Ground surface thickened oil steam injection pipe network extraction method and device based on high-definition image, electronic equipment and storage medium

Abstract

The invention relates to the technical field of pipeline identification, in particular to a surface heavy oil steam injection pipe network extraction method and device based on a high-definition image, electronic equipment and a storage medium, and the method comprises the steps: obtaining a to-be-identified oil field remote sensing image, inputting the to-be-identified oil field remote sensing image into a feature segmentation model, and obtaining a feature segmentation image; and analyzing the feature segmentation image based on edge detection and Hough transform to obtain a surface heavy oil steam injection pipeline detection result in the feature segmentation image. The feature segmentation model is established by taking the convolutional neural network as a trunk network and fusing multiple network structures, the feature representation capability is improved, the accurate feature segmentation image can be extracted from the to-be-identified oil field remote sensing image, and then the feature segmentation image is analyzed based on edge detection and Hough transform, so that the recognition accuracy of the oil field remote sensing image is improved. The accurate detection result of the surface heavy oil steam injection pipe in the feature segmentation image is obtained, and reliable data support is provided for management of the surface heavy oil steam injection pipe.

Description

Technical Field

[0001] This invention relates to the field of pipeline identification technology, specifically a method, apparatus, electronic device, and storage medium for extracting surface heavy oil steam injection pipeline networks based on high-definition images. Background Technology

[0002] Currently, in the field of oilfield surface engineering, when some old areas and newly built production capacity surface heavy oil steam injection pipelines have not been updated and put into storage in a timely manner, it is necessary to inspect and identify the surface heavy oil steam injection pipelines. Existing methods for inspecting and identifying surface heavy oil steam injection pipelines include: (1) Traditional physical detection methods are only applicable to the identification of buried pipelines and require manual execution, which has limited efficiency; (2) Traditional satellite positioning system mapping methods include: first, using GPS positioning technology, placing a handheld GPS device near or on the pipeline to collect data; second, using a total station to measure specific points on the pipeline, obtaining the three-dimensional coordinates of these points, and then calculating and converting them to obtain latitude and longitude coordinates. Other methods include combining RTK, laser rangefinders, or utilizing online map services. If GPS positioning equipment, total stations, RTK, laser rangefinders, etc., are used to identify heavy oil steam injection pipelines, it usually requires the assistance of a professional external team. Although the identification accuracy is high, the workload is large and the data collection cycle is long. If online map services are used for heavy oil steam injection pipeline identification, this method has the following problems: First, the acquisition accuracy depends on the resolution of the image data; in other words, the higher the resolution, the higher the acquisition accuracy. Second, manual point selection will introduce measurement errors. That is, if the screen error deviates from the pipeline by a few millimeters when selecting pipeline points on a 1:500 resolution map, it may deviate by 1-2 meters in the actual field. Therefore, although the work efficiency is high, the identification accuracy is easily affected by the image resolution and the accuracy of manual point selection. (3) Existing ground oil and gas pipeline detection methods based on algorithm models can detect and identify heavy oil injection pipelines on the surface of remote sensing images. However, compared with buildings, vehicles, vegetation and other land features, pipelines in oilfield remote sensing images have both small-scale local features, such as junctions, and large-scale features, such as long-distance oil gathering and transportation trunk lines. Therefore, pipelines have multi-scale features. As a result, oilfield remote sensing images contain pipeline targets with large differences in size, uneven density distribution, small proportion of pixels occupied by pipeline targets, and various types of land feature information. Existing single neural network models (such as the general CNN convolutional neural network) cannot improve the feature representation ability. Therefore, effective features cannot be extracted from oilfield remote sensing images without eliminating interference, which affects the accuracy of pipeline detection and identification. Summary of the Invention

[0003] This invention provides a method, apparatus, electronic device, and storage medium for extracting surface heavy oil steam injection pipeline networks based on high-definition images. It overcomes the shortcomings of the prior art and can effectively solve the problem that existing surface heavy oil steam injection pipeline detection and identification methods based on a single neural network model cannot improve the feature expression ability, resulting in the inability to extract effective features from oilfield remote sensing images and affecting the accuracy of pipeline detection and identification.

[0004] One of the technical solutions of this invention is achieved through the following measures: a method for extracting surface heavy oil steam injection pipeline networks based on high-definition imagery, comprising: The remote sensing image of the oilfield to be identified is acquired and input into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning on the fusion neural network model using several samples. Each sample in the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The surface heavy oil steam injection pipeline detection results are obtained by analyzing the feature segmentation image based on edge detection and Hough transform.

[0005] The following are further optimizations and / or improvements to the above-mentioned technical solution: The above also includes acquiring several remote sensing images of a designated area to be identified, obtaining the detection results of surface heavy oil steam injection pipelines for each remote sensing image to be identified, combining all the detection results of surface heavy oil steam injection pipelines according to coordinates, and extracting the surface heavy oil steam injection pipeline network image of the designated area.

[0006] The above feature segmentation model construction process includes: Acquire several historical oilfield remote sensing images and preprocess each historical oilfield remote sensing image, including binarization and image segmentation. The preprocessed historical oilfield remote sensing images were augmented, and all the augmented historical oilfield remote sensing images were divided into training sample set and test sample set according to the proportion. The R2AU-Net model is trained using a training sample set. A loss function is introduced during training. Training ends when the value of the loss function is stable, resulting in a feature segmentation model. The R2AU-Net model is a U-Net model that embeds a recurrent residual module and an attention module. The trained feature segmentation model is tested using a test set, the model parameters are optimized, and a feature segmentation model that meets the test evaluation requirements is output.

[0007] The above analysis of the feature segmentation image based on edge detection and Hough transform yields the detection results of surface heavy oil steam injection pipelines in the feature segmentation image, including: Edge detection operators are used to extract edges from feature segmentation images, and edge contour images are output. Hough transform is used to detect straight lines in edge contour images, and the results of surface heavy oil steam injection pipeline detection are output.

[0008] The above-mentioned method uses Hough transform to perform line detection on edge contour images and outputs the detection results of surface heavy oil steam injection pipelines, including: In the edge contour image, select any straight line, and transform the pixel coordinates of any point through which the line passes into the parameter space through parameter transformation to obtain an accumulator of a two-dimensional array. Iterate through all the points the line passes through and accumulate votes. Determine the cumulative value of all straight lines in the edge contour image, and select the straight lines that are greater than a set threshold as the detection result of the surface heavy oil steam injection pipeline.

[0009] The edge detection operator mentioned above is the Canny edge detection operator.

[0010] The second technical solution of the present invention is achieved through the following measures: a surface heavy oil steam injection pipeline network extraction device based on high-definition imagery, comprising: The segmentation unit acquires the remote sensing image of the oilfield to be identified and inputs it into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning the fusion neural network model using several samples. Each of the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The detection unit analyzes the feature segmentation image based on edge detection and Hough transform to obtain the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image.

[0011] The following are further optimizations and / or improvements to the above-mentioned technical solution: The above also includes a pipeline network image forming unit, which acquires several remote sensing images to be identified in a set area, obtains the detection results of surface heavy oil steam injection pipelines in each remote sensing image to be identified, combines all the detection results of surface heavy oil steam injection pipelines according to coordinates, and extracts the surface heavy oil steam injection pipeline network image of the set area.

[0012] The third technical solution of the present invention is achieved through the following measures: an electronic device, including a processor and a memory, wherein the memory stores a computer program, which is loaded and executed by the processor to implement the steps in the method for extracting surface heavy oil steam injection pipelines based on high-definition images.

[0013] The fourth technical solution of the present invention is achieved through the following measures: a storage medium storing a computer program that can be read by a computer, the computer program being configured to execute the steps in the method for extracting surface heavy oil steam injection pipelines based on high-definition images when running.

[0014] This invention establishes a feature segmentation model using a convolutional neural network as the backbone and incorporating multiple network structures to improve feature representation capabilities. This enables the extraction of accurate feature-segmented images from remote sensing images of oilfields to be identified. Further analysis of these segmented images based on edge detection and Hough transform yields accurate detection results for surface heavy oil steam injection pipelines, providing reliable data support for the management of these pipelines. Furthermore, based on the final results of the surface heavy oil steam injection pipeline detection, and following the rules of steam injection pipeline design, the invention interactively connects the surface pipelines represented by the detection results, forming a steam injection pipeline network for heavy oil production surface engineering. This eliminates the influence of non-steam injection pipelines and other straight-line features within the area, maintaining the accuracy and topological integrity of the steam injection pipeline network. Attached Figure Description

[0015] Appendix Figure 1 This is a schematic diagram of the implementation environment provided for one embodiment of the present invention.

[0016] Appendix Figure 2 This is a schematic diagram of a method for extracting heavy oil from a surface steam injection pipeline network, provided as an embodiment of the present invention.

[0017] Appendix Figure 3 This is a schematic diagram of a method for detecting surface heavy oil steam injection pipelines according to an embodiment of the present invention.

[0018] Appendix Figure 4 This is a schematic diagram of a method for extracting heavy oil from a surface steam injection pipeline network, as provided in one embodiment of the present invention.

[0019] Appendix Figure 5 This is a schematic diagram of a surface heavy oil steam injection pipeline extraction device provided in one embodiment of the present invention.

[0020] Appendix Figure 6 This is a schematic diagram of another surface heavy oil steam injection pipeline extraction device provided in an embodiment of the present invention. Detailed Implementation

[0021] The present invention is not limited to the following embodiments, and the specific implementation can be determined according to the technical solution of the present invention and the actual situation.

[0022] Those skilled in the art will understand that, unless specifically stated otherwise, in the embodiments of the present invention, a "module" or "unit" refers to a computer program or part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0023] In addition, in the embodiments of the present invention, "multiple" refers to two or more, and "first" and "second" are used to distinguish descriptions and should not be construed as implying relative importance.

[0024] Currently, compared to features like buildings, vehicles, and vegetation, pipelines in oilfield remote sensing images possess both small-scale local features (such as junctions) and large-scale features (such as long-distance oil gathering and transportation trunk lines). Therefore, pipelines exhibit multi-scale characteristics. Oilfield remote sensing images are characterized by significant differences in pipeline target size, uneven density distribution, a small proportion of pixels occupied by pipeline targets, and inclusion of various types of ground feature information. Existing single neural network models (such as the general CNN convolutional neural network) cannot improve feature representation capabilities and thus cannot eliminate interference, failing to extract effective features from oilfield remote sensing images. This affects the accuracy of pipeline detection and identification.

[0025] To improve the expressive power of features for detecting and identifying surface heavy oil steam injection pipelines in oilfield remote sensing images, this invention provides a method, apparatus, electronic device, and storage medium for extracting surface heavy oil steam injection pipeline networks based on high-definition images. With the goal of improving feature representation capabilities, a fusion neural network model is established, using a convolutional neural network as the backbone and incorporating multiple network structures. Through deep learning of the fusion neural network model, a feature segmentation model is obtained. The remote sensing image of the oilfield to be identified is acquired and input into the feature segmentation model to obtain a segmented image. The segmented image is then analyzed based on edge detection and Hough transform to obtain the detection results of surface heavy oil steam injection pipelines within the segmented image.

[0026] The method provided in this embodiment of the invention may involve artificial intelligence (AI) technology and may be implemented based on artificial intelligence technology, such as using deep learning to train a corresponding model using samples.

[0027] Deep learning (DL) specifically refers to machine learning based on deep neural network models and methods. It has developed from statistical machine learning, artificial neural network algorithms, and other algorithms, combined with the development of modern big data and high computing power. The most important technical feature of deep learning is its ability to automatically extract features.

[0028] The aforementioned machine learning and deep learning typically include techniques such as neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and instructional learning.

[0029] In deep learning, the loss function is used to predict the target value by comparing the predicted value with the target value. This is done by updating the weight vector of each layer of the neural network based on the difference between the two values ​​(usually with an initialization process before the first update, where parameters are pre-configured for each layer) until the network can predict the target value or a value very close to it. Therefore, deep learning requires pre-defining "how to compare the difference between the predicted value and the target value," which is the loss function.

[0030] As attached Figure 1 The diagram illustrates an implementation environment provided by an embodiment of the present invention. This implementation environment may include: training equipment and usage equipment.

[0031] Both the training equipment and the equipment used are computer devices; optionally, the computer device is a terminal device, such as a mobile phone, tablet computer, PC (Personal Computer) or other electronic devices; or, the computer device is a server, which can be a single server, a server cluster composed of multiple servers, or a cloud computing service center. This embodiment of the invention does not limit this.

[0032] Training equipment refers to computer equipment capable of training and learning neural networks. Optionally, the training equipment has the ability to acquire neural networks and train and learn them according to application requirements. For example, the training equipment acquires neural networks from other devices through a network and then trains them with training samples according to application requirements, so that the neural network has the ability to obtain feature segmentation images. Optionally, the training equipment has the ability to build neural networks. It can build neural networks itself according to application requirements and then train and learn them. For example, in order to obtain feature segmentation images from remote sensing images of oil fields to be identified, the training equipment builds its own neural network and then trains and learns it with samples according to application requirements.

[0033] The device used refers to a computer device that has the capability to use a neural network. Optionally, the device uses a graph neural network according to the application requirements and obtains the graph neural network from other devices through the network. For example, if the device has the capability to obtain a feature segmentation image, it can obtain a neural network that has been trained and learned to obtain the feature segmentation image from other devices through the network, and use the neural network to obtain the feature segmentation image corresponding to the remote sensing image of the oil field to be identified.

[0034] Based on this, the technical solution of the present invention will be described and explained below with reference to several examples.

[0035] Example 1: As shown in the attached document Figure 2 As shown, this embodiment of the invention discloses a method for extracting surface heavy oil steam injection pipeline networks based on high-definition imagery, including: Step S110: Obtain the remote sensing image of the oilfield to be identified and input it into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning the fusion neural network model using several samples. Each of the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. In this embodiment, the goal is to improve the representation ability of features by fusing neural network models. Therefore, a model is established with a convolutional neural network as the backbone network and incorporating multiple network structures. The specific structure can be, but is not limited to, the following: The fusion neural network model is the R2AU-Net model, which uses the U-Net model as the backbone network and incorporates embedded recurrent residual modules and attention modules. The U-Net model has good multi-scale characteristics, which can take into account both the detailed and global information of the image when extracting multi-scale features. The attention module can improve the feature representation ability of the model and enhance the attention to detailed information. The recurrent residual convolution structure can alleviate the gradient vanishing problem, and through the recurrent convolution operation, low-level features are repeatedly superimposed with high-level features, gradually enriching the expressive ability of low-level features, helping the network to better transmit and utilize information from different levels, thereby improving the feature representation ability.

[0036] The aforementioned fusion neural network model can be trained using semi-supervised learning or supervised deep learning methods.

[0037] Taking supervised learning as an example, the process of building a feature segmentation model includes: (1) Acquire several historical oilfield remote sensing images and preprocess each historical oilfield remote sensing image, including binarization and image segmentation; Specifically, binarization is the process of standardizing the pixel values ​​of an image to a range of 0 to 1 or -1 to 1. Specifically, image segmentation involves adjusting the size of the input image to match the expected size of the model's input layer. The input image size is typically 224x224 or 256x256. Specifically, data augmentation can help increase the diversity of training samples and improve the generalization ability of the model. Common data augmentation operations include random cropping, random flipping, rotation, scaling, etc.

[0038] (2) The preprocessed historical oilfield remote sensing images are augmented with data, and all the augmented historical oilfield remote sensing images are divided into training sample set and test sample set according to the proportion.

[0039] (3) The R2AU-Net model is trained using the training sample set. A loss function is introduced during training. When the value of the loss function is stable, the training is terminated to obtain the feature segmentation model. The R2AU-Net model is a U-Net model that embeds a cyclic residual module and an attention module. The loss function here can be, but is not limited to, cross-entropy.

[0040] (4) Test the trained feature segmentation model using the test set, optimize the model parameters of the feature segmentation model, and output the feature segmentation model that meets the test evaluation requirements.

[0041] The feature segmentation model established in this embodiment can extract features of different spatial scales and channel dimensions, coarse-scale features, and shallow low-level features from oilfield remote sensing images. These three features are then integrated by pixel addition to obtain a feature map with multi-scale spatial information and cross-channel attention, i.e., a feature segmentation image.

[0042] It should also be noted that the feature segmentation image includes not only the location features of the steam injection pipeline and the oilfield surface pipeline, but also the features of typical buildings such as surface heavy oil steam filling stations and oil production stations.

[0043] Step S120: Analyze the feature segmentation image based on edge detection and Hough transform to obtain the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image.

[0044] This invention discloses a method for extracting surface heavy oil steam injection pipeline networks based on high-definition imagery. By establishing a feature segmentation model with a convolutional neural network as the backbone and integrating multiple network structures, the representation capability of features is improved. This enables the extraction of accurate feature segmentation images from remote sensing images of oilfields to be identified. Then, based on edge detection and Hough transform, the feature segmentation images are analyzed to obtain accurate detection results of surface heavy oil steam injection pipelines in the feature segmentation images, providing reliable data support for the management of surface heavy oil steam injection pipelines.

[0045] Example 2: As shown in the attached document Figure 3As shown, the embodiment of the present invention is a further optimization of the above implementation, wherein the feature segmentation image is analyzed based on edge detection and Hough transform to obtain the detection result of the surface heavy oil steam injection pipeline in the feature segmentation image, including: Step S210: Use the edge detection operator to extract edges from the feature segmentation image and output the edge contour image.

[0046] The edge detection operator mentioned above is selected according to the actual situation. In this embodiment, the Canny edge detection operator is used as an example, and the corresponding edge extraction steps include: (1) Gaussian filtering is applied to the feature segmentation image to smooth the feature segmentation image while preserving key edge information; (2) Determine the intensity and direction of the edges by calculating the gradients of the feature segmentation image in the horizontal and vertical directions; (3) The gradient magnitude in the feature segmentation image is processed by the non-maximum suppression method to retain the pixel with the largest local gradient and suppress non-edge pixels; (4) Based on the set high and low thresholds, the pixels in the feature segmentation image are divided into strong edge pixels, weak edge pixels and non-edge pixels; (5) Use strong edge pixels to connect adjacent weak edge pixels to complete edge extraction.

[0047] Step S220: Use Hough transform to perform line detection on the edge contour image and output the detection results of the surface heavy oil steam injection pipeline.

[0048] In this embodiment, step S220 specifically includes: (1) Select any straight line in the edge contour image, and transform the pixel coordinates of any point through which the line passes into the parameter space through parameter transformation to obtain a two-dimensional array accumulator. Here, the pixel coordinates of any point along the line are transformed into the parameter space through parameter transformation, that is, the pixel coordinates are transformed from Cartesian coordinates to polar coordinates; (2) Traverse all the points passed by the line and accumulate the votes; (3) Determine the cumulative value of all straight lines in the edge contour image, and select the straight lines that are greater than the set threshold as the detection result of the surface heavy oil steam injection pipeline.

[0049] It should also be noted that after selecting a straight line that is greater than the set threshold, it is transformed back from the parameter space to the image space so that the detected straight line can be drawn in the original image.

[0050] Example 3: As shown in the attached document Figure 4 As shown, this embodiment of the invention discloses a method for extracting surface heavy oil steam injection pipeline networks based on high-definition imagery, including: Step S310: Obtain several remote sensing images of the designated area to be identified and input them into the feature segmentation model to obtain the corresponding feature segmentation image. The feature segmentation model is obtained by deep learning of the fusion neural network model using several samples. Each of the several samples includes an oilfield remote sensing image and a corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures.

[0051] Step S320: Analyze the feature segmentation image based on edge detection and Hough transform to obtain the detection result of the surface heavy oil steam injection pipeline in the feature segmentation image. Here, the detection result of the surface heavy oil steam injection pipeline is a straight line segment.

[0052] Step S330: Based on the knowledge and rules of steam injection pipeline network design for oil production surface engineering, the steam station and oil pumping unit in the feature segmentation image are used as the first and last ends of the steam injection pipeline network. The surface pipelines expressed by the surface heavy oil steam injection pipeline detection results are interactively connected to form the steam injection pipeline network for heavy oil production surface engineering.

[0053] This embodiment is based on the final results of the surface heavy oil steam injection pipeline detection. Based on the extracted straight segments, and in accordance with the rules of steam injection pipeline design, it semi-automatically or interactively constructs the steam injection pipeline network from the gas station to the oil production machine. This can eliminate the influence of other straight segments and features such as non-steam injection pipelines in the area, and maintain the accuracy and topological integrity of the steam injection pipeline network.

[0054] Example 4: As shown in the appendix Figure 5 As shown, this embodiment of the invention discloses a surface heavy oil steam injection pipeline network extraction device based on high-definition imagery, comprising: The segmentation unit acquires the remote sensing image of the oilfield to be identified and inputs it into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning the fusion neural network model using several samples. Each of the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The detection unit analyzes the feature segmentation image based on edge detection and Hough transform to obtain the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image.

[0055] Example 5: As shown in the attached document Figure 6 As shown, this embodiment of the invention discloses a surface heavy oil steam injection pipeline network extraction device based on high-definition imagery, comprising: The segmentation unit acquires the remote sensing image of the oilfield to be identified and inputs it into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning the fusion neural network model using several samples. Each of the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The detection unit analyzes the feature segmentation image based on edge detection and Hough transform to obtain the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image; The pipeline network image forming unit acquires several remote sensing images of a designated area to be identified, obtains the detection results of surface heavy oil steam injection pipelines for each remote sensing image to be identified, combines all the detection results of surface heavy oil steam injection pipelines according to coordinates, and extracts the surface heavy oil steam injection pipeline network image of the designated area.

[0056] Example 6: This embodiment of the invention discloses a storage medium storing a computer program that can be read by a computer. The computer program is configured to execute a method for extracting surface heavy oil steam injection pipelines based on high-definition images when it runs.

[0057] The aforementioned storage media may include, but are not limited to, USB flash drives, read-only memory, portable hard drives, magnetic disks, optical disks, and other media capable of storing computer programs.

[0058] Example 7: This embodiment of the invention discloses an electronic device, including a processor and a memory. The memory stores a computer program, which is loaded and executed by the processor to implement a method for extracting surface heavy oil steam injection pipelines based on high-definition images.

[0059] The processor described above can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. It can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The memory can include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, portable hard drives, magnetic disks, or optical disks.

[0060] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0061] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0062] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0063] The above technical features constitute the preferred embodiment of the present invention, which has strong adaptability and optimal implementation effect. Unnecessary technical features can be added or removed according to actual needs to meet the requirements of different situations.

[0064] The above content is only a specific embodiment of the present invention, which has strong adaptability and implementation effect. However, the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention. Therefore, equivalent changes made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for extracting surface heavy oil steam injection pipeline networks based on high-definition imagery, characterized in that, include: The remote sensing image of the oilfield to be identified is acquired and input into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning on the fusion neural network model using several samples. Each sample in the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The surface heavy oil steam injection pipeline detection results are obtained by analyzing the feature segmentation image based on edge detection and Hough transform.

2. The method for extracting surface heavy oil steam injection pipelines based on high-definition imagery according to claim 1, characterized in that, It also includes acquiring several remote sensing images of a designated area to be identified, obtaining the detection results of surface heavy oil steam injection pipelines for each remote sensing image to be identified, combining all the detection results of surface heavy oil steam injection pipelines according to coordinates, and extracting the surface heavy oil steam injection pipeline network image of the designated area.

3. The method for extracting surface heavy oil steam injection pipelines based on high-definition imagery according to claim 1 or 2, characterized in that, The feature segmentation model construction process includes: Acquire several historical oilfield remote sensing images and preprocess each historical oilfield remote sensing image, including binarization and image segmentation. The preprocessed historical oilfield remote sensing images were augmented, and all the augmented historical oilfield remote sensing images were divided into training sample set and test sample set according to the proportion. The R2AU-Net model is trained using a training sample set. A loss function is introduced during training. Training ends when the value of the loss function is stable, resulting in a feature segmentation model. The R2AU-Net model is a U-Net model that embeds a recurrent residual module and an attention module. The trained feature segmentation model is tested using a test set, the model parameters are optimized, and a feature segmentation model that meets the test evaluation requirements is output.

4. The method for extracting surface heavy oil steam injection pipelines based on high-definition imagery according to any one of claims 1 to 3, characterized in that, The analysis of the feature segmentation image based on edge detection and Hough transform yields the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image, including: Edge detection operators are used to extract edges from feature segmentation images, and edge contour images are output. Hough transform is used to detect straight lines in edge contour images, and the results of surface heavy oil steam injection pipeline detection are output.

5. The method for extracting surface heavy oil steam injection pipelines based on high-definition imagery according to claim 4, characterized in that, The method of using Hough transform to perform line detection on edge contour images and outputting the detection results of surface heavy oil steam injection pipelines includes: In the edge contour image, select any straight line, and transform the pixel coordinates of any point through which the line passes into the parameter space through parameter transformation to obtain an accumulator of a two-dimensional array. Iterate through all the points the line passes through and accumulate votes. Determine the cumulative value of all straight lines in the edge contour image, and select the straight lines that are greater than a set threshold as the detection result of the surface heavy oil steam injection pipeline.

6. The method for extracting surface heavy oil steam injection pipelines based on high-definition imagery according to claim 4, characterized in that, The edge detection operator is the Canny edge detection operator.

7. A surface heavy oil steam injection pipeline network extraction device based on high-definition imagery, employing the method described in any one of claims 1 to 6, characterized in that, include: The segmentation unit acquires the remote sensing image of the oilfield to be identified and inputs it into the feature segmentation model to obtain the feature segmentation image. The feature segmentation model is obtained by deep learning the fusion neural network model using several samples. Each of the several samples includes the remote sensing image of the oilfield and the corresponding feature segmentation image label. The fusion neural network model is a model that aims to improve the representation ability of features, with a convolutional neural network as the backbone network and a combination of multiple network structures. The detection unit analyzes the feature segmentation image based on edge detection and Hough transform to obtain the detection results of the surface heavy oil steam injection pipeline in the feature segmentation image.

8. The surface heavy oil steam injection pipeline extraction device based on high-definition imagery according to claim 7, characterized in that, It also includes a pipeline image forming unit, which acquires several remote sensing images to be identified in a set area, obtains the detection results of surface heavy oil steam injection pipelines in each remote sensing image to be identified, combines all the detection results of surface heavy oil steam injection pipelines according to coordinates, and extracts the surface heavy oil steam injection pipeline network image of the set area.

9. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which is loaded and executed by the processor to implement the steps of the method as claimed in any one of claims 1 to 6.

10. A storage medium, characterized in that, The storage medium stores a computer program that can be read by a computer, the computer program being configured to execute the steps of the method as described in any one of claims 1 to 6 when it is run.

Images