Mineral identification method and device, electronic equipment and storage medium

Abstract

The invention discloses a mineral identification method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a sample microscopic tomography image and a scanning electron microscope mineral detection image matched with the sample microscopic tomography image; according to the sample microscopic tomography image and the scanning electron microscope mineral detection image, training a U-Net model to obtain a mineral identification model; and inputting a to-be-recognized microscopic tomography image into the mineral recognition model to obtain a mineral recognition result output by the mineral recognition model. According to the technical scheme provided by the invention, accurate identification of core minerals is realized, so that high-precision reconstruction and quantitative characterization of the three-dimensional digital core are realized.

Description

Technical Field

[0001] The present invention relates to the technical field of geological exploration, and particularly relates to a method and device for mineral identification, an electronic device, and a storage medium. Background Art

[0002] The three-dimensional reconstruction of digital cores can reflect the pore structure characteristics and pore throat coordination relationships of real cores, and is the basis for quantitative analysis of rock pore structures and rock physics numerical simulations.

[0003] In the Micro-CT (Micro Computed Tomography) images and SEM (Scanning Electron Microscope) images of cores, there is a characteristic that the pixel gray value is positively correlated with the density of substances. Specifically, pores are of low gray level, various matrix minerals are of medium-high gray level, and metal ore crystals are of high gray level. Therefore, in the three-dimensional digital core reconstruction containing different material phases, a gray-scale segmentation method is usually adopted. Specifically, by statistically analyzing the gray-scale distribution intervals of different material phases in the images, the three-dimensional reconstruction of various minerals is carried out. However, when performing statistical analysis based on gray values, for minerals with relatively close densities, effective segmentation and identification cannot be performed. Summary of the Invention

[0004] The present invention provides a method and device for mineral identification, an electronic device, and a storage medium, so as to achieve accurate identification of core minerals, thereby realizing high-precision reconstruction and quantitative characterization of three-dimensional digital cores.

[0005] In a first aspect, an embodiment of the present invention provides a method for mineral identification, the method including:

[0006] Determine a sample micro-computed tomography image and a scanning electron microscope mineral detection image matching the sample micro-computed tomography image;

[0007] Train a U-Net model according to the sample micro-computed tomography image and the scanning electron microscope mineral detection image to obtain a mineral identification model;

[0008] Input the micro-computed tomography image to be identified into the mineral identification model to obtain a mineral identification result output by the mineral identification model.

[0009] In a second aspect, an embodiment of the present invention further provides a device for mineral identification, the device including:

[0010] A sample determination module, configured to determine a sample micro-computed tomography image and a scanning electron microscope mineral detection image matching the sample micro-computed tomography image;

[0011] A model training module, configured to train a U-Net model based on the sample micro-computed tomography image and the scanning electron microscope mineral detection image to obtain a mineral recognition model;

[0012] A mineral recognition module, configured to input the micro-computed tomography image to be recognized into the mineral recognition model to obtain a mineral recognition result output by the mineral recognition model.

[0013] In a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the mineral recognition method described in any one of the embodiments of the present invention is implemented.

[0014] In a fourth aspect, an embodiment of the present invention further provides a storage medium storing computer-executable instructions. When the computer-executable instructions are executed by a computer processor, the mineral recognition method described in any one of the embodiments of the present invention is executed.

[0015] The technical solution of the embodiment of the present invention determines a sample micro-computed tomography image and a matching scanning electron microscope mineral detection image, and trains a U-Net model based on the sample micro-computed tomography image and the matching scanning electron microscope mineral detection image to obtain a mineral recognition model, and uses the mineral recognition model to identify the mineral type of the micro-computed tomography image to be recognized. It solves the problem that in the prior art, the method of identifying mineral types based on gray value segmentation cannot effectively segment and identify minerals with relatively close densities, realizes the accurate identification of core minerals, and thus realizes the high-precision reconstruction and quantitative characterization of three-dimensional digital cores.

[0016] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Description of the Drawings

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 is a flowchart of a mineral recognition method provided in Embodiment 1 of the present invention;

[0019] Figure 2 is a schematic diagram of the image processing process of a mineral recognition model provided in Embodiment 1 of the present invention;

[0020] Figure 3 It is a flowchart of a mineral identification method provided in the second embodiment of the present invention;

[0021] Figure 4 It is a schematic structural diagram of a mineral identification device provided in the third embodiment of the present invention;

[0022] Figure 5 It is a schematic structural diagram of an electronic device provided in the fourth embodiment of the present invention. Detailed implementation manners

[0023] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the protection scope of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0025] Embodiment 1

[0026] Figure 1 A flowchart of a mineral identification method is provided for the first embodiment of the present invention. This embodiment is applicable to the situation of identifying the mineral type of a micro-computed tomography image for three-dimensional digital core reconstruction. This method can be executed by a mineral identification device, which can be implemented in the form of hardware and / or software, and the mineral identification device can be configured in an electronic device.

[0027] As Figure 1 shown, the method includes:

[0028] S110. Determine a sample micro-computed tomography image and a scanning electron microscope mineral detection image matching the sample micro-computed tomography image.

[0029] Among them, the micro-computed tomography image, i.e., the Micro-CT image. The micro-computed tomography technology is a non-destructive 3D imaging technology that can clearly understand the internal microstructure of a sample without damaging the sample. The sample micro-computed tomography image is the Micro-CT image used for model training. The scanning electron microscope mineral detection image, i.e., the QEMSCAN (Quantitative Evaluation of Minerals by SCANning electron microscopy) image. The QEMSCAN image is a two-dimensional mineral dissemination image obtained by scanning the surface of the sample.

[0030] In this embodiment, since the QEMSCAN image is a two-dimensional mineral dissemination image obtained by surface scanning, therefore, performing micro-computed tomography and scanning electron microscope mineral detection on the core simultaneously can obtain the Micro-CT image and the QEMSCAN image corresponding to the same end face of the core. The gray-scale characteristics of various minerals can be reflected in the Micro-CT image, and the size characteristics, morphological characteristics, and dissociation relationship characteristics of various minerals can be reflected in the QEMSCAN image.

[0031] Furthermore, determining the sample micro-computed tomography image can further include:

[0032] A1. Obtain the original micro-computed tomography image obtained by performing micro-computed tomography on the core end face.

[0033] A2. Perform image preprocessing on the original micro-computed tomography image, and use the image after preprocessing as the sample micro-computed tomography image.

[0034] In this embodiment, due to the influence of factors such as light, environment, and shooting hardware, there is a lot of noise and redundant information in the original micro-computed tomography image (i.e., the original Micro-CT image) obtained by performing micro-computed tomography on the core end face. If the original Micro-CT image is directly used for model training, it is easy to cause an increase in model parameters, data redundancy, and a decline in model performance. Therefore, in this embodiment, image preprocessing is performed on the original Micro-CT image to remove the redundancy and noise in the original Micro-CT image.

[0035] Furthermore, the image preprocessing can include image denoising processing, image dimensionality reduction processing, etc. The specific method of image preprocessing in this embodiment is not limited. Exemplarily, the principal component analysis method can be used for image preprocessing. The principal component analysis method measures the difference of data with variance and projects the high-dimensional data with large differences into a low-dimensional space for representation. It is a common linear dimensionality reduction algorithm.

[0036] Further, after A2, it may further include:

[0037] B1. In the image after preprocessing, select the first target pixel points and set their gray values to zero to generate the first preset number of images as sample micro-computed tomography (Micro-CT) images.

[0038] B2. In the image after preprocessing, select the second target pixel points and perform Gaussian noise processing on them to generate the second preset number of images as sample micro-computed tomography (Micro-CT) images.

[0039] In this embodiment, since the QEMSCAN image is an image obtained by scanning the core end face, the corresponding number of sample Micro-CT images is small. To solve the problem that the insufficient number of sample Micro-CT images is likely to cause insufficient model training and overfitting, this embodiment performs image enhancement on the sample Micro-CT images to expand the sample Micro-CT images required for training the model.

[0040] Specifically, the image enhancement algorithm may include setting the gray value to zero and / or Gaussian noise processing.

[0041] Specifically, for the image after preprocessing, select the first target pixel points among the pixels of the image, set their gray values to zero, and a total of the first preset number of images are generated. The first preset number of generated images are added to the sample Micro-CT images to enhance the generalization ability of the model.

[0042] Among them, when selecting the first target pixel points among the pixels of the image, it can either be to select a preset number of pixel points as the first target pixel points, or to select a preset proportion of the total number of pixel points as the first target pixel points. The selection method can be random selection or selection according to a selection algorithm. This embodiment does not limit this.

[0043] Exemplarily, for the image after preprocessing, randomly select 10% of the total number of pixel points among the pixels as the first target pixel points, set their pixel values to zero, repeat the above operation, and a total of 1000 times the number of images after preprocessing are generated, and add them to the sample Micro-CT images.

[0044] Specifically, for the image after preprocessing, select the second target pixel points among the pixels of the image, perform Gaussian noise processing on the second target pixel points, and a total of the second preset number of images are generated. The second preset number of generated images are added to the sample Micro-CT images to enhance the anti-interference ability of the model.

[0045] Similarly, second target pixel points are selected among the pixels of the image. It can be to select a preset number of pixel points as the second target pixel points, or to select a preset proportion of the total number of pixel points as the second target pixel points. The selection method can be random selection or can be made according to a selection algorithm. The number of the first target pixel points and the second target pixel points can be the same or different, and this embodiment does not limit this.

[0046] Exemplarily, for the preprocessed image, 20% of the total number of pixel points are randomly selected among the pixels as the second target pixel points, and Gaussian noise is added to their pixel features. The above operations are repeated, and a total of 1000 times the number of preprocessed images are generated and added to the sample Micro-CT images.

[0047] In this embodiment, the sample Micro-CT images are augmented through gray value setting to zero and Gaussian noise processing, which ensures that while fully training the model, the generalization ability and anti-interference ability of the model are enhanced.

[0048] S120. Train the U-Net model according to the sample micro-tomography images and the scanning electron microscope mineral detection images to obtain a mineral recognition model.

[0049] In this embodiment, the U-Net model is selected for model training. The U-Net model is a type of convolutional neural network model. The U-Net structure has no fully connected layer, which can greatly reduce the model parameters to be trained, enabling the model to perform calculations on inputs of different sizes, and the U-Net structure can well retain the information in the image. The U-Net model includes a downsampling path and an upsampling path. The downsampling path is used to capture the context information in the image, and the upsampling path is used to accurately locate the parts that need to be segmented in the image.

[0050] Figure 2 A schematic diagram of the image processing process of a mineral recognition model is provided, as Figure 2 shown. The downsampling path of the mineral recognition model continuously performs convolution and pooling operations on the input image, performs 3×3 convolution on the image, and processes it using ReLu (Rectified Linear Unit). Then, 2×2 pooling processing is performed through the pooling layer to quickly reduce the feature dimension of the image and enhance feature invariance. The upsampling path of the mineral recognition model performs upsampling processing on the image, and restores the extracted features to the image through upsampling. And in the upsampled image, the image features extracted during the downsampling path process are added to avoid degradation when the model depth is large. In the final convolution operation, the convolution kernel size is 1×1, and the image is restored pixel by pixel to improve the accuracy of segmentation and recognition.

[0051] In this embodiment, the Micro-CT images, QEMSCAN images, and the augmented Micro-CT images obtained by scanning the same end face of the core are used for training the mineral recognition model, so that the training samples include not only the gray-scale features of various minerals, but also the spatial information features such as the size features, morphological features, and contact relationship features of the minerals. Thus, the trained mineral recognition model will use more dimensional information to comprehensively judge the prediction target, thereby improving the accuracy and reliability of the prediction results.

[0052] S130. Input the micro-tomography image to be recognized into the mineral recognition model, and obtain the mineral recognition result output by the mineral recognition model.

[0053] Among them, the Micro-CT image to be recognized is other tomographic images in the Micro-CT scan results of the core except for the end-face Micro-CT image.

[0054] In this embodiment, the Micro-CT images and QEMSCAN images scanned from the same end face of the core are used for training the mineral recognition model, and the training set of the sample Micro-CT images is augmented. The model is trained until it converges, and the converged model is used as the mineral recognition model to perform mineral recognition and classification on other tomographic images in the subsequent Micro-CT scan results, and finally realize the reconstruction of the three-dimensional digital core containing various mineral information.

[0055] The technical solution of the embodiment of the present invention determines the sample micro-tomography image and the matching scanning electron microscope mineral detection image, and trains the U-Net model according to the sample micro-tomography image and the matching scanning electron microscope mineral detection image to obtain a mineral recognition model, and uses the mineral recognition model to identify the mineral type of the micro-tomography image to be recognized. It solves the problem that in the prior art, the method of identifying mineral types based on gray-value segmentation cannot effectively segment and identify minerals with relatively close densities, realizes the accurate identification of core minerals, and thus realizes the high-precision reconstruction and quantitative characterization of three-dimensional digital cores.

[0056] Embodiment 2

[0057] Figure 3 It is a flowchart of a mineral recognition method provided by Embodiment 2 of the present invention. On the basis of the above embodiment, the training process of the mineral recognition model and the process of the mineral recognition model for identifying mineral types are further specified.

[0058] As Figure 3 shown, the method includes:

[0059] S210. Determine the sample micro-computed tomography (micro-CT) scan image and the scanning electron microscope (SEM) mineral detection image that matches the sample micro-CT scan image.

[0060] The sample micro-CT scan image may include the image after image preprocessing of the original micro-CT scan image, the image after pixel value zeroing of the image after image preprocessing, and the image after Gaussian noise processing of the image after image preprocessing. The process of obtaining the sample micro-CT scan image and the SEM mineral detection image as the model training set will not be elaborated in this embodiment.

[0061] S220. Square the gray value of each pixel point of the target sample micro-CT scan image to obtain a target image with increased square difference, and take the square root of the gray value of each pixel point of the sample micro-CT scan image to obtain a target image with increased square root difference.

[0062] Among them, the gray value of the pixel point of the target image with increased square difference can be represented by the following formula: V” ij = V' ij 2 where V' ij represents the gray value of the pixel point with coordinates (i, j) in the target sample micro-CT scan image, and V” ij represents the gray value of the pixel point with coordinates (i, j) in the target image with increased square difference.

[0063] The gray value of the pixel point of the target image with increased square root difference can be represented by the following formula: where V' ij represents the gray value of the pixel point with coordinates (i, j) in the target sample micro-CT scan image, and V” ij represents the gray value of the pixel point with coordinates (i, j) in the target image with increased square root difference.

[0064] In this embodiment, by squaring the gray value of each pixel point of the target sample micro-CT scan image, the difference between pixel points with larger gray values in the target sample micro-CT scan image can be increased. By taking the square root of the gray value of each pixel point of the target sample micro-CT scan image, the difference between pixel points with smaller gray values in the target sample micro-CT scan image can be increased. Such a setting increases the pixel value difference of the target sample micro-CT scan image, which can improve the accuracy and precision of model training and recognition.

[0065] S230. Use the target sample micro-CT scan image, the target image with increased square difference, and the target image with increased square root difference as a group of input images.

[0066] In this embodiment, the microscopic tomography image of the target sample, the target squared difference increased image, and the target square root difference increased image are taken as a group and input into the model simultaneously. Specifically, the input of the model is m×n×3, where 3 indicates that a group of inputs includes three images, and m×n represents the pixel size of the three images.

[0067] S240. Train the U-Net model according to each group of input images and the identification results of various types of minerals in the scanning electron microscope mineral detection images that match the sample microscopic tomography images in each group of input images to obtain a mineral identification model.

[0068] In this embodiment, in the scanning electron microscope mineral detection image that matches the sample microscopic tomography image, the types of different minerals, as well as the sizes, morphologies, contact relationships, etc. of different types of minerals can be reflected.

[0069] In this embodiment, the U-Net model is fully trained according to the sample microscopic tomography images obtained by scanning the core end face, the greatly expanded sample microscopic tomography images, and the corresponding scanning electron microscope mineral detection images until the model converges to obtain a mineral identification model.

[0070] Further, S240 can further include:

[0071] C1. Input each group of input images into the U-Net model to obtain an output image marked with one-hot encoding.

[0072] C2. Determine the mineral identification prediction results that match each group of input images according to the output image marked with one-hot encoding.

[0073] C3. Determine the actual mineral identification results that match each group of input images according to the scanning electron microscope mineral detection images that match the sample microscopic tomography images in each group of input images.

[0074] C4. Train the U-Net model according to the mineral identification prediction results and the actual mineral identification results that match each group of input images to obtain a mineral identification model.

[0075] Among them, one-hot encoding refers to one-hot encoding, which is used to convert discrete classification labels into binary vectors. Exemplarily, if there are a total of three types, the one-hot encoding includes a total of three positions, which are represented by 100, 010, and 001 respectively to represent three different types.

[0076] In this embodiment, the input of the U-Net model is m×n×3, and the output is m×n×s. s represents that there are s positions in the one-hot encoding, and there are s types of minerals. If it is represented by 1 / 0 whether the category corresponding to this position is the predicted category, the category corresponding to the position where 1 is located is the predicted category.

[0077] In this embodiment, through one-hot encoding, the mineral recognition prediction results corresponding to each pixel point of the output image are represented, that is, the predicted mineral types. At the same time, through the scanning electron microscope mineral detection images matching the sample micro-computed tomography images in this group of input images, the actual mineral recognition results corresponding to each pixel point can be determined, that is, the actual mineral types. The predicted mineral types and the actual mineral types are compared to determine the prediction accuracy of the model, and the model parameters are optimized according to the comparison results, and the model training is continuously carried out until the model converges.

[0078] S250. Square the gray value of each pixel point of the micro-computed tomography image to be recognized to obtain an image with increased square difference to be recognized, and, take the square root of the gray value of each pixel point of the sample micro-computed tomography image to obtain an image with increased square root difference to be recognized.

[0079] The process of squaring the gray value of each pixel point and the process of taking the square root of the gray value of each pixel point have been described above, and will not be elaborated in this embodiment.

[0080] S260. Input the micro-computed tomography image to be recognized, the image with increased square difference to be recognized, and the image with increased square root difference to be recognized into the mineral recognition model to obtain the mineral recognition result output by the mineral recognition model.

[0081] Similarly, the input of the mineral recognition model is m×n×3, and the output is m×n×s. According to the one-hot encoding of each pixel point of the output image, the mineral type corresponding to each pixel point is determined. Further, according to the connected domain composed of pixel points of the same mineral type, the size, shape, and contact relationship of this mineral type can also be analyzed.

[0082] Further, S260 can further include:

[0083] D1. Input the micro-computed tomography image to be recognized, the image with increased square difference to be recognized, and the image with increased square root difference to be recognized into the mineral recognition model to obtain the output image marked with one-hot encoding output by the mineral recognition model.

[0084] D2. Determine the mineral recognition result according to the output image marked with one-hot encoding.

[0085] Similarly, the output image output by the mineral recognition model is represented by m×n×s. Exemplarily, if s is 10, it means there are a total of 10 mineral types. If the one-hot encoding of a certain pixel point is 1000000000, it means the mineral type of this pixel point is the mineral type corresponding to the first digit of the one-hot encoding.

[0086] The technical solution of the embodiment of the present invention is as follows: After preprocessing the original micro-computed tomography (micro-CT) scan image, the obtained image is augmented to obtain a sample micro-CT scan image, which solves the problems such as insufficient model training and overfitting caused by a small number of samples, and ensures that while fully training the model, the generalization ability and anti-interference ability of the model are enhanced. According to the sample micro-CT scan image and the scanning electron microscope (SEM) mineral detection image, the mineral recognition model is trained. During the training process of the mineral recognition model, it not only has the gray-scale features of various minerals, but also includes spatial information features such as the size, shape, and contact relationship features of the minerals. When the obtained mineral recognition model is used for mineral recognition and classification, it will use more dimensional information to comprehensively judge the prediction target, thereby improving the accuracy and reliability of the prediction result. The input of the mineral recognition model is m×n×3. The input image and the image obtained by squaring the gray-scale value of each pixel point and taking the square root of the gray-scale value of each pixel point of the input image are used as a group of input images, which increases the image difference of the input image. The output of the mineral recognition model is m×n×s, where s represents s positions of one-hot encoding. According to the one-hot encoding of each pixel point of the output image, the mineral recognition result in the output image is determined.

[0087] Embodiment III

[0088] Figure 4 FIG. is a schematic structural diagram of a mineral recognition device provided in Embodiment III of the present invention. As Figure 4 shown, the device includes:

[0089] A sample determination module 310, configured to determine a sample micro-CT scan image and a scanning electron microscope (SEM) mineral detection image matching the sample micro-CT scan image;

[0090] A model training module 320, configured to train a U-Net model according to the sample micro-CT scan image and the SEM mineral detection image to obtain a mineral recognition model;

[0091] A mineral recognition module 330, configured to input a micro-CT scan image to be recognized into the mineral recognition model to obtain a mineral recognition result output by the mineral recognition model.

[0092] The technical solution of the embodiment of the present invention determines a sample micro-computed tomography (micro-CT) image and a matched scanning electron microscope (SEM) mineral detection image, and trains a U-Net model based on the sample micro-CT image and the matched SEM mineral detection image to obtain a mineral identification model, and uses the mineral identification model to identify the mineral type of the micro-CT image to be identified. It solves the problem that in the prior art, the method of identifying mineral types based on gray value segmentation cannot effectively segment and identify minerals with relatively close densities, realizes the accurate identification of core minerals, and thus realizes the high-precision reconstruction and quantitative characterization of three-dimensional digital cores.

[0093] Based on the above embodiment, the sample determination module 310 includes:

[0094] An original micro-CT image acquisition unit for acquiring an original micro-CT image obtained by micro-CT scanning of the core end face;

[0095] An image preprocessing unit for preprocessing the original micro-CT image and using the preprocessed image as a sample micro-CT image.

[0096] Based on the above embodiment, the sample determination module 310 includes:

[0097] A gray value processing unit for selecting a first target pixel point in the preprocessed image to set the gray value to zero and generating a first preset number of images as sample micro-CT images;

[0098] A Gaussian noise processing unit for selecting a second target pixel point in the preprocessed image to perform Gaussian noise processing and generating a second preset number of images as sample micro-CT images.

[0099] Based on the above embodiment, the model training module 320 includes:

[0100] A sample image difference increasing processing unit for performing pixel gray value square processing on the target sample micro-CT image to obtain a target square difference increasing image, and performing pixel gray value square root processing on the sample micro-CT image to obtain a target square root difference increasing image;

[0101] A sample input image determination unit for using the target sample micro-CT image, the target square difference increasing image, and the target square root difference increasing image as a group of input images;

[0102] A model training unit for training a U-Net model based on each group of input images and the recognition results of various types of minerals in the scanning electron microscope mineral detection images matched with the sample micro-computed tomography images in each group of input images to obtain a mineral recognition model.

[0103] Based on the above embodiments, the model training unit is specifically configured to:

[0104] Input each group of input images into the U-Net model to obtain an output image marked with one-hot encoding;

[0105] Determine the mineral recognition prediction results matched with each group of input images according to the output image marked with one-hot encoding;

[0106] Determine the actual mineral recognition results matched with each group of input images according to the scanning electron microscope mineral detection images matched with the sample micro-computed tomography images in each group of input images;

[0107] Train the U-Net model according to the mineral recognition prediction results and the actual mineral recognition results matched with each group of input images to obtain a mineral recognition model.

[0108] Based on the above embodiments, the mineral recognition module 330 includes:

[0109] A processing unit for increasing the difference of the image to be recognized, which is used to perform pixel gray value squaring processing on the micro-computed tomography image to be recognized to obtain an image with increased square difference to be recognized, and perform pixel gray value square root processing on the sample micro-computed tomography image to obtain an image with increased square root difference to be recognized;

[0110] An input unit for the image to be recognized, which is used to input the micro-computed tomography image to be recognized, the image with increased square difference to be recognized, and the image with increased square root difference to be recognized into the mineral recognition model to obtain the mineral recognition result output by the mineral recognition model.

[0111] Based on the above embodiments, the input unit for the image to be recognized is specifically configured to:

[0112] Input the micro-computed tomography image to be recognized, the image with increased square difference to be recognized, and the image with increased square root difference to be recognized into the mineral recognition model to obtain the output image marked with one-hot encoding output by the mineral recognition model;

[0113] Determine the mineral recognition result according to the output image marked with one-hot encoding.

[0114] The mineral recognition device provided by the embodiments of the present invention can execute the mineral recognition method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method.

[0115] Example 4

[0116] Figure 5 FIG. shows a schematic structural diagram of an electronic device 10 that can be used to implement the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, personal digital processors, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0117] As Figure 5 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. The input / output (I / O) interface 15 is also connected to the bus 14.

[0118] A plurality of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0119] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the mineral identification method.

[0120] In some embodiments, the mineral identification method may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the mineral identification method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the mineral identification method by any other suitable means (e.g., by means of firmware).

[0121] The various embodiments of the systems and techniques described above in this document may be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs that may be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0122] The computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The computer programs may be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine, or entirely on the remote machine or server.

[0123] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0124] To provide for interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide for interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic, speech, or tactile input).

[0125] The systems and techniques described herein can be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0126] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The relationship between the client and the server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0127] It should be understood that various forms of processes shown above can be used, steps can be reordered, added or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0128] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A mineral identification method, characterized in that: include: Determining a sample microscopic tomographic image and a scanning electron microscope mineral detection image matching the sample microscopic tomographic image; According to the sample microscopic tomography image and the scanning electron microscope mineral detection image, a U-Net model is trained to obtain a mineral recognition model; The microscopic tomographic image to be identified is input into the mineral identification model to obtain the mineral identification result output by the mineral identification model.

2. The method according to claim 1, characterized in that Determine the sample micro-tomographic image, including: Acquire an original micro-tomography image obtained by micro-tomography of the core end face; The original microscopic tomography image is preprocessed, and the preprocessed image is used as a sample microscopic tomography image.

3. The method according to claim 2, characterized in that After the original microscopic tomography image is preprocessed, the method further includes: In the preprocessed image, a first target pixel is selected to set the gray value to zero, and a first preset number of images are generated as sample microscopic tomography images; And / or, in the image after preprocessing, a second target pixel point is selected to perform Gaussian noise processing to generate a second preset number of images as sample microscopic tomography images.

4. The method according to claim 1, characterized in that: According to the sample microscopic tomography image and the scanning electron microscope mineral detection image, the U-Net model is trained to obtain a mineral recognition model, including: Performing pixel grayscale value square processing on the target sample microscopic tomography image to obtain a target square difference increased image, and performing pixel grayscale value square root processing on the sample microscopic tomography image to obtain a target square root difference increased image; Taking the target sample microtomographic image, the target square difference augmented image and the target root difference augmented image as a group of input images; According to each group of input images and the various types of mineral recognition results of the scanning electron microscope mineral detection images matched with the sample microscopic tomography images in each group of input images, the U-Net model is trained to obtain a mineral recognition model.

5. The method according to claim 4, characterized in that According to each group of input images and the various types of mineral recognition results of the scanning electron microscope mineral detection images matched with the sample microscopic tomography images in each group of input images, the U-Net model is trained to obtain a mineral recognition model, including: Input each group of input images into the U-Net model to obtain the output image marked with one-hot encoding; According to the output images marked by one-hot encoding, the mineral identification prediction results matching each group of input images are determined; Determine the actual results of mineral identification matching each set of input images based on the SEM mineral detection images matching the sample microscopic tomography images in each set of input images; According to the mineral identification prediction results and actual results that match each group of input images, the U-Net model is trained to obtain a mineral identification model.

6. The method according to claim 1, characterized in that Inputting the microscopic tomography image to be identified into the mineral identification model to obtain the mineral identification result output by the mineral identification model includes: Performing pixel grayscale value square processing on the microscopic tomography image to be identified to obtain a square difference-enhanced image to be identified, and performing square root processing on the pixel grayscale value of the sample microscopic tomography image to obtain a square root difference-enhanced image to be identified; The microscopic tomography image to be identified, the square difference enlarged image to be identified and the square root difference enlarged image to be identified are input into the mineral identification model to obtain the mineral identification result output by the mineral identification model.

7. The method according to claim 6, characterized in that Inputting the microscopic tomography image to be identified, the square difference increase image to be identified, and the square root difference increase image to be identified into the mineral identification model to obtain the mineral identification result output by the mineral identification model, including: Input the microscopic tomography image to be identified, the square difference increase image to be identified, and the square root difference increase image to be identified into the mineral identification model to obtain an output image marked with a one-hot encoding output by the mineral identification model; The mineral identification result is determined based on the output image labeled with one-hot encoding.

8. A mineral identification device, characterized in that: include: A sample determination module, used to determine a sample microscopic tomography image and a scanning electron microscope mineral detection image matching the sample microscopic tomography image; A model training module is used to train a U-Net model according to the sample microscopic tomography image and the scanning electron microscope mineral detection image to obtain a mineral recognition model; The mineral identification module is used to input the microscopic tomography image to be identified into the mineral identification model to obtain the mineral identification result output by the mineral identification model.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the mineral identification method as described in any one of claims 1-7 is implemented.

10. A storage medium storing computer executable instructions, characterized in that: The computer executable instructions are used to perform the mineral identification method as described in any one of claims 1-7 when executed by a computer processor.

Images