Oil and gas facility identification model generation method, oil and gas facility identification method and device

Abstract

This specification relates to the field of oil and gas facility identification, and particularly to a method for generating an oil and gas facility identification model, an oil and gas facility identification method, and an apparatus. The generation method includes: acquiring multiple oil and gas facility images and a first associated ground feature image corresponding to each oil and gas facility image; generating an oil and gas facility identification dataset based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image; and generating an oil and gas facility identification model using the oil and gas facility images and their associated ground feature images as input, the oil and gas facilities as output, and the oil and gas facility identification dataset. This specification's embodiments solve the problem of poor oil and gas facility identification performance.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of oil and gas facility identification, and in particular to a method for generating an oil and gas facility identification model, an oil and gas facility identification method, and an apparatus. Background Technology

[0002] With the continuous expansion of oil and gas field exploration and development, the number of oil and gas facilities, such as stations and well sites, belonging to oil and gas field enterprises has been increasing year by year. The number of oil and gas facilities in some oil and gas fields has reached tens of thousands. How to effectively supervise them has become an urgent problem to be solved. Especially when it is necessary to determine the attribution of environmental pollution and air pollution incidents, timely and accurate understanding of the distribution of oil and gas facilities in oil and gas fields is particularly important.

[0003] Currently, deep learning technology has been widely used in the field of target detection, and some common land features (such as airplanes and football fields) can be effectively identified. However, when applied to target detection, existing deep learning technology only considers the features of the target object and does not learn the features of the surrounding environment. Therefore, it is not suitable for scenarios with many interfering factors. For example, oil and gas facilities and some land features (such as oil and gas facilities such as stations and well sites and industrial land exhibit similar features and textures in remote sensing images) have a high degree of similarity. When using existing deep learning algorithms to identify oil and gas facilities, there is a possibility of misidentifying industrial land as oil and gas facilities, resulting in a high false recognition rate. Therefore, it is difficult to widely apply it to the identification of oil and gas facilities. Summary of the Invention

[0004] This invention provides a method for generating an oil and gas facility identification model, an oil and gas facility identification method, and an apparatus to solve the problem of poor identification results for oil and gas facilities.

[0005] To address the aforementioned technical problems, a first aspect of this specification provides a method for generating an oil and gas facility identification model, the method comprising:

[0006] Acquire multiple images of oil and gas facilities and the first associated ground feature image corresponding to each oil and gas facility image;

[0007] Based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image, an oil and gas facility identification dataset is generated.

[0008] Using images of oil and gas facilities and their associated ground features as input, and the oil and gas facilities as output, an oil and gas facility identification model is generated using the oil and gas facility identification dataset.

[0009] Further, multiple oil and gas facility images and a first associated ground feature image corresponding to each oil and gas facility image are acquired, including:

[0010] Acquire remote sensing images containing oil and gas facilities and related features;

[0011] The remote sensing image is segmented to obtain multiple images with overlapping portions;

[0012] Based on the preset oil and gas facility integrity, the multiple images with overlapping parts are filtered to obtain multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

[0013] Further, the remote sensing image is segmented to obtain multiple images with overlapping portions, including:

[0014] The remote sensing image is input into a pre-trained segmentation model to obtain different segmentation regions and the confidence scores of oil and gas facilities corresponding to different segmentation regions.

[0015] The degree of overlap between different segmented regions is determined based on the confidence level of oil and gas facilities corresponding to different segmented regions.

[0016] The different segmented regions are segmented using the segmentation overlap degree to obtain multiple images with overlapping parts.

[0017] Further, based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image, an oil and gas facility identification dataset is generated, including:

[0018] A first associated sample set is generated based on each oil and gas facility image and its corresponding first associated ground feature image;

[0019] Determine the second associated feature image corresponding to each oil and gas facility image, wherein the second associated feature image is a feature image with the same work area type as the oil and gas facility image selected from all acquired associated feature images;

[0020] A second associated sample set is generated based on each oil and gas facility image and its corresponding second associated ground feature image;

[0021] An oil and gas facility identification dataset is generated based on the first and second associated sample sets.

[0022] Furthermore, using the oil and gas facility identification dataset, an oil and gas facility identification model is generated, including:

[0023] A multi-input machine learning model is constructed, comprising an extraction unit, a fusion unit, and a recognition unit. The extraction unit is used to extract features of oil and gas facilities from images of oil and gas facilities and to extract features of associated land features from images of associated land features. The fusion unit is used to generate fused features based on the features of oil and gas facilities and the features of associated land features. The recognition unit is used to recognize the fused features.

[0024] The multi-input machine learning model is trained using the oil and gas facility identification dataset.

[0025] The multi-input machine learning model that has been trained is adjusted to a single-input machine learning model, and the single-input machine learning model is determined as the oil and gas facility identification model.

[0026] A second aspect of this specification provides a method for identifying oil and gas facilities, the method comprising:

[0027] The image to be identified is input into the oil and gas facility identification model to obtain the oil and gas facility identification result.

[0028] A third aspect of this specification provides an apparatus for generating an oil and gas facility identification model, the apparatus comprising:

[0029] The acquisition module is used to acquire multiple images of oil and gas facilities and the first associated ground feature image corresponding to each image of an oil and gas facility;

[0030] The first generation module is used to generate an oil and gas facility identification dataset based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

[0031] The second generation module is used to take images of oil and gas facilities and their associated ground features as input, and oil and gas facilities as output, and generate an oil and gas facility identification model using the oil and gas facility identification dataset.

[0032] A fourth aspect of this specification provides an oil and gas facility identification device, the identification device comprising:

[0033] The recognition module is used to input the image to be recognized into the oil and gas facility recognition model to obtain the oil and gas facility recognition result.

[0034] The fifth aspect of this specification provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the computer program, when run by the processor, executes instructions for the method of generating an oil and gas facility identification model and the method of identifying oil and gas facilities as described in any of the foregoing embodiments.

[0035] A sixth aspect of this specification provides a computer storage medium having a computer program stored thereon, wherein the computer program, when run by a processor of a computer device, executes instructions for the method of generating an oil and gas facility identification model and the method of identifying oil and gas facilities as described in any of the foregoing embodiments.

[0036] The seventh aspect of this specification provides a computer program product, which includes a computer program that, when run by a processor of a computer device, executes instructions for the method of generating an oil and gas facility identification model and the method of identifying oil and gas facilities as described in any of the foregoing embodiments.

[0037] The method for generating an oil and gas facility identification model, the method for identifying oil and gas facilities, and the apparatus provided in this specification link images of oil and gas facilities with images of associated ground features. This provides samples that combine the characteristics of the oil and gas facilities themselves with the characteristics of the associated ground features. The oil and gas facility identification dataset constructed based on these samples can provide multi-dimensional reference information for model training, enabling the model to learn not only the characteristics of oil and gas facilities but also to pay attention to the environmental information in which the oil and gas facilities are located, thereby improving the model's oil and gas facility identification performance.

[0038] To make the above and other objects, features and advantages of this specification more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 A flowchart illustrating the method for generating an oil and gas facility identification model according to an embodiment of this specification is shown;

[0041] Figure 2 This specification illustrates a flowchart of an embodiment for acquiring images of oil and gas facilities and associated ground features;

[0042] Figure 3 A flowchart illustrating the segmentation of remotely sensed images according to an embodiment of this specification is shown;

[0043] Figure 4 This specification illustrates a flowchart of how an oil and gas facility identification dataset is generated according to an embodiment of the present specification.

[0044] Figure 5 A flowchart illustrating the generation of an oil and gas facility identification model according to an embodiment of this specification is shown;

[0045] Figure 6 A flowchart illustrating the training of the oil and gas facility identification model in an embodiment of this specification is shown;

[0046] Figure 7 A structural diagram of the apparatus for generating an oil and gas facility identification model according to an embodiment of this specification is shown;

[0047] Figure 8 A structural diagram of a computer device according to an embodiment of this specification is shown.

[0048] Explanation of symbols in the attached drawings:

[0049] 710. Acquisition Module;

[0050] 720. First generation module;

[0051] 730. Second generation module;

[0052] 802. Computer equipment;

[0053] 804, Processor;

[0054] 806. Memory;

[0055] 808. Drive mechanism;

[0056] 810. Input / Output Module;

[0057] 812. Input devices;

[0058] 814. Output devices;

[0059] 816. Presentation equipment;

[0060] 818. Graphical User Interface;

[0061] 820. Network interface;

[0062] 822. Communication link;

[0063] 824. Communication bus. Detailed Implementation

[0064] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0065] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0066] This specification provides the operational steps of the methods described in the embodiments or flowcharts, but based on conventional or non-inventive labor, more or fewer operational steps may be included. The order of steps listed in the embodiments is merely one possible execution order among many and does not represent the only possible execution order. In actual system or device products, the methods shown in the embodiments or drawings can be executed sequentially or in parallel.

[0067] It should be noted that the acquisition, storage, use, and processing of data in the technical solutions of the embodiments of this specification all comply with the relevant provisions of national laws and regulations.

[0068] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0069] In recent years, with in-depth research on satellite remote sensing detection of oil and gas facilities, the applicant has discovered a certain connection between oil and gas facilities and their surrounding environmental features. For example, well sites are typically located in flat areas, surrounded by low vegetation, deserts, farmland, etc., and are always accessible by roads. During manual interpretation and identification, interpreters also utilize these environmental features in their analysis. Therefore, this invention, based on the aid of environmental features in oil and gas facility identification, establishes a method for identifying oil and gas facilities that integrates environmental features with machine learning algorithms. This method not only learns the features of the oil and gas facilities themselves but also incorporates surrounding environmental features, significantly improving the accuracy of oil and gas facility identification.

[0070] In one embodiment of this specification, a method for generating an oil and gas facility identification model is provided to solve the problem of poor identification performance of oil and gas facilities.

[0071] Specifically, such as Figure 1As shown, the method for generating the oil and gas facility identification model includes:

[0072] Step 110: Obtain multiple images of oil and gas facilities and the first associated ground feature image corresponding to each image of an oil and gas facility;

[0073] Step 120: Generate an oil and gas facility identification dataset based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

[0074] Step 130: Using the oil and gas facility images and their associated ground feature images as input, and the oil and gas facilities as output, an oil and gas facility identification model is generated using the oil and gas facility identification dataset.

[0075] This embodiment links images of oil and gas facilities with images of associated land features, providing samples that combine the characteristics of the oil and gas facilities themselves with the characteristics of associated land features. The oil and gas facility identification dataset constructed based on these samples can provide multi-dimensional reference information for model training, enabling the model to learn not only the characteristics of oil and gas facilities but also to pay attention to the environmental information in which the oil and gas facilities are located, thereby improving the model's oil and gas facility identification performance.

[0076] It should be understood that oil and gas facilities can generally be divided into well site facilities and station facilities. Well site facilities include drilling platforms, wellheads, wellhead equipment, cranes, and pipelines, while station facilities include oil and gas processing plants, oil storage tank groups, compressor stations, oil tanks, and gas storage tanks. These oil and gas facilities are usually located in flat areas, surrounded by low vegetation, bare ground, deserts, farmland, etc., and have access roads leading to the well site or station. Additionally, the well site is typically whiter than the surrounding environment.

[0077] In order to make reasonable use of the information of the surrounding environment of oil and gas facilities, after studying the relationship between oil and gas facilities and surrounding related features, the images of oil and gas facilities and related features are combined into sample pairs and used together to train the machine learning model. During the training process, the oil and gas facilities in the images are used as labeled targets, and the related feature images are used as auxiliary features. Then, the images of oil and gas facilities and their related features are used as inputs, and the identification results of oil and gas facilities are used as outputs. After multiple iterations, the model training can be completed.

[0078] In some embodiments of this specification, the machine learning model may employ the YOLO series of algorithms, such as YOLOv3, YOLOv4, YOLOv5, YOLOX, YOLOv6, YOLOv7, etc., or other object detection algorithms, which are not limited in this specification.

[0079] In one embodiment of this specification, as Figure 2As shown, multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image are acquired, including:

[0080] Step 210: Obtain remote sensing images containing oil and gas facilities and related features;

[0081] Step 220: Segment the remote sensing image to obtain multiple images with overlapping portions;

[0082] Step 230: Based on the preset oil and gas facility integrity, the multiple images with overlapping parts are filtered to obtain multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

[0083] This embodiment segments the remote sensing image into multiple images with overlapping parts, which can better preserve the integrity of the oil and gas facility images. The overlapping images are then filtered according to the integrity of the oil and gas facilities, ensuring the quality of the oil and gas facility images and the corresponding related ground feature images.

[0084] In some embodiments of this specification, after obtaining satellite remote sensing images, the images are preprocessed. Preprocessing includes initial screening and segmentation. Initial screening primarily removes remote sensing images that do not meet the criteria, such as irrelevant images or images of poor quality (e.g., distorted, damaged, etc.). Segmentation is used to crop the remote sensing images into smaller-scale images to reduce the iterative transport load during model training. The size of the segmented images is determined according to actual needs, for example, it can be 1024*768 pixels; this specification does not impose a limitation.

[0085] During segmentation, to protect the integrity of oil and gas facility images, the segmented images should have some overlap. This ensures that even if the previous image only retains a partial outline of the oil and gas facility, the subsequent image can still retain the complete outline. This overlapping segmentation method prevents the loss of complete oil and gas facility samples due to each image retaining only a portion of the facility's outline. Furthermore, since associated feature images are primarily used to assist model training and provide reference features, and these images may contain multiple features, reflecting overall environmental information, ensuring the completeness of associated features is not required during segmentation.

[0086] In some embodiments of this specification, the segmented images can be filtered according to a preset oil and gas facility integrity. The preset oil and gas facility integrity can be the percentage of the structure retained by the oil and gas facility, or it can be the result of the calculation of the symmetry index of the oil and gas facility. This specification does not limit it.

[0087] In some embodiments of this specification, data augmentation is performed on oil and gas facility images and associated feature images. The Mosaic data augmentation method is used to improve the diversity of oil and gas facility images and associated feature images and ensure that the oil and gas facility images and associated feature images always maintain a corresponding relationship. At the same time, edge enhancement is also required for associated feature images to highlight the outline information of associated features, which helps to better focus on key features such as well site roads.

[0088] The edge enhancement process for associated ground feature images is as follows:

[0089] S1. Apply Gaussian filtering to the associated ground feature image. Gaussian filtering can smooth details in the image and suppress noise, while highlighting ground features and maximizing the accuracy of the spatial rate of change of image brightness obtained from gradient calculation, making subsequent edge detection more stable.

[0090] S2. Perform gradient calculation on the Gaussian-filtered associated ground feature image to obtain a gradient image. The gradient image can intuitively reflect the pixel value changes of the associated ground feature image.

[0091] S3. Perform non-maximum suppression on the gradient image. Non-maximum suppression preserves pixels with the largest local gradients to refine edges.

[0092] S4. Filter the image after non-maximum suppression. Pixels are divided into strong edges, weak edges, and non-edges by setting a dual threshold. The gradient magnitude of strong edge pixels is greater than the first threshold, the gradient magnitude of weak edge pixels is between the first and second thresholds, and the gradient magnitude of non-edge pixels is less than the second threshold. The first threshold is greater than the second threshold. Strong edges are used to highlight the main feature pixels, while weak edges and non-edges are used to distinguish other pixels. Weak edges usually represent pixels that are a mixture of the main feature and the background, while non-edges represent background pixels.

[0093] It should be understood that strong edges refer to pixels with very large gradient values, which are generally considered reliable edge pixels. These pixels have gradient values ​​above a first threshold, indicating that they are the most significant and reliable edge parts in the image, meaning the main feature is intact and the pixels are prominent. Weak edges refer to pixels with gradient values ​​between the first and second thresholds. These pixels have lower gradient values ​​but are still considered part of an edge, just with lower reliability. Weak edge pixels have gradient values ​​above the second threshold but below the first threshold, indicating that they may be part of an edge, but further confirmation is needed; that is, they may be a mixture of the main feature and a few background pixels or a mixture of a large background and a few main feature pixels, requiring interactive screening for usable edges. Non-edges refer to pixels with very small gradient values, which are generally considered background noise or non-edge parts. These pixels have gradient values ​​below the second threshold, indicating that they are unlikely to be edge pixels and are therefore marked as non-edges; this part will not be used. By distinguishing these three types of edges, useful information in the image can be extracted and processed more accurately, improving the accuracy and reliability of the processing results.

[0094] S5. Edge tracking is performed on the filtered images. The Hough transform is used to connect strong edge pixels into complete edge lines. The Hough transform first converts each point in the image into parameter space. Then, it iterates through each edge pixel in the image, accumulating the curves in its corresponding parameter space. Peak values ​​are found in the accumulator array, and the line parameters corresponding to these peak values ​​are converted back into image space to obtain the detected lines. The processed associated feature image, obtained by extracting edge information from the image using Canny edge enhancement, highlights the contours and structural features of associated features, facilitating subsequent oil and gas facility identification and analysis.

[0095] In one embodiment of this specification, as Figure 3 As shown, the remote sensing image is segmented to obtain multiple images with overlapping portions, including:

[0096] Step 310: Input the remote sensing image into the pre-trained segmentation model to obtain different segmentation regions and the confidence scores of oil and gas facilities corresponding to different segmentation regions;

[0097] Step 320: Determine the degree of overlap between different segmented regions based on the confidence levels of oil and gas facilities corresponding to different segmented regions;

[0098] Step 330: Use the segmentation overlap degree to segment the different segmentation regions to obtain multiple images with overlapping parts.

[0099] This embodiment takes into account that the number and quality of oil and gas facilities contained in different regions of the remote sensing image are different. Therefore, a pre-trained segmentation model is used to initially segment the entire remote sensing image to obtain different segmented regions and the confidence scores of oil and gas facilities corresponding to different segmented regions. The segmentation overlap of different segmented regions is flexibly determined based on the confidence scores, thus ensuring segmentation efficiency and effectiveness.

[0100] In some embodiments of this specification, a pre-trained UNet model is used as a segmentation model to identify potential facility areas. By setting the size of the segmentation area of ​​the UNet model, the initial segmentation of the remote sensing image is achieved, thereby obtaining several major segmentation areas. At the same time, the UNet model can also provide the confidence level of different segmentation areas. The confidence level can reflect the quality of different segmentation areas, that is, the accuracy and reliability of the oil and gas facilities contained in different segmentation areas.

[0101] After obtaining the confidence levels of different segmented regions, the segmentation overlap is determined based on the confidence levels. Regions with high confidence levels have higher segmentation overlap, and regions with low confidence levels have lower segmentation overlap. Confidence levels can be categorized as high, medium, and low confidence, or other categories may be used; this specification does not impose limitations on this. The segmentation overlap determines the size of the overlapping portion of the segmented images. For example, a segmentation overlap of 0.1 results in a 10% overlap area between the preceding and following segments; 0.2 results in a 20% overlap area; and 0.3 results in a 30% overlap area. By determining the segmentation overlap, it is possible to ensure that segmented regions with high confidence levels retain the integrity of oil and gas facilities to a greater extent, while ensuring that segmented regions with low confidence levels are segmented more quickly, thus improving segmentation efficiency.

[0102] In some embodiments of this specification, after obtaining the confidence levels of different segmented regions, the segmentation priority of different segmented regions can be determined based on the confidence levels. Regions with high confidence levels have higher segmentation priority, and regions with low confidence levels have lower segmentation priority. The order of further segmenting different regions is determined based on these priorities. Prioritizing the segmentation of high-priority regions allows for more efficient acquisition of high-quality oil and gas facility images and associated ground feature images. For example, by simultaneously segmenting multiple remote sensing images to obtain different segmented regions from multiple images, and determining the confidence levels of these segmented regions, the segmentation priority is determined after statistically analyzing the confidence levels of all segmented regions. Prioritizing the segmentation of high-confidence regions ensures that the efficiency of collecting oil and gas facility images and associated ground feature images is maintained while meeting sample size requirements.

[0103] In one embodiment of this specification, as Figure 4 As shown, an oil and gas facility identification dataset is generated based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image, including:

[0104] Step 410: Generate a first associated sample set based on each oil and gas facility image and its corresponding first associated ground feature image;

[0105] Step 420: Determine the second associated feature image corresponding to each oil and gas facility image, wherein the second associated feature image is a feature image with the same work area type as the oil and gas facility image selected from all the acquired associated feature images;

[0106] Step 430: Generate a second associated sample set based on each oil and gas facility image and the corresponding second associated ground feature image;

[0107] Step 440: Generate an oil and gas facility identification dataset based on the first associated sample set and the second associated sample set.

[0108] In this embodiment, oil and gas facility images and associated ground feature images are paired to generate a first associated sample set. At the same time, more associated ground feature images are matched to the oil and gas facility images according to the work area type to which the oil and gas facility belongs, generating a second associated sample set. Finally, the first associated sample set and the second associated sample set are combined to generate an oil and gas facility identification dataset, which greatly enriches the sample size of the oil and gas facility identification dataset.

[0109] In some embodiments of this specification, the background of the segmented oil and gas facility image may contain associated features. Therefore, an additional copy of the oil and gas facility image with associated features in the background can be made, resulting in two oil and gas facility images. One oil and gas facility image is labeled and used as an oil and gas facility image sample, while the other oil and gas facility image does not need labeling and is directly used as an associated feature image sample. If the background of the segmented oil and gas facility image does not contain suitable associated features, an image containing associated features can be selected from its surrounding images as its corresponding associated feature image. Through the above method, a set of oil and gas facility image samples and a set of associated feature image samples can be obtained, forming the first associated sample set. Each oil and gas facility image in the oil and gas facility image sample set has a clear correspondence with an associated feature image in the associated feature image sample set.

[0110] In other embodiments of this specification, to enrich the number of associated samples, images of associated features are filtered from the acquired images of different work areas, based on the work area type where the images of different oil and gas facilities are located. The filtered images of associated features are then paired with images of oil and gas facilities to form new sample pairs. A new set of associated samples (i.e., a second set of associated samples) is then formed based on a large number of new sample pairs. This allows for the reuse of associated feature images from different work areas belonging to the same work area type. Work area types can be divided into well site areas, industrial areas, and complex industrial areas. Well site areas typically have most oil and gas facilities that are well site facilities, and there are no large buildings nearby. Industrial areas usually have large stations, abundant access roads forming a road network, and are equipped with oil pipelines. Complex industrial areas are characterized by a combination of multiple small station facilities, and there are no large buildings nearby. Therefore, the distribution of oil and gas facilities can also be used as an auxiliary feature for training the model.

[0111] In one embodiment of this specification, as Figure 5 As shown, using the oil and gas facility identification dataset, an oil and gas facility identification model is generated, including:

[0112] Step 510: Construct a multi-input machine learning model;

[0113] The multi-input machine learning model includes an extraction unit, a fusion unit, and a recognition unit. The extraction unit is used to extract oil and gas facility features from oil and gas facility images and related feature features from related feature images. The fusion unit is used to generate fused features based on the oil and gas facility features and the related feature features. The recognition unit is used to recognize the fused features.

[0114] Step 520: Use the oil and gas facility identification dataset to train the multi-input machine learning model;

[0115] Step 530: Adjust the multi-input machine learning model that has been trained to a single-input machine learning model, and determine the single-input machine learning model as the oil and gas facility identification model.

[0116] This embodiment uses a multi-input machine learning model to process oil and gas facility images and related ground feature images, and learns the fusion features of oil and gas facility features and related ground feature features. Then, the trained model is adjusted to a single-input machine learning model to obtain an oil and gas facility identification model, which provides a foundation for high-quality oil and gas facility identification.

[0117] It should be understood that the oil and gas facility identification dataset contains separate sets of image samples of oil and gas facilities and images of associated ground features, such as... Figure 6As shown, when inputting into the machine learning model, images of oil and gas facilities and related land features are input into different input terminals of the model. The model's extraction unit extracts features from the two types of images separately to obtain oil and gas facility features and related land feature features. Then, the model's fusion unit fuses these two features, for example, by splicing oil and gas facility features and related land feature features to obtain fused features. After that, the model's recognition unit recognizes the fused features and provides predicted oil and gas facility information. The predicted information is then compared with the real information to calculate the loss, and the parameters of each unit are updated. After repeated iterations, the training of the model can be completed.

[0118] After training, the multi-input machine learning model cannot be directly used for oil and gas facility identification. It needs to be adjusted to a single-input model to meet common single-input identification tasks. This adjusted single-input model is the oil and gas facility identification model. Because the oil and gas facility identification model learns both oil and gas facility features and associated land feature features simultaneously, it exhibits excellent identification performance in remote sensing image-based oil and gas facility identification. When faced with similar industrial sites, associated land feature features can assist the model in making judgments, thereby avoiding misidentification of industrial sites without associated land feature features as oil and gas facilities.

[0119] In one embodiment of this specification, an oil and gas facility identification method is also provided to solve the problem of poor identification effect of oil and gas facilities.

[0120] Specifically, methods for identifying oil and gas facilities include:

[0121] The image to be identified is input into the oil and gas facility identification model to obtain the oil and gas facility identification result.

[0122] The images to be identified can be remote sensing images of different sizes. When faced with large images, the model can divide the large images into blocks by using a sliding window, thereby identifying all oil and gas facilities to the greatest extent possible.

[0123] Based on the same inventive concept, this specification also provides an apparatus for generating an oil and gas facility identification model, as described in the following embodiments. Since the principle behind the apparatus for generating an oil and gas facility identification model is similar to that of the method for generating an oil and gas facility identification model, the implementation of the apparatus for generating the oil and gas facility identification model can refer to the method for generating an oil and gas facility identification model; repeated details will not be elaborated further.

[0124] Specifically, such as Figure 7 As shown, the device for generating an oil and gas facility identification model includes:

[0125] The acquisition module 710 is used to acquire multiple oil and gas facility images and a first associated ground feature image corresponding to each oil and gas facility image;

[0126] The first generation module 720 is used to generate an oil and gas facility identification dataset based on the plurality of oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

[0127] The second generation module 730 is used to take images of oil and gas facilities and their associated ground features as input, and oil and gas facilities as output, and generate an oil and gas facility identification model using the oil and gas facility identification dataset.

[0128] Based on the same inventive concept, this specification also provides an oil and gas facility identification device, as described in the following embodiments. Since the principle of the oil and gas facility identification device in solving the problem is similar to that of the oil and gas facility identification method, the implementation of the oil and gas facility identification device can refer to the oil and gas facility identification method; repeated details will not be elaborated further.

[0129] Specifically, oil and gas facility identification devices include:

[0130] The recognition module is used to input the image to be recognized into the oil and gas facility recognition model to obtain the oil and gas facility recognition result.

[0131] The method for generating an oil and gas facility identification model, the method for identifying oil and gas facilities, and the apparatus provided in this specification link images of oil and gas facilities with images of associated ground features. This provides samples that combine the characteristics of the oil and gas facilities themselves with the characteristics of the associated ground features. The oil and gas facility identification dataset constructed based on these samples can provide multi-dimensional reference information for model training, enabling the model to learn not only the characteristics of oil and gas facilities but also to pay attention to the environmental information in which the oil and gas facilities are located, thereby improving the model's oil and gas facility identification performance.

[0132] In one embodiment of this specification, a computer device is also provided for implementing the methods described in any of the above embodiments, such as... Figure 8The diagram illustrates the structure of a computer device according to an embodiment of this specification. The computer device 802 may include one or more processors 804, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 802 may also include any memory 806 for storing information of any kind, such as code, settings, data, etc. Non-limitingly, for example, the memory 806 may include any type of RAM, any type of ROM, flash memory, hard disk, optical disk, etc. More generally, any memory can use any technology to store information. Furthermore, any memory can provide volatile or non-volatile retention of information. Furthermore, any memory can represent a fixed or removable component of the computer device 802. In one case, when the processor 804 executes associated instructions stored in any memory or combination of memories, the computer device 802 can perform any operation of the associated instructions. The computer device 802 also includes one or more drive mechanisms 808 for interacting with any memory, such as hard disk drive mechanisms, optical disk drive mechanisms, etc.

[0133] Computer device 802 may also include an input / output module 810 (I / O) for receiving various inputs (via input device 812) and providing various outputs (via output device 814). A specific output mechanism may include a presentation device 816 and an associated graphical user interface (GUI) 818. In other embodiments, the input / output module 810 (I / O), input device 810, and output device 814 may be omitted, and the device may function solely as a computer device within a network. Computer device 802 may also include one or more network interfaces 820 for exchanging data with other devices via one or more communication links 822. One or more communication buses 824 couple the components described above together.

[0134] Communication link 822 can be implemented in any way, such as via a local area network, a wide area network (e.g., the Internet), a point-to-point connection, or any combination thereof. Communication link 822 may include any combination of hardwired links, wireless links, routers, gateway functions, name servers, etc., governed by any protocol or combination of protocols.

[0135] Corresponding to Figures 1 to 6 In addition to the methods described above, embodiments of this specification also provide a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the methods described above.

[0136] This specification also provides computer-readable instructions, wherein when a processor executes the instructions, the program therein causes the processor to perform the following... Figures 1 to 6 The method shown.

[0137] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0138] It should also be understood that, in the embodiments of this specification, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this specification generally indicates that the preceding and following related objects have an "or" relationship.

[0139] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this specification can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this specification.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the several embodiments provided in this specification, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and 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 couplings or direct couplings or communication connections shown or discussed may be indirect couplings or communication connections through some interfaces, devices, or units, or they may be electrical, mechanical, or other forms of connection.

[0142] 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 the embodiments described in this specification, depending on actual needs.

[0143] Furthermore, the functional units in the various embodiments of this specification 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. The integrated unit can be implemented in hardware or as a software functional unit.

[0144] If the integrated unit 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 specification, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. 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 described in the various embodiments of this specification. 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.

[0145] This specification uses specific embodiments to illustrate the principles and implementation methods of this specification. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this specification. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this specification. Therefore, the content of this specification should not be construed as a limitation of this specification.

Claims

1. A method for generating an oil and gas facility identification model, characterized in that, The generation method includes: Acquire multiple images of oil and gas facilities and the first associated ground feature image corresponding to each oil and gas facility image; Based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image, an oil and gas facility identification dataset is generated. Using images of oil and gas facilities and their associated ground features as input, and the oil and gas facilities as output, an oil and gas facility identification model is generated using the oil and gas facility identification dataset.

2. The method as described in claim 1, characterized in that, Acquire multiple images of oil and gas facilities and the first associated ground feature image corresponding to each oil and gas facility image, including: Acquire remote sensing images containing oil and gas facilities and related features; The remote sensing image is segmented to obtain multiple images with overlapping portions; Based on the preset oil and gas facility integrity, the multiple images with overlapping parts are filtered to obtain multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image.

3. The method as described in claim 2, characterized in that, The remote sensing image is segmented to obtain multiple images with overlapping portions, including: The remote sensing image is input into a pre-trained segmentation model to obtain different segmentation regions and the confidence scores of oil and gas facilities corresponding to different segmentation regions. The degree of overlap between different segmented regions is determined based on the confidence level of oil and gas facilities corresponding to different segmented regions. The different segmented regions are segmented using the segmentation overlap degree to obtain multiple images with overlapping parts.

4. The method as described in claim 1, characterized in that, Based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image, an oil and gas facility identification dataset is generated, including: A first associated sample set is generated based on each oil and gas facility image and its corresponding first associated ground feature image; Determine the second associated feature image corresponding to each oil and gas facility image, wherein the second associated feature image is a feature image with the same work area type as the oil and gas facility image selected from all acquired associated feature images; A second associated sample set is generated based on each oil and gas facility image and its corresponding second associated ground feature image; An oil and gas facility identification dataset is generated based on the first and second associated sample sets.

5. The method as described in claim 1, characterized in that, Using the oil and gas facility identification dataset, an oil and gas facility identification model is generated, including: A multi-input machine learning model is constructed, comprising an extraction unit, a fusion unit, and a recognition unit. The extraction unit is used to extract features of oil and gas facilities from images of oil and gas facilities and to extract features of associated land features from images of associated land features. The fusion unit is used to generate fused features based on the features of oil and gas facilities and the features of associated land features. The recognition unit is used to recognize the fused features. The multi-input machine learning model is trained using the oil and gas facility identification dataset. The multi-input machine learning model that has been trained is adjusted to a single-input machine learning model, and the single-input machine learning model is determined as the oil and gas facility identification model.

6. A method for identifying oil and gas facilities, characterized in that, The identification method includes: The image to be identified is input into the oil and gas facility identification model generated using any one of claims 1 to 5 to obtain the oil and gas facility identification result.

7. A device for generating an oil and gas facility identification model, characterized in that, The generating apparatus includes: The acquisition module is used to acquire multiple images of oil and gas facilities and the first associated ground feature image corresponding to each image of an oil and gas facility; The first generation module is used to generate an oil and gas facility identification dataset based on the multiple oil and gas facility images and the first associated ground feature image corresponding to each oil and gas facility image. The second generation module is used to take images of oil and gas facilities and their associated ground features as input, and oil and gas facilities as output, and generate an oil and gas facility identification model using the oil and gas facility identification dataset.

8. An oil and gas facility identification device, characterized in that, The identification device includes: The recognition module is used to input the image to be recognized into the oil and gas facility recognition model generated by the method of any one of claims 1 to 5, and obtain the oil and gas facility recognition result.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method according to any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor of a computer device, it implements the method of any one of claims 1 to 6.

11. A computer program product, the computer program product comprising a computer program, characterized in that, When the computer program is executed by the processor of a computer device, it implements the method of any one of claims 1 to 6.

Images