet the requirements are eliminated; Perform color analysis: calculate the distribution of channel values ​​within the position information range obtained in the previous target detection step, the channel values ​​include the average value and the mode, so as to characterize the overall color tone tendency; or use a clustering algorithm to analyze the color feature vector of each pixel inside the detection frame to obtain the most likely color category; the color inside the laminae and the color of the laminae transition boundary are very different, usually the boundary color is darker and the inside of the laminae is lighter; then set an upper limit based on the extreme difference between the color inside the laminae and the color characteristics of the laminae transition boundary. If this upper limit is exceeded, it means that the laminae boundary is included in the range, and the model automatically narrows the range until the boundary is not included; then the laminae can be identified.

9. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 8, characterized in that: The process of automatic lamination division of concentric oolitic grains according to the brightness and thickness variation characteristics of the laminae is as follows: After completing the identification of concentric oolites, the model automatically divides the concentric oolite laminae according to the identified laminae characteristics, marks the boundaries of each layer, and generates analysis results based on the distribution of laminae; After the concentric oolites were divided into laminae, the analysis results were verified and optimized through relevant samples in the big data to correct them and ensure the accuracy and consistency of the analysis.

10. The method for statistically analyzing concentric oolitic limestone laminae based on brightness sensitivity according to claim 1, characterized in that: According to the statistical analysis results, YOLO was used to automatically generate the lamination map of concentric oolites, which clearly showed the light and dark lamination structure inside the concentric oolites and the distribution of each lamination.

Images