An automatic identification method of drilling rock debris lithology
By combining deep learning models and the KNN algorithm, the lithology of rock fragments can be automatically identified, solving the problems of low efficiency and accuracy in existing technologies and achieving efficient and accurate lithology identification.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA PETROLEUM & CHEMICAL CORP
- Filing Date
- 2025-01-02
- Publication Date
- 2026-07-03
AI Technical Summary
Existing methods for identifying rock cuttings lithology rely on human experience, resulting in low efficiency and accuracy, and cumbersome processes, which cannot meet the actual needs of rock cuttings logging.
By combining deep learning models and the KNN algorithm, an automatic identification model for rock debris lithology is established through image preprocessing, feature extraction, and classification labeling, thereby reducing the influence of subjective human factors.
It improves the efficiency and accuracy of lithology identification, simplifies the process, reduces costs, and enhances the model's generalization ability in complex scenarios.
Smart Images
Figure CN122336720A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of petroleum geological logging technology, and in particular to an automatic identification method for drilling cuttings lithology. Background Technology
[0002] Rock cuttings lithology identification is a key step in oil and gas exploration and development. By analyzing and identifying information such as the size, shape, density, and color of rock cuttings returned to the surface with drilling fluid during the drilling process, we can understand the geological structure of the target area and further provide accurate compositional analysis and lithology identification results for oil exploration.
[0003] Existing methods for identifying rock cuttings lithology, such as hand specimen identification, thin section identification, elemental analysis, and mineral analysis, typically rely on human experience, are susceptible to subjective factors, and involve cumbersome, time-consuming, costly, and low-efficiency and accuracy identification processes. With the development of logging technology, the number of rock cuttings logging images has increased dramatically, rendering traditional manual methods for identifying rock cuttings inadequate for practical needs. Summary of the Invention
[0004] The purpose of this invention is to provide at least one method for automatic identification of drilling cuttings lithology, which can at least solve the technical problems of low identification efficiency and accuracy in the prior art, and at least improve the efficiency and accuracy of lithology identification.
[0005] To address the aforementioned technical problems, at least one embodiment of the present invention provides an automatic identification method for drilling cuttings lithology, wherein the method includes:
[0006] Acquire drilling cuttings images from different well areas;
[0007] The drilling cuttings images are preprocessed to obtain target images that meet preset requirements;
[0008] The target image is divided into a training set, a validation set, and a test set. A deep learning model is used to identify the target image and segment out multiple individual rock debris images.
[0009] Feature extraction is performed on the single rock debris image based on a neural network algorithm;
[0010] Based on the extracted features, the KNN algorithm is used to classify and identify rock types in the feature space in order to establish an automatic identification model for rock debris lithology.
[0011] The target rock fragments to be identified are identified using the aforementioned automatic rock fragment lithology identification model.
[0012] In some optional embodiments, the preprocessing of the drilling cuttings image to obtain a target image that meets preset requirements includes: denoising the drilling cuttings image using a median filter.
[0013] In some optional embodiments, preprocessing the drilling cuttings image to obtain a target image that meets preset requirements includes: adjusting the contrast of the drilling cuttings image.
[0014] In some optional embodiments, the preprocessing of the drilling cuttings image to obtain a target image that meets preset requirements includes: using the Laplacian edge detection method to find the second-order local maximum of the drilling cuttings image, and determining the edge of the drilling cuttings image based on the second-order local maximum.
[0015] In some optional embodiments, the step of identifying the target image based on a deep learning model and segmenting it into multiple individual rock debris images includes: using a Mask R-CNN deep learning model to identify the target image and segment it into multiple individual rock debris images.
[0016] In some optional embodiments, the feature extraction of a single rock debris image based on a neural network algorithm includes: using a YOLOv9 network as a feature extractor to extract features from each rock debris image.
[0017] At least one embodiment of the present invention also provides an automatic identification device for drilling cuttings lithology, comprising:
[0018] The image acquisition module is used to acquire drilling cuttings images collected from different well areas;
[0019] The image processing module is used to preprocess the drilling cuttings image to obtain a target image that meets preset requirements;
[0020] The image segmentation module is used to divide the target image into a training set, a validation set, and a test set, and to use a deep learning model to identify the target image and segment out multiple individual rock debris images from it.
[0021] The feature extraction module is used to extract features from the single rock cutting image based on a neural network algorithm;
[0022] The model building module is used to classify and identify rock types in the feature space based on the extracted features using the KNN algorithm, so as to establish an automatic identification model for rock debris lithology.
[0023] The lithology identification module is used to identify the lithology of the target drilling cuttings using the automatic lithology identification model.
[0024] At least one embodiment of the present invention also provides a computer device, comprising:
[0025] At least one processor; and,
[0026] A memory communicatively connected to the at least one processor; wherein,
[0027] The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the automatic identification method for drilling cuttings lithology as described above.
[0028] At least one embodiment of the present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, implements the automatic identification method for drilling cuttings lithology as described above.
[0029] At least one embodiment of the present invention also provides a computer program product, including a computer program, wherein, when executed by a processor, the computer program implements the steps of the automatic identification method for drilling cuttings lithology as described above.
[0030] The present invention provides an automatic identification method for drilling cuttings lithology, which has the following beneficial technical effects:
[0031] 1. This method uses a combination of multiple image processing techniques, which can appropriately improve the quality of rock cuttings images, provide a high-quality data foundation for the constructed rock cuttings lithology identification model, and enhance the model's generalization ability in complex practical application scenarios.
[0032] 2. This method combines the K-Nearest Neighbor (KNN) classification algorithm with a deep learning network to construct a rock debris lithology identification model. It retains the intuitiveness of the KNN algorithm on small datasets while also possessing the powerful ability of deep learning networks to process large and complex data.
[0033] 3. This method introduces the YOLOv9 network to automatically extract image features, avoiding excessive human intervention and effectively reducing the impact of subjective human factors on the recognition results, while preserving important feature information in the original image.
[0034] The automatic identification method for drilling cuttings lithology disclosed in this invention is simple to operate, highly efficient in identification, accurate in lithology identification, and has a wide range of applications. Attached Figure Description
[0035] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.
[0036] Figure 1 This is a flowchart of the steps of the automatic identification method for drilling cuttings lithology provided in the embodiments of the present invention. Detailed Implementation
[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the various embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details are presented in the embodiments of the present invention to facilitate a better understanding of the invention. However, the technical solutions claimed in the present invention can be implemented even without these technical details and various variations and modifications based on the following embodiments. The division of the following embodiments is for ease of description and should not constitute any limitation on the specific implementation of the present invention. The various embodiments can be combined with and referenced by each other without contradiction.
[0038] Machine learning, as an important tool in data mining, has been introduced into cuttings lithology identification applications in recent years due to its ability to accurately characterize the complex nonlinear relationships between different lithologies and their logging data. This approach overcomes the shortcomings of traditional identification methods and improves the objectivity of the results. Deep learning, as a new research direction in the field of machine learning, not only possesses the advantages of traditional machine learning but also, given sufficient data, its deep neural network structure can learn the complex features of the input data, combining and abstracting these features at different levels. Especially when dealing with complex lithologies, deep learning can achieve more efficient and accurate cuttings lithology identification. Therefore, this invention proposes a novel automatic identification method for drilling cuttings lithology. This method utilizes machine learning algorithms to construct a lithology identification model, preserving the intuitiveness of image data while also possessing the data processing capabilities required for handling large amounts of image data.
[0039] like Figure 1 As shown, the main core idea of this method is to use machine learning algorithms to identify drilling cuttings images collected from different well areas, extract features from the cuttings images, establish an automatic lithology identification model for the cuttings based on the extracted features, and use the automatic lithology identification model to identify the lithology of the drilling cuttings to be tested.
[0040] The implementation details of the above method are described in detail below through examples. The following content is only for the convenience of understanding the implementation details and is not necessary for implementing this solution.
[0041] Example 1
[0042] This embodiment provides an automatic method for identifying the lithology of drilling cuttings, which mainly includes the following steps:
[0043] Step 1: Acquire drilling cuttings images from different well areas, especially high-resolution images of the cuttings.
[0044] The target image is identified using a deep learning model, and multiple individual rock debris images are segmented from it. Features are extracted from the individual rock debris images based on a neural network algorithm. Based on the extracted features, the rock type is classified and labeled in the feature space using the KNN algorithm to establish an automatic rock debris lithology identification model. The target rock debris to be identified is then identified using the automatic rock debris lithology identification model.
[0045] Step 2: Preprocess the drilling cuttings images to obtain target images that meet the preset requirements.
[0046] In this embodiment, the following image enhancement techniques are selected to calibrate and enhance the image's brightness, sharpness, contrast, and other indicators that affect image quality.
[0047] First, the image is denoised.
[0048] Read the input image, calculate the value of the median filter based on the filter radius, and assign the median filter value to the grayscale value of the output image.
[0049] G(x,y)=medianG in (xi,yj)|i,j∈[-k,k] (1)
[0050] Where G(x,y) is the gray value of the output image, G in (xi,yj) represents the grayscale values of the input image, and k is the radius of the filter.
[0051] Then, adjust the image contrast.
[0052] When k is a positive number and greater than 1, it stretches the gray levels of the image, increasing the image contrast, that is, the image becomes brighter or the details in the dark areas are clearer; when k is less than 1, it compresses the gray levels of the image, reducing the image contrast, that is, the image becomes softer.
[0053] G(x,y)=G in (x,y)+k×(G max -G min (2)
[0054] Where G(x,y) is the gray value of the output image, G in (x,y) represents the grayscale values of the original input image, G max G min are the maximum and minimum gray values of the input image, respectively, and k is a contrast control parameter.
[0055] Finally, the edges of the image are extracted.
[0056] Preferably, the Laplacian edge detection method is used, which uses the Laplacian operator of the image to find the second-order local maxima of the image. These local maxima usually correspond to the edges, so the edges of the image can be determined based on the second-order local maxima.
[0057] Lap(f(x,y))=f xx (x,y)+f yy (x,y) (3)
[0058] Where f(x,y) is the brightness value of the image at point (x,y), f xx (x,y), f yy (x, y) are the second-order partial derivatives at that point.
[0059] Step 3: Based on the edges detected in Step 2, the Mask R-CNN deep learning model is further used to simultaneously identify and segment multiple individual rock debris images for each image, perform pixel-level segmentation on each rock debris, and assign a unique label to it.
[0060] Step 4: Use the YOLOv9 network as a feature extractor to extract features from a single rock cuttings image. This network, based on programmable gradient information, can be used for various models ranging from lightweight to large-scale, and can be used to obtain complete rock cuttings image features, avoiding information loss during data transmission.
[0061] Step 5: Based on the features extracted in Step 4, use the KNN algorithm to classify rock types in the feature space.
[0062] This invention provides a novel solution for identifying the lithology of drilling cuttings. It preprocesses the original cuttings images based on image processing technology, performs image segmentation using the deep learning network Mask R-CNN, further extracts key image features using the deep learning network YOLOv9, and finally establishes an automatic lithology identification model for drilling cuttings based on the kNN algorithm. This has practical application significance for improving the efficiency of lithology identification and accurately obtaining stratigraphic geological structures.
[0063] Example 2
[0064] The technical solution and beneficial effects of the present invention will be further illustrated below with a specific example. This example is based on experiments and implementation using data collected from different well areas. The implementation steps of the entire process are as follows:
[0065] (1) High-resolution images of 30 common rock fragments from four major rock types in different well areas were acquired. To verify the generalization ability of this technique, mixed rocks were specifically selected. The image quality, such as brightness and sharpness, was calibrated and enhanced according to the image enhancement techniques described below.
[0066] ① Image denoising. Read the input image, calculate the value of the median filter based on the filter radius, and assign the median filter value to the grayscale value of the output image.
[0067] G(x,y)=medianG in (xi,yj)|i,j∈[-k,k] (1)
[0068] Where G(x,y) is the gray value of the output image, G in (xi,yj) represents the grayscale values of the input image, and k is the radius of the filter.
[0069] ② Image contrast adjustment. When k is a positive number and greater than 1, it stretches the gray levels of the image, increasing the image contrast, that is, the image becomes brighter or the details in the dark areas are clearer; when k is less than 1, it compresses the gray levels of the image, decreasing the image contrast, that is, the image becomes softer.
[0070] G(x,y)=G in (x,y)+k×(G max -G min (2)
[0071] Where G(x,y) is the gray value of the output image, G in (x,y) represents the grayscale values of the original input image, G max G min are the maximum and minimum gray values of the input image, respectively, and k is a contrast control parameter.
[0072] ③ Image edge extraction. The Laplacian edge detection method is used, which utilizes the Laplacian operator of the image to find the second-order local maxima of the image. These local maxima usually correspond to edges.
[0073] Lap(f(x,y))=f xx (x,y)+f yy (x,y) (3)
[0074] Where f(x,y) is the brightness value of the image at point (x,y), f xx (x,y), f yy (x, y) are the second-order partial derivatives at that point.
[0075] (2) Divide the preprocessed image dataset into training set, validation set and test set in a ratio of 7:3:1.
[0076] (3) Based on the Mask R-CNN deep learning model, multiple individual rock debris images are simultaneously identified and segmented for each image. Each rock debris is segmented at the pixel level and assigned a unique label. In this embodiment, the Mask R-CNN model uses the SGD optimizer with an initial learning rate of 0.001 and a weight decay of 0.0002.
[0077] (4) Feature extraction of a single rock debris image is performed based on the YOLOv9 network. In the example, the total number of training times of the model is set to 500 epochs, the initial learning rate is 0.01, and the weight decay is 0.0005.
[0078] (5) Based on the features extracted above, the KNN algorithm is used to classify rock types in the feature space. Specifically, the following steps are included:
[0079] Calculate the test data pixel x(A) x B x C x ,...,Z x ) and training data pixel points y(A y B y C y ,...,Z y The distance between them;
[0080]
[0081] Where A, B, C, ..., Z are feature points, and dis is the distance from the test data to the training data;
[0082] The training samples are sorted according to the calculated distances, and the k samples that are closest to the test data are found.
[0083] Select the k points with the smallest distance from the sorted samples;
[0084] Calculate the frequency of occurrence of each category in these k samples;
[0085] The category that appears most frequently is selected as the predicted category for the test data.
[0086] (5) Precision and recall are used as indicators to evaluate model performance.
[0087] Finally, experiments verified the feasibility and effectiveness of the invention, demonstrating that intelligent identification of rock debris lithology was achieved based on image processing technology and deep learning networks.
[0088] Example 3
[0089] Another embodiment of the present invention relates to an automatic identification device for drilling cuttings lithology, comprising:
[0090] The image acquisition module is used to acquire drilling cuttings images collected from different well areas;
[0091] The image processing module is used to preprocess the drilling cuttings image to obtain a target image that meets preset requirements;
[0092] The image segmentation module is used to divide the target image into a training set, a validation set, and a test set, and to use a deep learning model to identify the target image and segment out multiple individual rock debris images from it.
[0093] The feature extraction module is used to extract features from the single rock cutting image based on a neural network algorithm;
[0094] The model building module is used to classify and identify rock types in the feature space based on the extracted features using the KNN algorithm, so as to establish an automatic identification model for rock debris lithology.
[0095] The lithology identification module is used to identify the lithology of the target drilling cuttings using the automatic lithology identification model.
[0096] Example 4
[0097] Another embodiment of the present invention relates to a computer device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the automatic identification method for drilling cuttings lithology in the above embodiments.
[0098] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.
[0099] The processor manages the bus and handles general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory, on the other hand, is used to store data used by the processor during operation.
[0100] Example 5
[0101] Another embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the automatic identification method for drilling cuttings lithology described in the above embodiments.
[0102] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0103] Example 6
[0104] Another embodiment of the present invention relates to a computer program product, including a computer program that, when executed by a processor, implements the steps of the automatic identification method for drilling cuttings lithology described in the above embodiments.
[0105] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing the present invention, and in practical applications, various changes in form and detail may be made without departing from the spirit and scope of the present invention.
Claims
1. An automatic identification method for drilling cuttings lithology, characterized in that, include: Acquire drilling cuttings images from different well areas; The drilling cuttings images are preprocessed to obtain target images that meet preset requirements; The target image is divided into a training set, a validation set, and a test set. A deep learning model is used to identify the target image and segment out multiple individual rock debris images. Feature extraction is performed on the single rock debris image based on a neural network algorithm; Based on the extracted features, the KNN algorithm is used to classify and identify rock types in the feature space in order to establish an automatic identification model for rock debris lithology. The target rock fragments to be identified are identified using the aforementioned automatic rock fragment lithology identification model.
2. The automatic identification method for drilling cuttings lithology according to claim 1, characterized in that, The preprocessing of the drilling cuttings image to obtain a target image that meets preset requirements includes: The drilling cuttings image was denoised using a median filter.
3. The automatic identification method for drilling cuttings lithology according to claim 2, characterized in that, The preprocessing of the drilling cuttings image to obtain a target image that meets preset requirements includes: The contrast of the drilling cuttings image is adjusted.
4. The automatic identification method for drilling cuttings lithology according to claim 3, characterized in that, The preprocessing of the drilling cuttings image to obtain a target image that meets preset requirements includes: The Laplacian edge detection method is used to find the second-order local maxima of the drilling cuttings image, and the edges of the drilling cuttings image are determined based on the second-order local maxima.
5. The automatic identification method for drilling cuttings lithology according to claim 1, characterized in that, The process of identifying the target image based on a deep learning model and segmenting it into multiple individual rock debris images includes: The target image is identified using the Mask R-CNN deep learning model, and multiple individual rock debris images are segmented from it.
6. The automatic identification method for drilling cuttings lithology according to claim 1, characterized in that, The feature extraction of a single rock cuttings image based on a neural network algorithm includes: The YOLOv9 network was used as a feature extractor to extract features from each rock debris image.
7. An automatic identification device for drilling cuttings lithology, characterized in that, include: The image acquisition module is used to acquire drilling cuttings images collected from different well areas; The image processing module is used to preprocess the drilling cuttings image to obtain a target image that meets preset requirements; The image segmentation module is used to divide the target image into a training set, a validation set, and a test set, and to use a deep learning model to identify the target image and segment out multiple individual rock debris images from it. The feature extraction module is used to extract features from the single rock cutting image based on a neural network algorithm; The model building module is used to classify and identify rock types in the feature space based on the extracted features using the KNN algorithm, so as to establish an automatic identification model for rock debris lithology. The lithology identification module is used to identify the lithology of the target drilling cuttings using the automatic lithology identification model.
8. A computer device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the automatic identification method for drilling cuttings lithology as described in any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the automatic identification method for drilling cuttings lithology as described in any one of claims 1 to 6.
10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the automatic identification method for drilling cuttings lithology as described in any one of claims 1 to 6.