Strata fracture identification method, device, equipment, medium and program product

Abstract

Embodiments of the present application provide a stratum fracture identification method, device, equipment, medium and program product. The method comprises: obtaining an imaging logging image corresponding to a target stratum; obtaining a plurality of input sub-images corresponding to the imaging logging image; generating output sub-images corresponding to the plurality of input sub-images based on the plurality of input sub-images corresponding to the imaging logging image and a stratum fracture identification model trained to convergence, to complete fracture identification; the output sub-images are generated after the stratum fracture identification model extracts local features and global features from the input sub-images; and the target stratum in the present application includes deep and ultra-deep complex tight sandstone oil and gas reservoir strata, so that the output sub-images of local features and global features can be fused in the face of complex fractures and harsh conditions, and the fractures can be more accurately identified.

Description

Technical Field

[0001] This application relates to the field of oil and gas reservoir development technology, and in particular to a method, apparatus, equipment, medium and program product for identifying formation fractures. Background Technology

[0002] In recent years, the global oil and gas reservoir exploration and development field has entered a period of rapid development. Deep and ultra-deep tight sandstone oil and gas reservoirs have attracted much attention in the energy sector, demonstrating their excellent exploration potential. However, the presence of fractures in the formation can disrupt the integrity of the caprock, leading to oil and gas loss. Therefore, studying the identification and development characteristics of fractures in the formation can provide reliable geological and engineering basis for the exploration and development of oil and gas in tight sandstone reservoirs.

[0003] In existing technologies, oil-based drilling fluid is added to the formation wellbore, filling the formation fractures. The oil-based drilling fluid and the wellbore skeleton form a resistivity difference, and the formation fractures are identified through resistivity difference imaging.

[0004] However, oil-based drilling fluids actually reduce the conductivity of the fluid inside the wellbore, resulting in poor conductivity and increased resistivity. Consequently, the resistivity difference between the oil-based drilling fluid and the wellbore skeleton within the formation fractures becomes smaller, leading to poor micro-resistivity imaging and reduced formation fracture identification capabilities, making it impossible to accurately identify formation fractures. Summary of the Invention

[0005] This application provides a method, apparatus, device, medium, and program product for identifying formation fractures, so as to achieve the effect of accurately identifying formation fractures.

[0006] In a first aspect, embodiments of this application provide a method for identifying formation fractures, including:

[0007] Acquire imaging logging images corresponding to the target formation; at least one imaging logging image is required; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging images include oil-based drilling fluid subsurface micro-resistivity imaging logging images.

[0008] For each imaging logging image, multiple corresponding input sub-images are obtained based on the imaging logging image;

[0009] Based on multiple input sub-images corresponding to imaging logging images, and a formation fracture identification model that has been trained to convergence, multiple output sub-images corresponding to the input sub-images are generated to complete fracture identification. The output sub-images include the fracture conditions corresponding to the input sub-images. The output sub-images are generated by the formation fracture identification model after extracting local and global features from the input sub-images. The formation fracture identification model is generated by training on a training set. The training set includes at least one training image and its corresponding label image. The label image includes the fractures present in the training image. The label image is obtained by labeling the training image using a preset labeling software and then binarizing it.

[0010] For each imaging logging image, multiple output sub-images corresponding to the imaging logging image are stitched together to obtain the formation fracture image corresponding to the imaging logging image.

[0011] Optionally, acquire imaging logging images corresponding to the target formation, including:

[0012] Acquire electrical imaging data corresponding to the target strata;

[0013] Imaging logging images are obtained based on electrical imaging data.

[0014] Optionally, multiple input sub-images can be obtained based on imaging logging images, including:

[0015] Fill the blank areas to obtain the filled image;

[0016] The padded image is then segmented to obtain multiple input sub-images.

[0017] Optionally, the padded image can be segmented to obtain multiple input sub-images, including:

[0018] Keeping the cutting width consistent with the image capture width, the filled image is vertically overlapped and cut within the image capture height according to the preset step size and preset cutting height to obtain multiple input sub-images; the overlap height is the difference between the preset cutting height and the preset step size.

[0019] Optionally, based on multiple input sub-images corresponding to the imaging logging images and a formation fracture identification model that has been trained to convergence, multiple output sub-images corresponding to the input sub-images are generated to complete fracture identification, including:

[0020] Multiple input sub-images corresponding to the imaging logging images are batch input into the formation fracture identification model;

[0021] A formation fracture identification model is used to generate low-resolution feature maps corresponding to multiple input sub-images;

[0022] A formation fracture identification model is used to generate multiple output sub-images corresponding to each input sub-image based on each low-resolution feature map to complete fracture identification; the output sub-images are binarized images.

[0023] Optionally, a formation fracture identification model is used to generate low-resolution feature maps corresponding to multiple input sub-images, including:

[0024] The first convolutional layer in the geological fracture identification model is used to extract local features from each input sub-image to obtain a local feature map corresponding to each input sub-image; the local features include edge and corner features in the input sub-image; the first convolutional layer is located in the encoder;

[0025] The pooling layer in the formation fracture identification model is used to downsample each local feature map to capture abstract features and generate corresponding low-resolution feature maps.

[0026] Optionally, a formation fracture identification model is used to generate multiple output sub-images corresponding to each input sub-image based on each low-resolution feature map to complete fracture identification, including:

[0027] The low-resolution feature map is upsampled using an upsampling layer in the formation fracture identification model in order to identify the upsampled feature map; the resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image.

[0028] The second convolutional layer in the geological fracture identification model is used to extract global features from each input sub-image in a hierarchical order; the second convolutional layer is located in the decoder; the global features include the appearance and shape features of the input sub-image; the second convolutional layer includes at least one;

[0029] Each global feature is fused into the corresponding upsampled feature map to generate a fused feature map for each input sub-image; the fused feature map includes local features and global features.

[0030] Each fused feature map is determined as the corresponding output sub-image to complete crack identification.

[0031] Secondly, embodiments of this application provide a formation fracture identification device, comprising:

[0032] The acquisition module is used to acquire imaging logging images corresponding to the target formation; at least one imaging logging image is required; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging images include oil-based drilling fluid subsurface micro-resistivity imaging logging images.

[0033] The acquisition module is used to obtain multiple corresponding input sub-images based on each imaging logging image;

[0034] The identification module is used to generate multiple output sub-images corresponding to the input sub-images based on the imaging logging images and a formation fracture identification model that has been trained to convergence, in order to complete fracture identification. The output sub-images include the fracture conditions corresponding to the input sub-images. The output sub-images are generated by the formation fracture identification model after extracting local and global features from the input sub-images. The formation fracture identification model is generated by training on a training set. The training set includes at least one training image and its corresponding label image. The label image includes the fractures present in the training image. The label image is obtained by labeling the training image using a preset labeling software and then binarizing it.

[0035] The stitching module is used to stitch together multiple output sub-images corresponding to each imaging logging image to obtain a formation fracture image corresponding to the imaging logging image.

[0036] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;

[0037] The memory stores the instructions that the computer executes;

[0038] The processor executes computer execution instructions stored in memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0039] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0040] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0041] This application provides a method, apparatus, device, medium, and program product for identifying formation fractures. First, multiple input sub-images are obtained from the acquired imaging logging images. Then, based on the multiple input sub-images and a converged formation fracture identification model, corresponding output sub-images are identified. The output sub-images include the corresponding fractures. Since the output sub-images in this application are generated by the formation fracture identification model after extracting local and global features from the corresponding input sub-images, the output sub-images integrate local and global features, making them more closely aligned with the actual situation and thus more accurate. Therefore, the fractures in the output sub-images are more accurate, reducing the possibility of missed or incorrect identification. Furthermore, the formation fracture identification model in this application is generated based on at least one training image and its corresponding label image. Since the formation fracture identification model can extract local and global features, the local and global features of the training image are also considered when training the model. The training set in this application includes at least one training image and its corresponding label image. The label image is obtained by labeling the training image using a preset labeling software and performing binarization processing. Therefore, the label image is more accurate, which in turn makes the formation fracture identification model more optimized. Thus, the formation fracture identification model is an improved model that can more accurately obtain the output sub-image. The target formation in this application includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations. Therefore, for complex formations, this application can still use imaging logging images to identify the fractures included therein. Moreover, since it can identify local and global features, it can identify the fractures more accurately. In addition, the imaging logging images in this application can include micro-resistivity imaging logging images under oil-based drilling fluid. Therefore, for oil-based drilling fluid, this application can also accurately identify the fractures in the target formation. Attached Figure Description

[0042] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0043] Figure 1 A schematic diagram illustrating the scenario for identifying formation fractures provided in this application;

[0044] Figure 2 This is a schematic diagram of a formation fracture identification method provided in Embodiment 1;

[0045] Figure 3 This is a schematic diagram of a formation fracture identification method provided in Example 3;

[0046] Figure 4 This is a schematic diagram of a formation fracture identification method provided in Example 6;

[0047] Figure 5 A schematic diagram of overlapping cutting provided in Embodiment Seven;

[0048] Figure 6 This is a complete schematic diagram of a formation fracture identification method provided in Example 7;

[0049] Figure 7 This is a schematic diagram of a formation fracture identification device provided in Embodiment 8;

[0050] Figure 8 A schematic diagram of the structure of the electronic device provided in this application.

[0051] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0052] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0053] In existing technologies, oil-based drilling fluid is added to the formation wellbore, filling the formation fractures. The oil-based drilling fluid and the wellbore skeleton create a resistivity difference, which is then used for imaging to identify formation fractures. However, the oil-based drilling fluid actually reduces the conductivity of the fluid within the wellbore, resulting in poor conductivity. This further reduces the resistivity difference between the oil-based drilling fluid and the wellbore skeleton within the formation fractures, leading to poor micro-resistivity imaging. The small resistivity difference can cause false fracture identification or even prevent fracture detection altogether, thus reducing the ability to identify formation fractures. To overcome the shortcomings of existing technologies, the inventors of this solution have conducted inventive research and designed a new approach. This solution provides a method for identifying formation fractures. To address the low accuracy of existing methods for identifying formation fractures, this solution obtains multiple input sub-images based on the acquired imaging logging images. Furthermore, it identifies output sub-images corresponding to these input sub-images using a pre-trained and converged formation fracture identification model. In this solution, the output sub-images are generated by the formation fracture identification model through the extraction of local and global features from the input sub-images, including the fractures corresponding to the input sub-images. Because this solution integrates local and global features, the output sub-images are more realistic, resulting in more accurate fracture identification. Moreover, due to the extraction of local features, the edges and corners of the fractures can be accurately represented. Additionally, the formation fracture identification model in this application is trained based on at least one training image and its corresponding label image. The formation fracture identification model can extract global and local features. Therefore, during model training, global and local features are also extracted from the training images, enabling the model to learn to extract multiple features. Furthermore, the training set in this application includes at least one training image and its corresponding label image. The label image is obtained by labeling the training image using pre-defined labeling software and then binarizing it. The label image includes the fractures present in the training image. Because the label image is obtained using pre-defined labeling software, it is more accurate, thus optimizing the formation fracture identification model. This is an improved fracture identification model. Therefore, using the above-mentioned formation fracture identification model can accurately identify the output sub-image, making the identification of formation fractures more accurate. In addition, the target formations in this scheme include deep and ultra-deep complex tight sandstone oil and gas reservoirs. Therefore, for complex formations, this application can also accurately identify fractures using the formation fracture identification model.

[0054] In other related technologies, well logging software is used to process and obtain fracture images. While these images can identify fractures, the identified fractures are one-dimensional (including fracture shape and length), lacking the intuitiveness of two-dimensional images. The output sub-images in this solution show two-dimensional fractures (including fracture shape, length, and width), making them more intuitive and accurate.

[0055] Figure 1 This application provides a schematic diagram of a scenario for identifying formation fractures, as shown below. Figure 1 As shown, the specific application scenario of this application includes electronic device 101.

[0056] Among them, electronic device 101 can be a computer or a server, and there is no restriction here.

[0057] The electronic device 101 includes a formation fracture identification model 102 that has been trained to convergence.

[0058] In this scenario, the electronic device 101 acquires the imaging logging image corresponding to the target formation, segments the imaging logging image, and obtains multiple corresponding input sub-images.

[0059] Furthermore, multiple input sub-images are input into the formation fracture recognition model 102, and local and global features are extracted to identify the output sub-images corresponding to each input sub-image.

[0060] Furthermore, multiple output sub-images are stitched together to obtain the formation fracture image corresponding to the imaging logging image.

[0061] Based on the above scenario, it is clear that existing methods for identifying cracks using resistivity differences cannot accurately determine crack patterns. This solution, however, includes both local and global features corresponding to the input sub-image in its output sub-image. By addressing multiple features, the output sub-image becomes more accurate. The extraction of local features allows for the identification of prominent and minute features within the cracks of the target formation in a real-world environment, thus accurately representing the true state of the cracks in the target formation. Local and global features can be accurately extracted from the crack's color and pattern, enabling accurate crack identification. Furthermore, the formation crack identification model in this solution is trained based on at least one training image and its corresponding label image. During training, both global and local features in the training image are extracted, allowing the model to learn this function. Therefore, the formation crack identification model in this solution is an improved crack identification model, enabling accurate crack identification.

[0062] It should be noted that when using the formation fracture identification model to extract features from the input sub-image in this application, local and global features are extracted based on the color, texture, and other feature information of the fractures in the input sub-image, thereby making the output sub-image more accurate.

[0063] It should be noted that the formation fracture identification method of this application can be applied to deep and ultra-deep complex tight sandstone oil and gas reservoirs under oil-based drilling fluids. The fractures in this application can be structural fractures and occurrence fractures, where occurrence fractures include near-horizontal fractures, low-angle oblique fractures, medium-angle oblique fractures, and high-angle oblique fractures. In this application, for different types of fractures, the formation fracture identification model can extract local and global features, accurately identifying the fractures. Therefore, for complex fractures, this application can use the formation fracture identification model to extract local and global features in multiple layers to obtain more accurate local and global features. This application can also capture abstract features to improve accuracy.

[0064] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0065] Example 1

[0066] The execution subject of Embodiments 1 to 7 of this application is a formation fracture identification device (hereinafter referred to as the identification device), which is located in an electronic device.

[0067] Figure 2 This is a schematic flowchart of a formation fracture identification method provided in Embodiment 1. Figure 2 As shown, it specifically includes:

[0068] S201, acquire imaging logging images corresponding to the target formation; at least one imaging logging image is required; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging images include oil-based drilling fluid micro-resistivity imaging logging images.

[0069] It should be noted that before obtaining the imaging logging image corresponding to the target formation, the user can input basic data into the recognition device, so that the recognition device can generate and obtain the imaging logging image based on the obtained basic data.

[0070] The basic data includes core data, test data, and imaging logging data corresponding to the target formation. The core data includes surface gamma-ray data, transverse and longitudinal scans. The imaging logging data includes electrical imaging data.

[0071] The target strata include deep and ultra-deep complex tight sandstone oil and gas reservoirs, or shallow strata (or conventional strata), without any restrictions.

[0072] Among them, the imaging logging images include oil-based drilling fluid subsurface microresistivity imaging logging images, which can be obtained or generated from electrical imaging data.

[0073] It should be noted that imaging logging images include both the image capture width and the image capture height. The image capture width represents the width range of the target formation, while the image capture height represents the height range of the target formation. For example, assuming an image capture width of 0-5m, the target formation corresponds to a 0-5m range in the lateral direction. This lateral direction is defined by point O, located at a height of 0m on the right wall of the wellbore when facing the wellbore directly, and extends to the right of the right wall as the lateral direction (i.e., the X direction). The image capture height is the vertical downward height of the formation, extending downward from point O on the right wall of the wellbore as the starting point (i.e., the longitudinal direction, the Z direction).

[0074] Among them, the image capture width refers to the width of the target stratum in the actual scene of the target stratum.

[0075] Among them, the image capture height refers to the height of the target stratum in the actual scene where the image is captured.

[0076] It should be noted that an image has a certain width and height. The image width is proportional to the image capture width, and similarly, the image height is proportional to the image capture height.

[0077] It should be noted that imaging logging images can be of any resolution; there are no restrictions here.

[0078] It should be noted that there can be at least one imaging logging image, and each image corresponds to a different image capture width. It should also be noted that the image capture height for each imaging logging image can be the same. For example, the first imaging logging image has an image capture width of 0-5m and an image capture height of 0-2000m, and the second imaging logging image has an image capture width of 5-10m and an image capture height of 0-2000m. The starting height within the 0-2000m range is 0m, and the ending height is 2000m.

[0079] It should be noted that the imaging logging images in this application can also be images corresponding to the downcut plane at any angle within a 360° radius along the Z direction.

[0080] S202, for each imaging logging image, obtain multiple corresponding input sub-images based on the imaging logging image.

[0081] In one approach, the imaging logging image can be divided into multiple input sub-images corresponding to the imaging logging image by dividing it into equal parts according to a preset cutting height.

[0082] For example, assuming the image acquisition height of the imaging logging image is 0-2000m and the preset cutting height is 1m, the image is evenly divided according to the preset cutting height, resulting in 2000 input sub-images. The first input sub-image occupies an acquisition height of 0-1m. The initial acquisition height of the imaging logging image is 0m, and the final acquisition height is 2000m.

[0083] For example, in this approach, the input sub-image can be represented as: ,in, It refers to the first One input sub-image, The initial height is the image capture height corresponding to the imaging logging image. This refers to the shooting height occupied by the input sub-image in the imaging logging image. This refers to the preset cutting height. Based on the above example, the shooting height occupied by the first input sub-image is... That is, the shooting height occupied by the first input sub-image in the imaging logging image is 0-1m.

[0084] S203, based on multiple input sub-images corresponding to the imaging logging images and the formation fracture recognition model that has been trained to convergence, multiple output sub-images corresponding to the input sub-images are generated to complete fracture recognition; the output sub-images include the fracture situation corresponding to the input sub-images; the output sub-images are generated by the formation fracture recognition model after extracting local and global features from the input sub-images; the formation fracture recognition model is generated by training on a training set; the training set includes at least one training image and its corresponding label image; the label image includes the fractures present in the training image; the label image is obtained by labeling the training image with a preset labeling software and then binarizing it.

[0085] Local features refer to the prominent and minute features of cracks in the input sub-image that represent the shallow layer. These features can include crack edge features and corner features, and there are no restrictions on them here.

[0086] Among them, global features refer to the appearance and shape of the cracks in the input sub-image, which represent the features of the crack depth.

[0087] The crack situation corresponding to the input sub-image includes the presence of cracks and the absence of cracks. If cracks exist, the output sub-image describes the cracks corresponding to the input sub-image; if cracks do not exist, the output sub-image is a blank image.

[0088] The cracks in the output sub-image are two-dimensional cracks, including the shape, length and width of the cracks.

[0089] In one approach, for each imaging logging image, multiple corresponding input sub-images are input into the formation fracture identification model to obtain the output sub-image corresponding to each input sub-image.

[0090] In one approach, for each imaging logging image, multiple corresponding input sub-images are batch-input into the formation fracture identification model, thereby obtaining the output sub-images corresponding to each input sub-image in batches.

[0091] In this step, the formation fracture identification model is generated based on at least one training image and its corresponding label image. The label image is a binary image, i.e., a monochrome image, which includes the fractures corresponding to the training image. The fractures in the label images of this application are also two-dimensional fractures.

[0092] Among them, the formation fracture identification model that has been trained to convergence can be trained on the basis of the medical image segmentation model (U-net neural network).

[0093] It should be noted that the labeled images were obtained by labeling using pre-set labeling software. Specifically, the pre-set labeling software was used to label the cracks in the training images to outline their appearance, shape, length, and width, and then binarization was performed to obtain the labeled images.

[0094] It should be noted that the resolution of the training images in this step is the same as that of the input sub-images.

[0095] It should be noted that the training images in this step are obtained by filling in the training imaging images and then cutting them out. The cutting method in the training phase is the same as that in the application phase, as detailed in step S202 or Example 3, and will not be repeated here.

[0096] S204: For each imaging logging image, multiple output sub-images corresponding to the imaging logging image are stitched together to obtain the formation fracture image corresponding to the imaging logging image.

[0097] In one approach, the shooting height of the output sub-image is consistent with that of the corresponding input sub-image. For each imaging logging image, multiple output sub-images are stitched together according to the shooting height of each output sub-image from top to bottom, thereby obtaining the formation fracture image corresponding to the imaging logging image.

[0098] It should be noted that the image capture height of the formation fracture image is the same as that of the imaging logging image, and similarly, the image capture width of the two is also the same.

[0099] In one approach, this embodiment stitches together formation fracture images corresponding to each imaging logging image. Simultaneously, the fracture types included in the formation fracture images can be identified, specifically filling fractures, semi-open fractures, and open fractures. More specifically, corresponding fracture type images can be generated based on the formation fracture images, and the type of each fracture can be labeled in the fracture type images.

[0100] This embodiment provides a method for identifying formation fractures. First, multiple input sub-images are obtained from the acquired imaging logging images. Then, based on these multiple input sub-images and a converged formation fracture identification model, corresponding output sub-images are identified. The output sub-images include the corresponding fractures. Since the output sub-images in this embodiment are generated by the formation fracture identification model after extracting local and global features from the corresponding input sub-images, they incorporate both local and global features, making the output sub-images more accurate. Therefore, the fractures in the output sub-images are more accurate, reducing the likelihood of missed or incorrect identification. Furthermore, the formation fracture identification model in this embodiment is generated based on at least one training image and its corresponding label image. Since the formation fracture identification model can extract both local and global features, the training process also considers both local and global features of the training image, thus optimizing the model. Therefore, the formation fracture identification model is an improved model that can more accurately obtain output sub-images. Additionally, the target formation in this embodiment can be deep and ultra-deep tight sandstone oil and gas reservoirs under oil-based drilling fluids. Therefore, in complex formations, facing complex fractures… This embodiment uses a formation fracture identification model to identify the input sub-image. Even if the fracture situation represented in the input sub-image is complex, local and global features can still be extracted and fused to obtain a more accurate output sub-image, which more accurately represents the fracture situation. In this embodiment, the label image is obtained by labeling the training image with a preset labeling software and performing binarization. Therefore, the label image outlines the fractures included in the training image, making the label image accurate. As a result, the formation fracture identification model trained based on the training set is more accurate.

[0101] Example 2

[0102] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for obtaining imaging logging images corresponding to the target formation, specifically including:

[0103] Step 1:

[0104] Obtain electrical imaging data corresponding to the target strata.

[0105] The aforementioned electrical imaging data may include the resistivity detected by eight electrodes around the wellbore at a 360° angle.

[0106] In one approach, noise data from acoustic imaging data can also be used to obtain imaging logging images. Here, acoustic imaging data is noise data acquired by a sound detector. In another approach, imaging logging images are generated and obtained from the noise data of the fluid in the acoustic imaging data.

[0107] Step 2: Obtain imaging logging images based on electrical imaging data.

[0108] In one approach, the resistivity in the fractures and wellbore is obtained from electrical imaging data, and a microresistivity imaging logging map is drawn based on each resistivity. The aforementioned microresistivity imaging logging map is the imaging logging image, thereby obtaining the imaging logging image.

[0109] This embodiment provides a method for identifying formation fractures. In this embodiment, imaging logging images can be accurately generated and obtained based on electrical imaging data or acoustic imaging data.

[0110] Example 3

[0111] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for obtaining multiple input sub-images based on imaging logging images.

[0112] Figure 3 This is a schematic diagram of a formation fracture identification method provided in Example 3. Figure 3 As shown, it specifically includes:

[0113] It should be noted that due to the harsh testing environment, some areas in the imaging logging images may not be detected, which are called blank areas.

[0114] S301, fill the blank area to obtain the filled image.

[0115] It should be noted that filling blank areas can ensure the integrity of the image.

[0116] It should be noted that imaging logging images are three-channel images, which are filled with the colors near the blank areas to obtain the filled image.

[0117] Specifically, an interpolation mechanism is used to fill the blank areas.

[0118] S302, the filled image is segmented to obtain multiple input sub-images.

[0119] In one approach, the filled image can be divided into equal sections or cut in overlapping sections.

[0120] The overlapping cutting method is detailed in Example 7 and will not be repeated here.

[0121] This embodiment provides a method for identifying formation fractures. In order to ensure completeness, blank areas in the imaging logging image are filled in, thereby obtaining a filled image that provides a more complete description of the target formation.

[0122] Example 4

[0123] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for generating output sub-images corresponding to multiple input sub-images based on multiple input sub-images corresponding to imaging logging images and a formation fracture identification model that has been trained to convergence, in order to complete fracture identification.

[0124] Step 1: Input multiple input sub-images corresponding to the imaging logging images into the formation fracture identification model in batches.

[0125] In one approach, multiple input sub-images are batch-input into a formation fracture identification model.

[0126] In another approach, multiple input sub-images are divided into batches to obtain multiple batches of input sub-images. Each batch of input sub-images includes at least two input sub-images. Then, the input sub-images included in the corresponding batches are input into the formation fracture identification model in batches.

[0127] Step 2: Use the formation fracture identification model to generate low-resolution feature maps corresponding to multiple input sub-images.

[0128] In one approach, an encoder in a formation fracture identification model is used to batch extract local features from multiple input sub-images and downsample them to generate a low-resolution feature map.

[0129] Low-resolution feature maps refer to feature maps with a resolution lower than that of the input sub-image. These low-resolution feature maps include local features, which are extracted based on the color and texture of cracks in the input sub-image.

[0130] Step 3: The formation fracture identification model is used to generate multiple output sub-images corresponding to the input sub-images based on each low-resolution feature map to complete the fracture identification; the output sub-images are binarized images.

[0131] In one approach, a decoder in a formation fracture identification model is used to upsample each low-resolution feature map in batches and extract global features to generate corresponding output sub-images. The output sub-images include the corresponding fractures, thereby completing fracture identification.

[0132] This embodiment provides a method for identifying formation fractures. In this embodiment, multiple input sub-images can be batch-input into the formation fracture identification model, thereby achieving batch processing and improving processing speed.

[0133] Example 5

[0134] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for generating low-resolution feature maps corresponding to multiple input sub-images using a formation fracture identification model, specifically including:

[0135] Step 1: The first convolutional layer in the formation fracture identification model is used to extract local features from each input sub-image to obtain the local feature map corresponding to each input sub-image; the local features include edge and corner features in the input sub-image; the first convolutional layer is located in the encoder.

[0136] It should be noted that the formation fracture identification model includes an encoder and a decoder, which are symmetrical. The encoder includes a first convolutional layer and a pooling layer; the decoder includes an upsampling layer and a second convolutional layer. The first convolutional layer can include multiple layers.

[0137] Specifically, for any input sub-image, the first convolutional layer extracts local features from the input sub-image in ascending order of hierarchy, thereby extracting local characteristic information such as the color and texture of cracks in the input sub-image and obtaining a local feature map. In this embodiment, the local feature extraction is performed in ascending order of hierarchy, achieving multi-layer local feature extraction, making the local features more accurate and more consistent with the actual situation.

[0138] It should be noted that when generating local feature maps, binarization can be performed, resulting in a binarized image of the local feature map.

[0139] The local feature map is the image generated after extracting local features from the input sub-image.

[0140] It should be noted that local features are prominent, minute features, and are a type of shallow feature.

[0141] Step 2: Use the pooling layer in the formation fracture identification model to downsample each local feature map to capture abstract features and generate corresponding low-resolution feature maps.

[0142] In one approach, for any input sub-image, the pooling layer downsamples the local feature map at a preset resolution. During downsampling, key features are preserved while abstract features are captured, thereby generating a low-resolution feature map. The preset resolution is smaller than the resolution corresponding to the input sub-image.

[0143] This embodiment provides a method for identifying formation fractures. In this embodiment, the first convolutional layer extracts local features from the input sub-images. These local features include edge and corner features, which are prominent, small, and shallow features, thus enabling more accurate identification of fractures in the target formation. Furthermore, during downsampling, abstract features are captured to generate corresponding low-resolution feature maps, which in turn combine multiple features, resulting in greater accuracy.

[0144] Example 6

[0145] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for completing fracture identification by using a formation fracture identification model to generate multiple output sub-images corresponding to each input sub-image based on each low-resolution feature map.

[0146] Figure 4 This is a schematic flowchart of a formation fracture identification method provided in Example 6. Figure 4 As shown, it specifically includes:

[0147] S401, the upsampling layer in the formation fracture identification model is used to upsample the low-resolution feature map in order to identify the upsampled feature map; the resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image.

[0148] It should be noted that the purpose of upsampling is to restore the resolution of the low-resolution feature map so that its resolution is consistent with that of the corresponding input sub-image.

[0149] Among them, the upsampled feature map is the feature map obtained by upsampling to restore the resolution.

[0150] S402, the second convolutional layer in the geological fracture identification model is used to extract global features from each input sub-image in a hierarchical order; the second convolutional layer is located in the decoder; the global features include the appearance and shape features of the input sub-image; the second convolutional layer includes at least one.

[0151] For any input sub-image, the second convolutional layer extracts the global features of the cracks in the input sub-image to extract the global characteristic information of the cracks' color and texture, thus obtaining the global features.

[0152] It should be noted that in this embodiment, the second convolutional layer may include at least one. Therefore, when extracting global features from the input sub-image, the global features can be extracted sequentially by the second convolutional layer in each layer according to the hierarchical order of the second convolutional layer, so as to obtain more complete and realistic global features, that is, multi-layer global feature extraction is realized.

[0153] S403, fuse each global feature into the corresponding upsampled feature map to generate a fused feature map for each input sub-image; the fused feature map includes local features and global features.

[0154] The fused feature map combines both local and global features.

[0155] S404, each fused feature map is determined as the corresponding output sub-image to complete crack recognition.

[0156] It should be noted that after obtaining the output sub-image, it is output from the formation fracture identification model.

[0157] This embodiment provides a method for identifying formation fractures. In this embodiment, the low-resolution feature map is upsampled to restore the resolution, so that the resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image, thereby ensuring that the resolution of the output sub-image is consistent with the resolution of the input sub-image. In this embodiment, the fusion feature map fuses local and global features, so it can better describe the fractures in the target formation and thus identify the fractures more accurately.

[0158] Example 7

[0159] This embodiment is a further refinement of any of the above embodiments. This embodiment is an optional method for segmenting the filled image to obtain multiple input sub-images, specifically including:

[0160] Keeping the cutting width consistent with the image capture width, the filled image is vertically overlapped and cut within the image capture height according to the preset step size and preset cutting height to obtain multiple input sub-images; the overlap height is the difference between the preset cutting height and the preset step size.

[0161] The preset step size refers to the preset distance between the initial height of the previous input sub-image and the initial height of the next input sub-image.

[0162] Here, overlap height refers to the height of overlap between two adjacent input sub-images.

[0163] The preset cutting height refers to the pre-set cutting height for each input sub-image. For example, it can be 1m or other heights, and the preset cutting height is less than the image capture height.

[0164] Specifically, the cutting width is kept consistent with the image capture width; that is, the horizontal width of the filled image remains unchanged while it is cut vertically.

[0165] For example, assuming the width of the filled image is 0-5m, occupies 5m, the image height is 0-200m, the preset cutting height is 2m, and the preset step size is 1m, the filled image is vertically overlapped and cut to obtain the first input sub-image with an image width of 0-5m and an image height of 0-2m. According to the preset step size (1m), the second input sub-image has an image width of 0-5m and an image height of 1-3m.

[0166] In this embodiment, the input sub-image obtained by overlapping cutting can be represented as: ,in, It refers to the first One input sub-image, The initial height is the image capture height corresponding to the imaging logging image. This refers to the shooting height occupied by the input sub-image in the imaging logging image. This refers to the preset cutting height. This refers to the preset step size. Based on the above example, the shooting height occupied by the first input sub-image is... That is, the first input sub-image occupies a shooting height of 0-2m in the imaging logging image; the second input sub-image occupies a shooting height of... .

[0167] Figure 5 This is a schematic diagram of an overlapping cut provided in Embodiment Seven. (See attached diagram.) Figure 5 As shown, it includes the filled image 501 and multiple cut input sub-images 502.

[0168] exist Figure 5 In this example, assuming the filled image has a capture width of 0-5m, occupies 5m in the X direction, and a capture height of 0-20m, with a preset cutting height of 2m and a preset step size of 1m, the first input sub-image has a capture width of 0-5m and a capture height of 0-2m. The second input sub-image has a capture width of 0-5m and a capture height of 1-3m. The overlap height between the first and second input sub-images is 1-2m (the capture height), totaling 1m, which is the preset step size. This process continues until the final input sub-image has a capture width of 0-5m and a capture height of 18-20m.

[0169] The filled image and the input sub-image are both three-channel images.

[0170] It should be noted that this embodiment also includes the training method of the already converged formation fracture identification model, as follows:

[0171] Step 1: Obtain the initial model; the initial model includes the initial model parameters.

[0172] The initial model can be any version of a medical image segmentation model (U-net neural network) that has been trained to convergence.

[0173] Step 2: Input at least one training image and its corresponding label image into the initial model for training to obtain a converged formation fracture recognition model.

[0174] In one approach, at least one training image and its corresponding label image are input into an initial model for training to obtain a converged formation fracture identification model, specifically including the following steps:

[0175] Step 1: Divide at least one training image and its corresponding label image into batches to obtain multiple batches of training image sets.

[0176] Each batch of training image sets includes at least one training image and its corresponding label image. Each batch of training image sets includes a preset number of training data sets. It should be noted that the training data includes the training image and its corresponding label image. The next batch of training image sets includes a portion of the training data from the previous batch. The number of these portions of training data is a preset overlap number, which refers to the preset overlap number of training data sets at the end of the sorted sequence from the previous batch. For example, assuming there are a total of 5 training images and their corresponding label images (i.e., 5 training data sets, arranged by sequence number), a preset overlap number of 2, and a preset batch number of 3, then the training image sets are divided into batches based on the preset overlap number and the preset batch number. The first batch of training image sets includes 1-3 training data sets, the second batch includes 2-4 training data sets, and the third batch includes 3-5 training data sets. The second batch includes 2 training data sets that overlap with the first batch, namely the second and third training data sets.

[0177] It should be noted that the label images in this embodiment are obtained by labeling the corresponding training images with preset labeling software and performing binarization processing. The label images are binarized images.

[0178] The labeled image may include a crack corresponding to the training image. The crack is a two-dimensional crack and includes information such as the shape, length, and width of the crack.

[0179] The labeling process involves marking the cracks in the training images to delineate them.

[0180] Step 2: Input the current batch of training images into the initial model in batch order to train the model and obtain the loss value corresponding to each training data in the current batch of images.

[0181] Step 3: Calculate the average loss value corresponding to each training data in the current batch of images.

[0182] Step 4: If the average loss value of the current batch of training images is less than the saved optimal average loss value, then update the saved optimal average loss value to the current average loss value, take the current average loss value as the optimal average loss value, and retain the optimized model parameters corresponding to the current average loss value.

[0183] In one approach, when the current batch of training images is used to train the initial model, each training data in the current batch of training images can optimize the model parameters. Therefore, each training data can correspond to its own optimized model parameters (i.e., correspond to at least one optimized model parameter). When retaining the optimized model parameters corresponding to the current average loss value, the optimal loss value can be determined from the loss values ​​corresponding to each training data. The optimized model parameters of the training data corresponding to the optimal loss value are determined as the optimized model parameters corresponding to the current average loss value to be retained, and then retained.

[0184] In another approach, for the current batch of training image sets, the optimized model parameters corresponding to the last training data can be determined as the optimized model parameters corresponding to the current average loss value to be retained, and then retained.

[0185] Step 5: Continue to input the next batch of training image sets into the model with optimized model parameters from Step 4. Optimize the model parameters based on the optimized model parameters until the next batch of training image sets is trained. Obtain the average loss value and corresponding optimized model parameters for the next batch. Use the next batch of training image sets as the current training image set and continue to execute Step 4 and Step 5 until all batches of training image sets are trained. Obtain the optimal optimized model parameters. The optimal optimized model parameters are the formation fracture recognition model trained to convergence.

[0186] It should be noted that, in order to improve the generalization and accuracy of the formation fracture identification model, some training images may not include fractures, and their corresponding label images will be blank images. Therefore, in practical applications, if there are input sub-images that do not include fractures, the formation fracture identification model can accurately generate an output sub-image that is a blank image, that is, it will not make incorrect identifications.

[0187] The loss value can be calculated based on the binary cross-entropy loss function.

[0188] This embodiment provides a method for identifying formation fractures. In this embodiment, overlapping cutting can ensure the integrity of fracture information in the input sub-image.

[0189] This embodiment provides a complete schematic diagram for identifying formation fractures. Figure 6 This is a complete schematic diagram of a formation fracture identification method provided in Example 7.

[0190] like Figure 6 The process includes an application phase and a training phase. The application phase includes an imaging logging image 601 of the target formation, an image 602 after filling the blank areas in the imaging logging image 601, an input sub-image 603 after overlapping and cutting the filled image 602, a formation fracture recognition model 604 trained to convergence in this application, and an output sub-image 605 obtained by inputting the input sub-image 603 into the formation fracture recognition model 604.

[0191] exist Figure 6 In this process, the input sub-image 603 is input into the formation fracture identification model 604, and the corresponding output sub-image 605 is output from the formation fracture identification model 604.

[0192] The crack in the output sub-image 605 is a two-dimensional and complete curve.

[0193] It should be noted that, in Figure 6 In this example, we will use a single input sub-image 603 as an example. In practical applications, after overlapping and segmenting the filled image 602, multiple input sub-images 603 can be obtained. The processing method for multiple input sub-images 603 is as follows: Figure 6 The input sub-image 603 is the same, so it will not be elaborated here.

[0194] exist Figure 6 The dataset also includes another input sub-image 606, which is obtained by filling blank areas and overlapping and cutting the remaining imaging logging images. This other input sub-image 606 exhibits more complete fractures.

[0195] exist Figure 6 In this process, another input sub-image 606 is input into the formation fracture recognition model 604 to obtain a first sub-image 607 of a two-dimensional fracture. The first sub-image 607 is a two-dimensional fracture with a complete fracture morphology.

[0196] In one input sub-image, 606, a more complete crack is observed, while in another input sub-image, 603, a complete crack is not observed. However, when input sub-image 603 is input into the formation crack recognition model 604, this application can still output a complete crack morphology output sub-image, 605. Therefore, this application can also identify complex cracks in complex formations and accurately identify the corresponding cracks, recognizing the complete morphology of complex cracks.

[0197] exist Figure 6 The process also includes a traditional recognition model 608, which inputs the input sub-image 603 into the traditional recognition model 608 and outputs a second sub-image 609.

[0198] In particular, the crack morphology in the second sub-image 609 is incomplete, appearing as a discontinuous crack. The crack morphology in the second sub-image 609 differs greatly from that in the output sub-image 605, demonstrating that the output sub-image 605 can identify the complete crack morphology, thereby improving accuracy.

[0199] exist Figure 6 It also includes the well logging software 610.

[0200] The input sub-image 603 is input into the logging software 610, and a third sub-image 611 is output.

[0201] The crack in the third sub-image is a one-dimensional crack, which includes the crack shape and length.

[0202] exist Figure 6 As can be seen, the crack in the output sub-image 605 is a two-dimensional crack, including the crack shape, length and width; while the crack in the third sub-image 611 is a one-dimensional crack, including the crack shape and length. Since the output sub-image 605 is two-dimensional, the output sub-image 605 is more accurate than the third sub-image 611.

[0203] exist Figure 6 The image also includes a formation fracture image 612, which is a stitched image of the formation fractures formed by stitching together the output sub-image 605 and other output sub-images 605 (not shown in the figure) corresponding to the imaging logging image 601.

[0204] Furthermore, the fracture width, fracture density, and fracture length included in the formation fracture image 612 are analyzed to obtain the fracture analysis diagram 613.

[0205] exist Figure 6 The training phase also includes a medical image segmentation model 614, at least one training image 615 and its corresponding label image 616.

[0206] exist Figure 6In the training image 615, the training image 617 is obtained by filling the blank areas and cutting it.

[0207] Among them, training image 617 is an imaging logging image of the training formation.

[0208] In this process, the training image 615 is labeled using a pre-set labeling software to obtain the training labeled image 618.

[0209] Specifically, the training labeled image 618 is binarized to obtain the label image 616 corresponding to the training image 615.

[0210] Furthermore, training data is obtained according to the above method, resulting in at least one training image 615 and its corresponding label image 616. The medical image segmentation model 614 is then trained using at least one training image 615 and its corresponding label image 616 to obtain a converged formation fracture recognition model 604. This converged model 604 is then used to identify fractures in the target formation. Figure 6 In this example, we will use one training data point as an example. In the actual training phase, at least one training data point is required.

[0211] Example 8

[0212] The following is an embodiment of the device of this application. Figure 7 This is a schematic diagram of a formation fracture identification device provided in Embodiment 8. Figure 7 As shown, the formation fracture identification device 700 includes the following modules:

[0213] The acquisition module 701 is used to acquire imaging logging images corresponding to the target formation; the imaging logging image is at least one; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging image includes oil-based drilling fluid subsurface micro-resistivity imaging logging images.

[0214] The module 702 is used to obtain multiple corresponding input sub-images based on each imaging logging image.

[0215] The identification module 703 is used to generate multiple output sub-images corresponding to the input sub-images based on multiple input sub-images corresponding to the imaging logging images and a formation fracture identification model that has been trained to convergence, so as to complete fracture identification. The output sub-images include the fracture conditions corresponding to the input sub-images. The output sub-images are generated by the formation fracture identification model after extracting local and global features from the input sub-images. The formation fracture identification model is generated by training on a training set. The training set includes at least one training image and its corresponding label image. The label image includes the fractures present in the training image. The label image is obtained by labeling the training image using a preset labeling software and then binarizing it.

[0216] The stitching module 704 is used to stitch together multiple output sub-images corresponding to each imaging logging image to obtain a formation fracture image corresponding to the imaging logging image.

[0217] The acquisition module 701, when acquiring the imaging logging image corresponding to the target formation, is specifically used for:

[0218] Acquire electrical imaging data corresponding to the target strata;

[0219] Imaging logging images are obtained based on electrical imaging data.

[0220] The module 702, when obtaining multiple input sub-images based on imaging logging images, is specifically used for:

[0221] Fill the blank areas to obtain the filled image;

[0222] The padded image is then segmented to obtain multiple input sub-images.

[0223] The module 702, when segmenting the padded image to obtain multiple input sub-images, is specifically used for:

[0224] Keeping the cutting width consistent with the image capture width, the filled image is vertically overlapped and cut within the image capture height according to the preset step size and preset cutting height to obtain multiple input sub-images; the overlap height is the difference between the preset cutting height and the preset step size.

[0225] The identification module 703, when generating output sub-images corresponding to multiple input sub-images based on multiple input sub-images corresponding to imaging logging images and a formation fracture identification model that has been trained to convergence, is specifically used for:

[0226] Multiple input sub-images corresponding to the imaging logging images are batch input into the formation fracture identification model;

[0227] A formation fracture identification model is used to generate low-resolution feature maps corresponding to multiple input sub-images;

[0228] A formation fracture identification model is used to generate multiple output sub-images corresponding to each input sub-image based on each low-resolution feature map to complete fracture identification; the output sub-images are binarized images.

[0229] The recognition module 703, when generating low-resolution feature maps corresponding to multiple input sub-images using the formation fracture recognition model, is specifically used for:

[0230] The first convolutional layer in the geological fracture identification model is used to extract local features from each input sub-image to obtain a local feature map corresponding to each input sub-image; the local features include edge and corner features in the input sub-image; the first convolutional layer is located in the encoder;

[0231] The pooling layer in the formation fracture identification model is used to downsample each local feature map to capture abstract features and generate corresponding low-resolution feature maps.

[0232] The identification module 703, when using a formation fracture identification model to generate multiple output sub-images corresponding to each input sub-image based on each low-resolution feature map to complete fracture identification, is specifically used for:

[0233] The low-resolution feature map is upsampled using an upsampling layer in the formation fracture identification model in order to identify the upsampled feature map; the resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image.

[0234] The second convolutional layer in the geological fracture identification model is used to extract global features from each input sub-image in a hierarchical order; the second convolutional layer is located in the decoder; the global features include the appearance and shape features of the input sub-image; the second convolutional layer includes at least one;

[0235] Each global feature is fused into the corresponding upsampled feature map to generate a fused feature map for each input sub-image; the fused feature map includes local features and global features.

[0236] Each fused feature map is determined as the corresponding output sub-image to complete crack identification.

[0237] The formation fracture identification device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0238] Figure 8 A schematic diagram of the structure of the electronic device provided in this application. Figure 8 As shown, the electronic device 800 provided in this embodiment includes at least one processor 801 and a memory 802. Optionally, the electronic device 800 further includes a communication component 803. The processor 801, memory 802, and communication component 803 are connected via a bus 804.

[0239] In a specific implementation, at least one processor 801 executes computer execution instructions stored in memory 802, causing at least one processor 801 to perform the above-described method.

[0240] The specific implementation process of processor 801 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0241] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0242] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.

[0243] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0244] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0245] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.

[0246] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0247] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.

[0248] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0249] The units described as separate components may or may not be physically separate. 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 the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0250] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0251] If a function is implemented as 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 this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this 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.

[0252] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0253] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for identifying formation fractures, applied to well logging engineering in deep and ultra-deep tight sandstone oil and gas reservoirs under oil-based drilling fluid conditions, characterized in that... include: Acquire imaging logging images corresponding to the target formation; the imaging logging image is at least one; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging image includes oil-based drilling fluid subsurface micro-resistivity imaging logging images. For each imaging logging image, multiple corresponding input sub-images are obtained based on the imaging logging image; Multiple input sub-images corresponding to the imaging logging images are batch input into the formation fracture recognition model. The first convolutional layer in the formation fracture recognition model extracts local features from each input sub-image to obtain a local feature map corresponding to each input sub-image. The local features include edge and corner features in the input sub-image. The first convolutional layer is located in the encoder. The pooling layer in the formation fracture recognition model downsamples each local feature map to capture abstract features and generate a corresponding low-resolution feature map. The upsampling layer in the formation fracture recognition model upsamples the low-resolution feature map to identify the upsampled feature map. The resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image. The second convolutional layer in the formation fracture recognition model extracts global features from each input sub-image in a hierarchical, multi-layered manner. The second convolutional layer is located in the decoder. The global features include the appearance and shape features of the input sub-image. The second convolutional layer includes at least one component. Each global feature is fused into the corresponding upsampled feature map to generate a fused feature map corresponding to each input sub-image. The fused feature map includes local and global features. Each fused feature map is determined as a corresponding output sub-image to complete crack identification; the output sub-image is a binarized image; the output sub-image includes the crack situation corresponding to the input sub-image; the two-dimensional crack situation corresponding to the input sub-image includes the presence of cracks and the absence of cracks. If cracks exist, the output sub-image includes the shape, length, and width of the two-dimensional cracks corresponding to the input sub-image; if cracks do not exist, the output sub-image is a blank image; the formation crack identification model is generated by training on a training set; the training set includes at least one training image and its corresponding label image; the label image includes the cracks present in the training image; the label image is obtained by labeling the training image using preset labeling software and then binarizing it; the label image includes the shape, length, and width of the two-dimensional cracks; For each imaging logging image, multiple output sub-images corresponding to the imaging logging image are stitched together to obtain the formation fracture image corresponding to the imaging logging image.

2. The method according to claim 1, characterized in that, The acquisition of imaging logging images corresponding to the target formation includes: Acquire electrical imaging data corresponding to the target stratum; Imaging logging images are obtained based on the electrical imaging data.

3. The method according to claim 1, characterized in that, The process of obtaining multiple input sub-images based on the imaging logging images includes: Fill the blank areas to obtain the filled image; The filled image is segmented to obtain multiple input sub-images; The step of segmenting the filled image to obtain multiple input sub-images includes: Keeping the cutting width consistent with the image capture width, the filled image is vertically overlapped and cut within the image capture height according to a preset step size and a preset cutting height to obtain multiple input sub-images; the overlap height is the difference between the preset cutting height and the preset step size.

4. A formation fracture identification device, applied to formation logging engineering in deep and ultra-deep tight sandstone oil and gas reservoirs under oil-based drilling fluid environment, characterized in that, include: The acquisition module is used to acquire imaging logging images corresponding to the target formation; the imaging logging image is at least one; the target formation includes deep and ultra-deep complex tight sandstone oil and gas reservoir formations; the imaging logging image includes oil-based drilling fluid subsurface micro-resistivity imaging logging images. The acquisition module is used to obtain multiple corresponding input sub-images based on each imaging logging image; The recognition module is used to batch input multiple input sub-images corresponding to the imaging logging images into the formation fracture recognition model; the first convolutional layer in the formation fracture recognition model is used to extract local features from each input sub-image to obtain a local feature map corresponding to each input sub-image; the local features include edge and corner features in the input sub-image; the first convolutional layer is located in the encoder; the pooling layer in the formation fracture recognition model is used to downsample each local feature map to capture abstract features and generate a corresponding low-resolution feature map; the upsampling layer in the formation fracture recognition model is used... The low-resolution feature map is upsampled to identify the upsampled feature map; the resolution of the upsampled feature map is consistent with the resolution of the corresponding input sub-image; the second convolutional layer in the formation fracture identification model is used to extract global features from each input sub-image in a hierarchical manner; the second convolutional layer is located in the decoder; the global features include the appearance and shape features of the input sub-image; the second convolutional layer includes at least one; each of the global features is fused into the corresponding upsampled feature map to generate a fused feature map corresponding to each input sub-image; the fused feature map includes local features and global features; Each fused feature map is determined as a corresponding output sub-image to complete crack identification; the output sub-image is a binarized image; the output sub-image includes the crack situation corresponding to the input sub-image; the two-dimensional crack situation corresponding to the input sub-image includes the presence of cracks and the absence of cracks. If cracks exist, the output sub-image includes the shape, length, and width of the two-dimensional cracks corresponding to the input sub-image; if cracks do not exist, the output sub-image is a blank image; the formation crack identification model is generated by training on a training set; the training set includes at least one training image and its corresponding label image; the label image includes the cracks present in the training image; the label image is obtained by labeling the training image using preset labeling software and then binarizing it; the label image includes the shape, length, and width of the two-dimensional cracks; The stitching module is used to stitch together multiple output sub-images corresponding to each imaging logging image to obtain a formation fracture image corresponding to the imaging logging image.

5. An electronic device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-3.

7. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-3.

Images