Rock debris identification method and device, electronic equipment and computer storage medium

Abstract

The embodiment of the invention discloses a rock debris recognition method and device, electronic equipment and a computer storage medium. The rock debris recognition method and device are used for improving rock debris recognition efficiency. The method provided by the embodiment of the invention comprises the following steps: acquiring terahertz spectrum data of a target rock debris sample through a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data comprises an absorption spectrum, a transmission spectrum and a refractive index; inputting the terahertz spectrum data into a pre-trained rock debris classification model to obtain the probability of the target rock debris sample output by the rock debris classification model on each preset category; and determining a classification result according to the probability of the target rock debris sample on each preset category.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of rock cuttings identification technology, and in particular to a rock cuttings identification method, device, electronic device, and computer storage medium. Background Art

[0002] The classification and identification of rock chip samples is a fundamental and critical task in geological exploration and lithologic analysis. Existing techniques rely primarily on manual identification, typically performed by experienced geologists through visual observation of macroscopic features such as color, grain structure, density, and mineralogy. While this method demonstrates some accuracy for certain common rock chip types, it is susceptible to subjective influence and can result in significant errors when rock chip samples exhibit complex morphology, similar composition, or are severely weathered. Furthermore, manual identification is inefficient, making it difficult to meet the demands of large-scale, rapid analysis.

[0003] To overcome the uncertainties of manual identification, some studies have begun to explore the use of spectral analysis, particularly infrared or near-infrared spectroscopy, to determine the physical properties of rock cuttings. However, these methods are limited in penetration, resolution, and ability to identify complex structures, making it difficult to accurately capture the microscopic electromagnetic response characteristics within rock cuttings. Furthermore, most existing spectral applications still rely on manual interpretation of spectra by experts, failing to achieve truly intelligent classification and thus exhibiting significant deficiencies in efficiency and consistency. Summary of the Invention

[0004] The embodiments of the present application provide a rock cuttings identification method, device, electronic device, and computer storage medium, which can improve the efficiency of rock cuttings identification.

[0005] A first aspect of an embodiment of the present application provides a rock debris identification method, comprising:

[0006] Acquiring terahertz spectrum data of the target rock cuttings sample by a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes an absorption spectrum, a transmission spectrum, and a refractive index;

[0007] Inputting the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample in each preset category output by the rock chip classification model;

[0008] The classification result is determined according to the probability of the target rock cuttings sample in each preset category.

[0009] Optionally, before inputting the terahertz spectral data into a pre-trained rock debris classification model, the method further includes:

[0010] constructing an initial classification model, wherein the initial classification model is an untrained rock cuttings classification model;

[0011] The initial classification model is trained using standard rock chip samples, and after the training is completed, a trained rock chip classification model is obtained; wherein the standard rock chip samples are rock chip samples marked with categories, and the trained rock chip classification model includes category information of all the standard rock chip samples.

[0012] Optionally, the constructing of the initial classification model includes:

[0013] The number of samples from which standard rock cuttings samples were obtained;

[0014] determining a model category based on the sample size;

[0015] An initial classification model is constructed according to the model category.

[0016] Optionally, determining the model category according to the sample size includes:

[0017] Determining whether the sample size is greater than or equal to a preset threshold;

[0018] If yes, the model category is determined to be a convolutional neural network; if no, the model category is determined to be a support vector machine.

[0019] Optionally, when the model type is a support vector machine, inputting the terahertz spectral data into a pre-trained rock debris classification model includes:

[0020] Extracting terahertz spectrum features according to the terahertz spectrum data, wherein the terahertz spectrum features include absorption peak change, transmission spectrum phase delay and refractive index change;

[0021] The terahertz spectral features are input into a pre-trained rock debris classification model.

[0022] Optionally, acquiring terahertz spectral data of the target rock cuttings sample by a terahertz time-domain spectroscopy system includes:

[0023] The target rock cuttings sample is scanned in the range of 0.1 THz to 3 THz by a terahertz time-domain spectroscopy system to obtain terahertz spectrum data of the target rock cuttings sample.

[0024] Optionally, the target rock cuttings sample is spread flat in a target tray, the thickness of the spread target rock cuttings sample is less than or equal to 3 mm, and the particle size of each particle in the target rock cuttings sample ranges from 0.5 mm to 2 mm.

[0025] A second aspect of an embodiment of the present application provides a rock debris identification device, comprising:

[0026] an acquisition unit, configured to acquire terahertz spectrum data of the target rock cuttings sample through a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes an absorption spectrum, a transmission spectrum, and a refractive index;

[0027] An input unit, configured to input the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample in each preset category output by the rock chip classification model;

[0028] The determination unit is configured to determine a classification result according to the probability of the target rock cuttings sample in each preset category.

[0029] Optionally, the rock cuttings identification device further includes:

[0030] A construction unit, configured to construct an initial classification model, wherein the initial classification model is an untrained rock debris classification model;

[0031] The training unit is used to train the initial classification model using standard rock chip samples, and obtain a trained rock chip classification model after the training is completed; wherein the standard rock chip samples are rock chip samples marked with categories, and the trained rock chip classification model includes category information of all the standard rock chip samples.

[0032] Optionally, the construction unit includes:

[0033] an acquisition module for acquiring the sample quantity of a standard rock cutting sample;

[0034] A determination module, configured to determine a model category according to the number of samples;

[0035] A construction module is used to construct an initial classification model according to the model category.

[0036] Optionally, the determining module is specifically configured to:

[0037] Determining whether the sample size is greater than or equal to a preset threshold;

[0038] If yes, the model category is determined to be a convolutional neural network; if no, the model category is determined to be a support vector machine.

[0039] Optionally, when the model type is a support vector machine, the input unit is specifically used to:

[0040] Extracting terahertz spectrum features according to the terahertz spectrum data, wherein the terahertz spectrum features include absorption peak change, transmission spectrum phase delay and refractive index change;

[0041] The terahertz spectral features are input into a pre-trained rock debris classification model.

[0042] Optionally, the acquiring unit is specifically configured to:

[0043] The target rock cuttings sample is scanned in the range of 0.1 THz to 3 THz by a terahertz time-domain spectroscopy system to obtain terahertz spectrum data of the target rock cuttings sample.

[0044] Optionally, the target rock cuttings sample is spread flat in a target tray, the thickness of the spread target rock cuttings sample is less than or equal to 3 mm, and the particle size of each particle in the target rock cuttings sample ranges from 0.5 mm to 2 mm.

[0045] A third aspect of the embodiments of the present application provides an electronic device, including:

[0046] processor, memory, input and output units, and buses;

[0047] The processor is connected to the memory, the input and output unit, and the bus;

[0048] A program is stored in the memory, and the processor calls the program to execute the method in the first aspect and any possible implementation of the first aspect.

[0049] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium, on which a program is stored. When the program is executed on a computer, the computer executes the method in the first aspect and any possible implementation of the first aspect.

[0050] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0051] In an embodiment of the present application, terahertz spectral data including absorption spectrum, transmission spectrum and refractive index are first obtained through a terahertz time-domain spectroscopy system, and the terahertz spectral data is input into a pre-trained rock chip classification model, so that the model can judge the probability distribution of the target rock chip samples in each preset category based on different spectral feature dimensions, and then automatically output the classification results based on the maximum probability, thereby realizing rapid, non-destructive and high-precision identification of the rock chip sample type, and improving the efficiency of rock chip identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 This is a flow chart of an embodiment of a rock cuttings identification method in an embodiment of the present application;

[0053] Figure 2 This is a flow chart of another embodiment of the rock cuttings identification method in the embodiments of the present application;

[0054] Figure 3 This is a schematic structural diagram of an embodiment of a rock debris identification device in an embodiment of the present application;

[0055] Figure 4 This is a structural diagram of an embodiment of an electronic device in the embodiments of the present application. DETAILED DESCRIPTION

[0056] The embodiments of the present application provide a rock cuttings identification method, device, electronic device, and computer storage medium for improving the efficiency of rock cuttings identification.

[0057] The method of the present application can be applied to a server, a terminal or other device with logic processing capabilities, and the present application does not limit this. For the convenience of description, the following description is based on an example in which the execution subject is a server.

[0058] The embodiments of the present application will be described below with reference to the accompanying drawings.

[0059] See also Figure 1 In one embodiment of the present application, a method for identifying rock debris includes:

[0060] 101. Acquire terahertz spectrum data of the target rock cuttings sample through a terahertz time-domain spectroscopy system. The terahertz spectrum data includes absorption spectrum, transmission spectrum, and refractive index;

[0061] The server controls the terahertz time-domain spectroscopy system to collect the spectral response of the target rock cuttings sample in the terahertz band. By modulating and demodulating the generated terahertz pulse signal and allowing it to penetrate the target rock cuttings sample, the sample's absorption capacity (absorption spectrum) and transmission degree (transmission spectrum) for terahertz waves of different frequencies are determined, as well as the refractive index distribution derived from the corresponding changes in the complex dielectric function. The server synchronizes, collects, digitizes, and preprocesses this raw measurement data to form structured terahertz spectral data for subsequent analysis.

[0062] 102. Inputting the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample being in each preset category output by the rock chip classification model;

[0063] The server calls a pre-trained rock chip classification model and uses the multi-dimensional spectral data such as the absorption spectrum, transmission spectrum, and refractive index obtained in step 101 as model input. The multi-layer neural network constructed in the model extracts the rock chip spectral characteristics and compares and calculates the distribution pattern of the training set samples. Finally, the model outputs the classification probability distribution of the target rock chip sample in all preset rock chip categories (such as sandstone, shale, limestone, etc.).

[0064] 103. Determine the classification result according to the probability of the target rock cuttings sample being in each preset category.

[0065] The server analyzes the probability results of each category output by the rock chip classification model and selects the category corresponding to the maximum probability as the final classification result of the target rock chip sample. The classification result will be used for geological identification or engineering decision support.

[0066] In this embodiment, the server first obtains multidimensional spectral data including absorption spectrum, transmission spectrum and refractive index through the terahertz time-domain spectroscopy system, and inputs it into the pre-trained rock chip classification model, so that the model can judge the probability distribution of the target rock chip samples in each preset category based on different spectral feature dimensions. The server then automatically outputs the classification results based on the maximum probability, thereby realizing rapid, non-destructive and high-precision identification of the rock chip sample type, thereby improving the efficiency of rock chip identification.

[0067] See also Figure 2 Another embodiment of the rock debris identification method in the embodiment of the present application includes:

[0068] 201. The number of samples obtained for standard rock cuttings samples;

[0069] The server accesses the rock chip sample database used for model training, retrieves standard rock chip samples with clear category labels, and counts the total number of these samples. Standard rock chip samples are those whose types have been confirmed by professional geologists and assigned unique category labels, making them representative and authoritative. This number not only determines the scale of data the model can learn from but also provides the necessary categorization foundation for the subsequent construction of the classification system.

[0070] 202. Determine the model category based on the sample size;

[0071] The server determines the number of standard rock cuttings samples obtained, and first compares the number with a preset threshold. This threshold is determined based on a large number of experiments and is used to distinguish the critical point for using a deep learning model or a traditional machine learning model. If the number of samples is greater than or equal to the threshold, it means that the sample size is sufficient to support the training of a complex model. At this time, the server determines that the model category is a convolutional neural network (CNN). Convolutional neural networks have powerful feature extraction capabilities and can automatically learn local patterns and differences between categories from two-dimensional or multi-dimensional inputs of terahertz spectra. They are suitable for large sample scenarios. If the number of samples is less than the threshold, it means that the training data is relatively limited. The server determines that the model category is a support vector machine (SVM). Support vector machines have good generalization capabilities under small sample conditions. They achieve category division by constructing an optimal hyperplane and are suitable for simple classification problems of spectral data. The choice of this model category directly affects the network structure and training strategy of subsequent model construction.

[0072] 203. Construct an initial classification model according to the model category;

[0073] Based on the number of rock fragment categories determined, the server constructs an initial classification model with corresponding output channels. This model is typically a multi-classification architecture that accepts multi-dimensional features as input and outputs a probability value corresponding to each category.

[0074] In one possible embodiment, when the determined model category is CNN, the model structure of the initial classification model constructed by the server may include an input layer, multiple hidden layers and a Softmax output layer, wherein the input layer is used to receive terahertz spectral feature data, the hidden layer is used to extract deep information, and the output layer performs probability prediction for each category. The loss function generally adopts a cross entropy loss function to adapt to multi-category discrimination tasks. It should be noted that the above description is only exemplary and does not limit the embodiment of the present application to only this embodiment.

[0075] In one possible embodiment, when the determined model category is CNN, the server can first set the kernel function type to radial basis function (RBF kernel) to introduce nonlinear mapping capabilities in the low-dimensional input feature space, and at the same time set the penalty parameter C to 1.0 to balance the training error and model complexity, and initialize the γ parameter to 1 / N (N is the feature dimension) to adapt to the distribution characteristics of the terahertz spectral features; then, the server constructs a multi-classification structure according to a one-vs-one strategy, that is, trains a two-classification SVM sub-model for every two categories, and finally combines all sub-models into a complete initial classifier framework to prepare for the subsequent training of standard rock cuttings samples. It should be noted that the above description is only exemplary and does not limit the embodiment of the present application to only this embodiment.

[0076] 204. Train the initial classification model using standard rock chip samples. After the training is completed, a trained rock chip classification model is obtained. The standard rock chip samples are rock chip samples labeled with categories, and the trained rock chip classification model includes category information of all standard rock chip samples.

[0077] The server uses standard rock cuttings samples with class labels as training data, feeds their corresponding terahertz spectral data into an initial classification model, and uses the true class of the samples as a supervisory signal to iteratively update the model parameters. During training, the model gradually learns the characteristic distributions of different rock cuttings categories in terms of terahertz absorption spectra, transmission spectra, and refractive index. Training continues until the model's performance on the validation set is stable and meets the set accuracy requirements, ultimately forming a mature classification model that can be used to identify unknown rock cuttings.

[0078] 205. Acquire terahertz spectrum data of the target rock cuttings sample through a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes absorption spectrum, transmission spectrum and refractive index;

[0079] The server invokes and controls a terahertz time-domain spectroscopy (THz-TDS) system to perform optical inspection on the rock chip sample to be identified. The system emits a terahertz pulse wave, which penetrates the rock chip sample and records its energy changes and propagation behavior at different frequencies, generating a set of frequency-domain spectral data. The absorption spectrum reflects the material's ability to absorb electromagnetic waves, the transmission spectrum describes the intensity changes of the wave passing through the sample, and the refractive index indicates the sample's influence on the wave's propagation speed. Together, these parameters constitute the sample's optical fingerprint.

[0080] Specifically, before scanning with the terahertz time-domain spectroscopy system, a target rock chip sample is first obtained. During this process, the sample surface is kept clean and as pristine as possible to avoid chemical contamination due to manual handling or environmental factors. Basic information such as sample size, particle size, and humidity is recorded to ensure consistent sample conditions during subsequent spectral analysis. To control the particle size range, standard screening methods are used to screen the cuttings, keeping the sample size between 0.5 mm and 2 mm, thereby improving signal stability and repeatability during terahertz detection. After initial screening, the cuttings are pre-treated, including using clean air purging or ultrasonic cleaning to remove dust, sediment, and other surface deposits. If high moisture content is detected, the sample is further dehydrated using a low-temperature vacuum drying process to prevent strong absorption interference from moisture on the terahertz signal. The cleaned and dried cuttings are evenly spread on a specially designed target tray, with a thickness of no more than 3 mm to minimize the effects of multiple scattering of the terahertz wave within the sample. Subsequently, the server controls the terahertz time-domain spectroscopy system to scan the rock cuttings samples in the tray. The scanning frequency range covers 0.1THz to 3THz, and comprehensively collects the absorption spectrum, transmission spectrum and refractive index data of the rock cuttings samples in this frequency band, ensuring that rich information characterizing the electromagnetic properties of the rock cuttings materials is captured, and the collected terahertz spectral data is stored for subsequent analysis and classification model use.

[0081] 206. Inputting the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample being in each preset category output by the rock chip classification model;

[0082] The server preprocesses the acquired terahertz absorption spectrum, transmission spectrum, and refractive index data, converting them into model-acceptable feature vectors and feeding them into the trained rock debris classification model. The model then performs a matching operation on the input data based on the learned spectral signature patterns of each rock debris type and outputs a probability vector representing the likelihood that the target sample belongs to each rock debris category. This output format provides interpretability and quantification for the classification process.

[0083] The server first executes the corresponding data processing flow based on the selected model category. It should be noted that when inputting terahertz spectral data into the pre-trained rock debris classification model, data preprocessing is required, including noise reduction, outlier detection and elimination, etc.

[0084] After data preprocessing, when the model category is CNN, the server can directly input the terahertz spectral data into the rock debris classification model. When the model category is SVM, the server does not directly use the original spectral data for judgment, but instead performs feature extraction operations on the terahertz spectral data to extract key spectral features that can significantly distinguish rock debris categories. Specifically, the server extracts the variation characteristics of the absorption peak from the absorption spectrum and analyzes the amplitude and position differences of energy absorption at each frequency; extracts phase delay information from the transmission spectrum to evaluate the time lapse of the terahertz wave as it propagates within the rock debris; and calculates its variation trend from the refractive index curve to reflect the influence of the material on the wave velocity at different frequencies. These features constitute a set of highly representative and relatively low-dimensional spectral description vectors, which are suitable for support vector machine processing. The server passes this spectral feature vector as input to the trained SVM rock debris classification model. The model uses the boundary decision function to perform multi-category discrimination on the input sample and outputs the discrimination results of the target sample on each preset rock debris category, expressed in the form of probability. If the model category is a convolutional neural network, the server directly uses the normalized complete terahertz absorption spectrum, transmission spectrum, and refractive index sequence as input, and automatically completes feature extraction and classification prediction with the help of the network.

[0085] 207. Determine the classification result according to the probability of the target rock cuttings sample in each preset category.

[0086] The server sorts the probability results returned by the model, identifies the category with the highest probability, and records it as the final classification result for the target rock cuttings sample. This result represents the model's most likely classification based on the degree of match between the current spectral characteristics and the various types of rock cuttings used in training. It can serve as a basis for subsequent geological research, lithology analysis, and engineering material selection.

[0087] In this embodiment, the server constructs a classification model covering all rock chip types based on standard rock chip samples. By collecting the terahertz absorption spectrum, transmission spectrum, and refractive index of the target rock chip samples and inputting them into the model for automatic judgment, the classification process can be completed without relying on manual experience. Because the response of terahertz spectra in different lithologic structures is highly variable, combined with the precise output of the classification model, it can achieve efficient and non-destructive identification of unknown rock chips, effectively improving the intelligent level of rock chip sample classification and geological exploration efficiency.

[0088] See also Figure 3 In one embodiment of the present application, a rock debris identification device includes:

[0089] An acquisition unit 301 is configured to acquire terahertz spectrum data of a target rock cutting sample through a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes an absorption spectrum, a transmission spectrum, and a refractive index;

[0090] An input unit 302 is used to input the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample in each preset category output by the rock chip classification model;

[0091] The determination unit 303 is configured to determine a classification result according to the probability of the target rock cuttings sample being in each preset category.

[0092] In this embodiment, the acquisition unit 301 first acquires terahertz spectral data including absorption spectrum, transmission spectrum and refractive index through the terahertz time-domain spectroscopy system. The input unit 302 inputs the acquired terahertz spectral data into the pre-trained rock chip classification model, so that the model can judge the probability distribution of the target rock chip samples in each preset category in combination with different spectral feature dimensions. The determination unit 303 then automatically outputs the classification result based on the maximum probability value, thereby realizing rapid, non-destructive and high-precision identification of the rock chip sample type, thereby improving the efficiency of rock chip identification.

[0093] Optionally, the rock cuttings identification device further includes:

[0094] A construction unit, used for constructing an initial classification model, where the initial classification model is an untrained rock fragment classification model;

[0095] The training unit is used to train the initial classification model using standard rock chip samples. After the training is completed, a trained rock chip classification model is obtained. The standard rock chip samples are rock chip samples marked with categories, and the trained rock chip classification model includes category information of all standard rock chip samples.

[0096] Optionally, the building blocks include:

[0097] an acquisition module for acquiring the sample quantity of a standard rock cutting sample;

[0098] A determination module is used to determine the model category based on the number of samples;

[0099] The building module is used to build an initial classification model based on the model category.

[0100] Optionally, determine whether the module is specifically used for:

[0101] Determine whether the sample size is greater than or equal to a preset threshold;

[0102] If yes, the model category is determined to be a convolutional neural network; if no, the model category is determined to be a support vector machine.

[0103] Optionally, when the model type is a support vector machine, the input unit 302 is specifically used to:

[0104] Extracting terahertz spectrum features based on terahertz spectrum data, where the terahertz spectrum features include absorption peak change, transmission spectrum phase delay, and refractive index change;

[0105] The terahertz spectral features are fed into a pre-trained rock fragments classification model.

[0106] Optionally, the acquiring unit 301 is specifically configured to:

[0107] The target rock cuttings sample is scanned in the range of 0.1 THz to 3 THz by a terahertz time-domain spectroscopy system to obtain terahertz spectrum data of the target rock cuttings sample.

[0108] Optionally, the target rock cuttings sample is spread flat in a target tray, the thickness of the spread target rock cuttings sample is less than or equal to 3 mm, and the particle size of each particle in the target rock cuttings sample ranges from 0.5 mm to 2 mm.

[0109] In this implementation, the functions of each unit and module are the same as those mentioned above. Figures 1 to 2 The steps in the illustrated embodiment correspond to each other and will not be repeated here.

[0110] See also Figure 4 In the embodiments of the present application, an electronic device includes:

[0111] Processor 401, memory 402, input and output unit 403 and bus 404;

[0112] The processor 401 is connected to the memory 402, the input and output unit 403 and the bus 404;

[0113] The memory 402 stores a program, and the processor 401 calls the program to execute Figures 1 to 2 Steps in the illustrated embodiment.

[0114] In this embodiment, the function of the processor 401 is the same as that of the aforementioned Figures 1 to 2 The steps in the illustrated embodiment correspond to each other and will not be repeated here.

[0115] The embodiment of the present application further provides a computer-readable storage medium having a program stored thereon, which, when executed on a computer, causes the computer to execute the aforementioned Figures 1 to 2 A method in any possible embodiment.

[0116] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0117] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0118] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0119] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0120] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, read-only memory), random access memory (RAM, random access memory), disk or optical disk, and other media that can store program code.

Claims

1. A rock cuttings identification method, characterized in that: include: Acquiring terahertz spectrum data of the target rock cuttings sample by a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes an absorption spectrum, a transmission spectrum, and a refractive index; Inputting the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample in each preset category output by the rock chip classification model; The classification result is determined according to the probability of the target rock cuttings sample in each preset category.

2. The method according to claim 1, characterized in that Before inputting the terahertz spectral data into a pre-trained rock debris classification model, the method further includes: constructing an initial classification model, wherein the initial classification model is an untrained rock cuttings classification model; The initial classification model is trained using standard rock chip samples, and after the training is completed, a trained rock chip classification model is obtained; wherein the standard rock chip samples are rock chip samples marked with categories, and the trained rock chip classification model includes category information of all the standard rock chip samples.

3. The method according to claim 2, characterized in that The constructing of the initial classification model comprises: The number of samples from which standard rock cuttings samples were obtained; determining a model category based on the sample size; An initial classification model is constructed according to the model category.

4. The method according to claim 3, characterized in that Determining the model category according to the sample quantity includes: Determining whether the sample size is greater than or equal to a preset threshold; If yes, the model category is determined to be a convolutional neural network; if no, the model category is determined to be a support vector machine.

5. The method according to claim 4, characterized in that When the model type is a support vector machine, inputting the terahertz spectrum data into a pre-trained rock debris classification model includes: Extracting terahertz spectrum features according to the terahertz spectrum data, wherein the terahertz spectrum features include absorption peak change, transmission spectrum phase delay and refractive index change; The terahertz spectral features are input into a pre-trained rock debris classification model.

6. The method according to any one of claims 1 to 5, characterized in that The method of obtaining terahertz spectral data of the target rock cuttings sample by using a terahertz time-domain spectroscopy system includes: The target rock cuttings sample is scanned in the range of 0.1 THz to 3 THz by a terahertz time-domain spectroscopy system to obtain terahertz spectrum data of the target rock cuttings sample.

7. The method according to claim 6, characterized in that The target rock cuttings sample is spread in a target tray, the thickness of the spread target rock cuttings sample is less than or equal to 3 mm, and the particle size of each particle in the target rock cuttings sample ranges from 0.5 mm to 2 mm.

8. A rock chip identification device, characterized in that: include: an acquisition unit, configured to acquire terahertz spectrum data of the target rock cuttings sample through a terahertz time-domain spectroscopy system, wherein the terahertz spectrum data includes an absorption spectrum, a transmission spectrum, and a refractive index; An input unit, configured to input the terahertz spectrum data into a pre-trained rock chip classification model to obtain the probability of the target rock chip sample in each preset category output by the rock chip classification model; The determination unit is configured to determine a classification result according to the probability of the target rock cuttings sample in each preset category.

9. An electronic device, characterized in that: include: processor, memory, input and output units, and buses; The processor is connected to the memory, the input and output unit, and the bus; A program is stored in the memory, and the processor calls the program to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, and when the program is executed on a computer, the computer is caused to execute the method according to any one of claims 1 to 7.

Images