Calculation method and system for thermal radiation of multi-angle jet fire in natural gas pipelines

Abstract

The present invention provides a method and system for calculating the thermal radiation of multi-angle jet fires in a natural gas pipeline. The method comprises the following steps: obtaining visible light images of jet fires at different angles at the current moment and the jet angles of the jet fires; determining the segmentation window size of any jet fire visible light image based on the jet angle of the corresponding jet fire and performing segmentation; binarizing each visible light sub-image to obtain an initial segmented image; inputting the initial segmented image and any jet fire visible light image into an image segmentation model to obtain a segmented image of the jet fire visible light image; performing multi-view stereo matching based on the segmented images of each jet fire visible light image to obtain a depth map of the jet fire; and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire, thereby determining the surface area of ​​the jet fire. The thermal radiation at any position is calculated based on the surface area of ​​the jet fire, thereby improving the calculation accuracy of the thermal radiation.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal radiation calculation, and in particular to a method and system for calculating thermal radiation of multi-angle jet fires in a natural gas pipeline. Background Art

[0002] During service, natural gas pipelines may rupture due to third-party vandalism, corrosion, human error, or poor construction quality, leading to natural gas leaks and potentially serious consequences. When natural gas leaks from a pipeline, it forms a high-speed jet that spreads. If the leak ignites, it produces a jet flame. Jet flames are the most common and representative type of gas pipeline fire. They primarily affect the surrounding environment through thermal radiation, causing surrounding objects to burn and deform, and even resulting in casualties. Therefore, accurate calculation of thermal radiation is crucial.

[0003] Flame surface area is a key parameter in thermal radiation calculations. However, since flames are often irregular in shape and the different injection angles of leak ports significantly affect the shape of the injected flame, this makes it difficult to estimate the flame surface area. Therefore, a method that can accurately extract flame morphology to improve the accuracy of thermal radiation calculations is urgently needed. Summary of the Invention

[0004] The present invention provides a method and system for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline, which are used to improve the calculation accuracy of thermal radiation.

[0005] The present invention provides a method for calculating heat radiation of multi-angle jet fire in a natural gas pipeline, comprising:

[0006] Obtaining visible light images of the jet fire at different angles at the current moment, and determining the jet angle of the jet fire in each visible light image of the jet fire;

[0007] For any jet fire visible light image, determining a segmentation window size of the jet fire visible light image based on a jet angle of the corresponding jet fire, and segmenting the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images;

[0008] Determine the segmentation threshold of each visible light sub-image respectively, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image;

[0009] Inputting the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model;

[0010] performing multi-view stereo matching based on segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire;

[0011] Based on the three-dimensional point cloud of the jet fire, the surface area of ​​the jet fire is determined, and the thermal radiation at any position is calculated based on the surface area of ​​the jet fire.

[0012] According to a method for calculating thermal radiation of multi-angle jet fires in a natural gas pipeline provided by the present invention, determining the size of a segmentation window of a visible light image of any jet fire based on the jet angle of the corresponding jet fire includes:

[0013] The segmentation window size of any jet fire visible light image is determined based on the following formula:

[0014]

[0015] in, is the segmentation window size of any jet fire visible light image, is the minimum window size, is the maximum window size, is the jet fire's injection angle, It is the maximum spray angle of jet fire.

[0016] According to a method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline provided by the present invention, the image segmentation model is constructed based on an encoder-decoder model;

[0017] Inputting the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model, comprising:

[0018] performing feature extraction on the initial segmented image and any one of the jet fire visible light images based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map;

[0019] performing feature extraction on the infrared image registered with any jet fire visible light image based on multiple downsampling layers of different scales of the second encoder in the image segmentation model to obtain an infrared feature map;

[0020] Based on the feature fusion module in the image segmentation model, the visible light feature map and the infrared feature map are fused to obtain a jet fire fusion feature map;

[0021] The jet fire fusion feature map is decoded based on multiple upsampling layers of different scales in a decoder in the image segmentation model to obtain the segmented image.

[0022] According to a method for calculating thermal radiation of a multi-angle jet fire in a natural gas pipeline provided by the present invention, feature extraction is performed on the initial segmented image and any jet fire visible light image based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map, including:

[0023] Performing feature extraction on the initial segmented image and any one of the jet fire visible light images based on the first downsampling layer to obtain an initial segmentation feature map and an initial visible light feature map;

[0024] Based on the feature value of each feature point in the initial segmentation feature map, weighting the feature value of the corresponding feature point in the initial visible light feature map to obtain a fused visible light feature map;

[0025] Feature extraction is performed on the fused visible light feature map based on the remaining downsampling layers to obtain the visible light feature map.

[0026] According to a method for calculating thermal radiation from multi-angle jet fires in a natural gas pipeline provided by the present invention, the downsampling layer of the second encoder includes multiple convolutional layers and a fusion layer, wherein the multiple convolutional layers use convolution kernels in different directions; the fusion layer is used to perform weighted fusion on the feature maps output by the multiple convolutional layers based on the weights corresponding to the multiple convolutional layers; wherein the weights corresponding to the multiple convolutional layers are determined based on the jet angles of the corresponding jet fires.

[0027] According to a method for calculating thermal radiation of multi-angle jet fires in a natural gas pipeline provided by the present invention, any upsampling layer is jump-connected to a downsampling layer of the same scale in the first encoder and a downsampling layer of the same scale in the second encoder.

[0028] According to a method for calculating thermal radiation of multi-angle jet fires in a natural gas pipeline provided by the present invention, the segmentation threshold of any visible light sub-image is the average value of the pixel value of each pixel point in the any visible light sub-image.

[0029] The present invention also provides a natural gas pipeline multi-angle jet fire heat radiation calculation system, comprising:

[0030] an image acquisition unit, configured to acquire visible light images of the jet fire at different angles at the current moment, and determine the jet angle of the jet fire in each visible light image of the jet fire;

[0031] a segmentation unit, configured to determine, for any jet fire visible light image, a segmentation window size of the jet fire visible light image based on a jet angle of the corresponding jet fire, and segment the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images;

[0032] A binarization unit is used to determine the segmentation threshold of each visible light sub-image respectively, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image;

[0033] a segmentation unit, configured to input the initial segmented image and the any jet fire visible light image into an image segmentation model, and obtain a segmented image of the any jet fire visible light image output by the image segmentation model;

[0034] a 3D reconstruction unit, configured to perform multi-view stereo matching based on the segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and perform 3D reconstruction based on the depth map of the jet fire to obtain a 3D point cloud of the jet fire;

[0035] The heat radiation calculation unit is used to determine the surface area of ​​the jet fire based on the three-dimensional point cloud of the jet fire, and calculate the heat radiation at any position based on the surface area of ​​the jet fire.

[0036] The present invention also provides an electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the program, any of the above-described methods for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline is implemented.

[0037] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned methods for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline.

[0038] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned methods for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline.

[0039] The method and system for calculating the thermal radiation of multi-angle jet fires in a natural gas pipeline provided by the present invention obtain visible light images of jet fires at different angles at the current moment and determine the jet angle of the jet fire in each visible light image of the jet fire; for any visible light image of the jet fire, based on the jet angle of the corresponding jet fire, determine the segmentation window size of any visible light image of the jet fire, and segment any visible light image of the jet fire based on the segmentation window size to obtain multiple visible light sub-images; determine the segmentation threshold of each visible light sub-image respectively, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; and convert the initial segmented image into a binary image. and any jet fire visible light image are input into an image segmentation model to obtain a segmented image of any jet fire visible light image output by the image segmentation model; multi-view stereo matching is performed based on the segmented images of each jet fire visible light image to obtain a depth map of the jet fire, and three-dimensional reconstruction is performed based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; based on the three-dimensional point cloud of the jet fire, the surface area of ​​the jet fire is determined, and the thermal radiation at any position is calculated based on the surface area of ​​the jet fire. This scheme can be applied to the analysis of jet fire field test results as a data reference for optimizing the theoretical model for jet fire thermal radiation calculation, thereby improving the calculation accuracy of thermal radiation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 It is a flow chart of the method for calculating the thermal radiation of multi-angle jet fire in a natural gas pipeline provided by the present invention;

[0042] Figure 2 It is a structural diagram of the image segmentation model provided by the present invention;

[0043] Figure 3 This is a schematic diagram of the structure of the device for calculating the thermal radiation of multi-angle jet fire in a natural gas pipeline provided by the present invention;

[0044] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0045] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0046] Figure 1 The figure is a flow chart of the method for calculating the thermal radiation of multi-angle jet fire in a natural gas pipeline provided by the present invention. Figure 1 As shown, the method includes:

[0047] Step 110 , obtaining visible light images of the jet fire at different angles at the current moment, and determining the jet angle of the jet fire in each visible light image of the jet fire;

[0048] Step 120: for any jet fire visible light image, determine a segmentation window size of the jet fire visible light image based on the injection angle of the corresponding jet fire, and segment the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images;

[0049] Step 130 , determining a segmentation threshold for each visible light sub-image, and binarizing each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image;

[0050] Step 140: Input the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model;

[0051] Step 150 , performing multi-view stereo matching based on the segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire;

[0052] In step 160 , the surface area of ​​the jet fire is determined based on the three-dimensional point cloud of the jet fire, and the thermal radiation at any position is calculated based on the surface area of ​​the jet fire.

[0053] Here, a high-speed camera can be used to capture visible light images of the jet fire at different angles (i.e., different directions of the jet fire scene) at the current moment in real time, and the jet fire's injection angle relative to the horizontal in each visible light image can be determined. It should be noted that the visible light images of the jet fire can be obtained through field testing or CFD simulation, and this is not specifically limited in this embodiment of the present invention. The injection angle of each visible light image of the jet fire can be estimated based on the direction of the natural gas pipeline leak and the position and camera parameters of the camera that captured each visible light image of the jet fire. Furthermore, an edge detection algorithm can be used to extract edges from the flame region in the visible light image of the jet fire to determine the main direction of the flame region and the injection angle of the jet fire, and this is not specifically limited in this embodiment of the present invention.

[0054] Considering that the presence of cluttered background in visible-light jet fire images interferes with flame region extraction, and that flame edges (e.g., the flame tip) are blurred, reducing the accuracy of flame morphology capture and flame surface area calculation, a more precise flame segmentation scheme is needed to accurately extract the flame region for subsequent quantification of the flame surface area. For any visible-light jet fire image, a coarse flame region extraction is performed first. A deep learning mechanism is then used to jointly learn the jet fire visible-light image and the coarsely extracted flame region. The shape, edge, and texture information of the flame provided in the coarsely extracted flame region is used to assist in extracting image semantics, thereby improving the accuracy of flame region segmentation.

[0055] Specifically, to improve the accuracy of flame region coarse extraction, considering the varying edge morphology (e.g., edge smoothness) and background noise levels of jet fires at different injection angles, the segmentation window size for any jet fire visible light image can be determined based on the jet fire's injection angle. The jet fire visible light image is then segmented based on this segmentation window size to produce multiple visible light sub-images. The larger the injection angle, the smaller the segmentation window size.

[0056] In some embodiments, the segmentation window size of any jet fire visible light image may be determined based on the following formula:

[0057]

[0058] in, is the segmentation window size of the jet fire visible light image, is the minimum window size, is the maximum window size, is the jet angle of the jet fire, It is the maximum spray angle of jet fire.

[0059] Subsequently, a segmentation threshold is determined for each visible light sub-image, and each visible light sub-image is binarized based on the corresponding segmentation threshold to obtain an initial segmented image. The segmentation threshold for any visible light sub-image is the average pixel value of each pixel in that visible light sub-image. Based on the segmentation threshold for any visible light sub-image, the binarized pixel value of pixels in that visible light sub-image with a value above the segmentation threshold can be set to 1, and the flame region is determined from that visible light sub-image based on the pixels with a binarized pixel value of 1. The pixel values ​​of pixels in the non-flame region of each visible light sub-image are set to 0 and then integrated to obtain the initial segmented image corresponding to the jet fire visible light image, thereby achieving a rough extraction of the flame region.

[0060] The initial segmented image corresponding to any jet fire visible light image and the jet fire visible light image are input together into an image segmentation model based on a deep learning mechanism to obtain a segmented image of the jet fire visible light image output by the image segmentation model.

[0061] In some embodiments, an image segmentation model can be constructed based on an encoder-decoder model. It should be noted that in order to improve the accuracy of flame segmentation, the rich temperature information in the infrared image can be introduced to assist in segmenting the flame area. Figure 2 As shown, an encoder branch can be introduced into the encoder-decoder model, forming a model structure consisting of two encoders and one decoder. For ease of explanation, the encoder responsible for semantic encoding of visible light images is called the first encoder, and the encoder responsible for semantic encoding of infrared images is called the second encoder.

[0062] Multiple downsampling layers at different scales in the first encoder are used to extract features from any jet fire visible light image and its corresponding initial segmented image, generating a visible light feature map. Multiple downsampling layers at different scales (consistent with the scales of the downsampling layers in the first encoder) in the second encoder are used to extract features from an infrared image registered with the jet fire visible light image, generating an infrared feature map. The image segmentation model also includes a feature fusion module that fuses the visible light feature map and the infrared feature map to generate a fused jet fire feature map. In some embodiments, the visible light feature map and the infrared feature map can be fused channel by channel, by superimposing feature submaps from the same channel in the visible light and infrared feature maps to achieve feature map fusion. Subsequently, multiple upsampling layers at different scales (corresponding to the scales of the downsampling layers) in the decoder decode the fused jet fire feature map to generate a segmented image of the jet fire visible light image. In some embodiments, reference can be made to the UNet network, establishing skip connections between the decoder and encoder to improve decoder performance. However, unlike the UNet network, in the image segmentation model of the present invention, any upsampling layer is skip-connected to both the downsampling layer of the same scale in the first encoder and the downsampling layer of the same scale in the second encoder, thereby utilizing the image features of the visible light image and the image features of the infrared image generated by both encoders. Specifically, the input to the upsampling layer includes the feature map output by the previous upsampling layer (or, in the case of the first upsampling layer, the jetfire fusion feature map), the feature map output by the downsampling layer of the same scale in the first encoder, and the feature map output by the downsampling layer of the same scale in the second encoder.

[0063] In some embodiments, when the first encoder performs feature extraction on the jet fire visible light image and its corresponding initial segmented image using multiple downsampling layers of different scales, feature extraction can be performed on the initial segmented image and the jet fire visible light image based on the first downsampling layer, respectively, to obtain an initial segmentation feature map and an initial visible light feature map. Subsequently, based on the feature values ​​of each feature point in the initial segmentation feature map, the feature values ​​of the corresponding feature points in the initial visible light feature map are weighted to obtain a fused visible light feature map. Feature extraction is then performed on the fused visible light feature map based on the remaining downsampling layers to obtain a visible light feature map. It should be noted that the image semantics (e.g., edge information, shape information, and texture information) contained in the coarsely extracted flame region are relatively low-level. Therefore, incorporating these into the encoding process of only shallow downsampling layers can better guide the semantic encoding of the first encoder and reduce the encoding computational complexity of the first encoder, thereby improving both semantic encoding capability and encoding efficiency.

[0064] In other embodiments, considering that the temperature distribution in infrared images has directional characteristics and that the direction of change in the temperature distribution is strongly correlated with the jet fire's injection angle, to enhance the second encoder's infrared image encoding capabilities, the second encoder's downsampling layer can be designed as a combination of multiple convolutional layers and a fusion layer. The convolutional layers employ convolution kernels with different orientations, while the fusion layer performs a weighted fusion of the feature maps output by the multiple convolutional layers based on their corresponding weights. The weights of the convolutional layers are determined based on the corresponding jet fire's injection angle. The closer the orientation of the convolution kernel in any convolutional layer is to the corresponding jet fire's injection angle, the higher the weight assigned to that convolutional layer.

[0065] After obtaining the segmented images of each jet fire visible light image output by the image segmentation model, a multi-view stereo matching model, such as CasMVSNet, can be used to perform multi-view stereo matching on the segmented images of each jet fire visible light image to obtain a depth map of the jet fire. Subsequently, the jet fire depth map is 3D reconstructed using a 3D model tool library, such as Open3D, and the jet fire depth map is converted into a 3D point cloud. The surface area of ​​the jet fire is then determined based on the 3D point cloud of the jet fire. After obtaining the surface area of ​​the jet fire, the thermal radiation at any location can be calculated based on a thermal radiation calculation formula. For example, thermal radiation calculation can be performed based on the thermal radiation formula of the Thornton model. It should be noted that based on the image segmentation and 3D reconstruction schemes provided in the above embodiments, the morphology of the jet flame can be more accurately captured, and thermal radiation results can be calculated based on a theoretical model for thermal radiation calculation. This thermal radiation result can serve as a reference for the theoretical model, helping to optimize the accuracy of the theoretical model's estimation of the jet fire surface area, thereby improving the theoretical model's calculation accuracy of the jet fire thermal radiation.

[0066] In summary, the method provided by the embodiment of the present invention obtains visible light images of jet fires at different angles at the current moment, and determines the jet angle of the jet fire in each visible light image of the jet fire; for any visible light image of the jet fire, based on the jet angle of the corresponding jet fire, determines the segmentation window size of any visible light image of the jet fire, and segments any visible light image of the jet fire based on the segmentation window size to obtain multiple visible light sub-images; determines the segmentation threshold of each visible light sub-image, and binarizes each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; compares the initial segmented image with any visible light image of the jet fire; The visible light image is input into an image segmentation model to obtain a segmented image of any jet fire visible light image output by the image segmentation model; multi-view stereo matching is performed based on the segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and three-dimensional reconstruction is performed based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; the surface area of ​​the jet fire is determined based on the three-dimensional point cloud of the jet fire, and the thermal radiation at any position is calculated based on the surface area of ​​the jet fire. This scheme can be applied to the analysis of jet fire field test results as a data reference for optimizing the theoretical model for jet fire thermal radiation calculation, thereby improving the calculation accuracy of thermal radiation.

[0067] The following describes a device for calculating the thermal radiation of a multi-angle jet fire in a natural gas pipeline provided by the present invention. The device for calculating the thermal radiation of a multi-angle jet fire in a natural gas pipeline described below and the method for calculating the thermal radiation of a multi-angle jet fire in a natural gas pipeline described above can be used for reference in correspondence with each other.

[0068] Based on any of the above embodiments, Figure 3 This is a schematic diagram of the structure of the natural gas pipeline multi-angle jet fire heat radiation calculation device provided by the present invention. Figure 3 As shown, the device includes:

[0069] The image acquisition unit 310 is used to acquire visible light images of the jet fire at different angles at the current moment, and determine the jet angle of the jet fire in each visible light image of the jet fire;

[0070] a segmentation unit 320 configured to determine, for each jet fire visible light image, a segmentation window size of the jet fire visible light image based on a jet angle of the corresponding jet fire, and segment the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images;

[0071] A binarization unit 330 is used to determine a segmentation threshold for each visible light sub-image, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image;

[0072] a segmentation unit 340 configured to input the initial segmented image and the any jet fire visible light image into an image segmentation model, and obtain a segmented image of the any jet fire visible light image output by the image segmentation model;

[0073] a 3D reconstruction unit 350 for performing multi-view stereo matching based on the segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and performing 3D reconstruction based on the depth map of the jet fire to obtain a 3D point cloud of the jet fire;

[0074] The heat radiation calculation unit 360 is configured to determine the surface area of ​​the jet fire based on the three-dimensional point cloud of the jet fire, and calculate the heat radiation at any position based on the surface area of ​​the jet fire.

[0075] The device provided by the embodiment of the present invention obtains visible light images of jet fires at different angles at the current moment and determines the jet angle of the jet fire in each visible light image of the jet fire; for any visible light image of the jet fire, based on the jet angle of the corresponding jet fire, determines the segmentation window size of any visible light image of the jet fire, and segments any visible light image of the jet fire based on the segmentation window size to obtain multiple visible light sub-images; determines the segmentation threshold of each visible light sub-image respectively, and binarizes each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; compares the initial segmented image with any visible light image of the jet fire, and performs binarization on the image based on the corresponding segmentation threshold to obtain an initial segmented image; The optical image is input into an image segmentation model to obtain a segmented image of any jet fire visible light image output by the image segmentation model; multi-view stereo matching is performed based on the segmented images of each jet fire visible light image to obtain a depth map of the jet fire, and three-dimensional reconstruction is performed based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; based on the three-dimensional point cloud of the jet fire, the surface area of ​​the jet fire is determined, and the thermal radiation at any position is calculated based on the surface area of ​​the jet fire. This scheme can be applied to the analysis of jet fire field test results as a data reference for optimizing the theoretical model for jet fire thermal radiation calculation, thereby improving the calculation accuracy of thermal radiation.

[0076] Based on any of the foregoing embodiments, determining the segmentation window size of any jet fire visible light image based on the injection angle of the corresponding jet fire includes:

[0077] The segmentation window size of any jet fire visible light image is determined based on the following formula:

[0078]

[0079] in, is the segmentation window size of any jet fire visible light image, is the minimum window size, is the maximum window size, is the jet fire's injection angle, It is the maximum spray angle of jet fire.

[0080] Based on any of the above embodiments, the image segmentation model is constructed based on an encoder-decoder model;

[0081] Inputting the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model, comprising:

[0082] performing feature extraction on the initial segmented image and any jet fire visible light image based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map;

[0083] performing feature extraction on the infrared image registered with any jet fire visible light image based on multiple downsampling layers of different scales of the second encoder in the image segmentation model to obtain an infrared feature map;

[0084] Based on the feature fusion module in the image segmentation model, the visible light feature map and the infrared feature map are fused to obtain a jet fire fusion feature map;

[0085] The jet fire fusion feature map is decoded based on multiple upsampling layers of different scales in a decoder in the image segmentation model to obtain the segmented image.

[0086] Based on any of the foregoing embodiments, performing feature extraction on the initial segmented image and any jet fire visible light image based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map includes:

[0087] Performing feature extraction on the initial segmented image and any one of the jet fire visible light images based on the first downsampling layer to obtain an initial segmentation feature map and an initial visible light feature map;

[0088] Based on the feature value of each feature point in the initial segmentation feature map, weighting the feature value of the corresponding feature point in the initial visible light feature map to obtain a fused visible light feature map;

[0089] Feature extraction is performed on the fused visible light feature map based on the remaining downsampling layers to obtain the visible light feature map.

[0090] Based on any of the above embodiments, the downsampling layer of the second encoder includes multiple convolutional layers and a fusion layer, the multiple convolutional layers use convolution kernels in different directions; the fusion layer is used to perform weighted fusion on the feature maps output by the multiple convolutional layers based on the weights corresponding to the multiple convolutional layers; wherein the weights corresponding to the multiple convolutional layers are determined based on the injection angle of the corresponding jet fire.

[0091] Based on any of the foregoing embodiments, any upsampling layer is jump-connected to a downsampling layer of the same scale in the first encoder and a downsampling layer of the same scale in the second encoder.

[0092] Based on any of the above embodiments, the segmentation threshold of any visible light sub-image is the average value of the pixel value of each pixel point in the any visible light sub-image.

[0093] Figure 4 Schematic diagram of the structure of the electronic device provided by the present invention, such as Figure 4 As shown, the electronic device may include: a processor 410, a memory 420, a communications interface 430 and a communication bus 440, wherein the processor 410, the memory 420 and the communications interface 430 communicate with each other via the communications bus 440. The processor 410 may call the logic instructions in the memory 420 to execute a method for calculating the thermal radiation of a multi-angle jet fire in a natural gas pipeline, the method comprising: obtaining visible light images of the jet fire at different angles at the current moment, and determining the jet angle of the jet fire in each visible light image of the jet fire; for any visible light image of the jet fire, based on the jet angle of the corresponding jet fire, determining the segmentation window size of the visible light image of the jet fire, and segmenting the visible light image of the jet fire based on the segmentation window size to obtain a plurality of visible light sub-images; determining the segmentation threshold of each visible light sub-image respectively, and segmenting the visible light image based on the segmentation window size. Binarizing each visible light sub-image at a corresponding segmentation threshold to obtain an initial segmented image; inputting the initial segmented image and any one of the jet fire visible light images into an image segmentation model to obtain a segmented image of any one of the jet fire visible light images output by the image segmentation model; performing multi-view stereo matching based on the segmented images of each of the jet fire visible light images to obtain a depth map of the jet fire, and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; determining the surface area of ​​the jet fire based on the three-dimensional point cloud of the jet fire, and calculating the thermal radiation at any position based on the surface area of ​​the jet fire.

[0094] Furthermore, the logic instructions in the aforementioned memory 420 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0095] On the other hand, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the method for calculating the thermal radiation of multi-angle jet fires in a natural gas pipeline provided by the above methods, the method including: obtaining visible light images of jet fires at different angles at the current moment, and determining the injection angle of the jet fire in each visible light image of the jet fire; for any jet fire visible light image, based on the injection angle of the corresponding jet fire, determining the segmentation window size of the any jet fire visible light image, and segmenting the any jet fire visible light image based on the segmentation window size. The method comprises the steps of: performing segmentation on the jet fire image to obtain a plurality of visible light sub-images; determining a segmentation threshold for each visible light sub-image, and binarizing each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; inputting the initial segmented image and any one of the jet fire visible light images into an image segmentation model to obtain a segmented image of any one of the jet fire visible light images output by the image segmentation model; performing multi-view stereo matching based on the segmented images of each of the jet fire visible light images to obtain a depth map of the jet fire, and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; determining the surface area of ​​the jet fire based on the three-dimensional point cloud of the jet fire, and calculating the thermal radiation at any position based on the surface area of ​​the jet fire.

[0096] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the above-mentioned methods for calculating thermal radiation of multi-angle jet fires in natural gas pipelines, the methods comprising: obtaining visible light images of jet fires at different angles at the current moment, and determining the jet angle of the jet fire in each visible light image of the jet fire; for any visible light image of the jet fire, based on the jet angle of the corresponding jet fire, determining a segmentation window size of the visible light image of the jet fire, and segmenting the visible light image of the jet fire based on the segmentation window size to obtain a plurality of visible light sub-images; respectively Determine a segmentation threshold for each visible light sub-image, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; input the initial segmented image and any one of the jet fire visible light images into an image segmentation model to obtain a segmented image of any one of the jet fire visible light images output by the image segmentation model; perform multi-view stereo matching based on the segmented images of each jet fire visible light image to obtain a depth map of the jet fire, and perform three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; determine the surface area of ​​the jet fire based on the three-dimensional point cloud of the jet fire, and calculate the thermal radiation at any position based on the surface area of ​​the jet fire.

[0097] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0098] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline, characterized in that: include: Obtaining visible light images of the jet fire at different angles at the current moment, and determining the jet angle of the jet fire in each visible light image of the jet fire; For any jet fire visible light image, determining a segmentation window size of the jet fire visible light image based on a jet angle of the corresponding jet fire, and segmenting the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images; Determine the segmentation threshold of each visible light sub-image respectively, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; Inputting the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model; performing multi-view stereo matching based on segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and performing three-dimensional reconstruction based on the depth map of the jet fire to obtain a three-dimensional point cloud of the jet fire; determining a surface area of ​​the jet fire based on a three-dimensional point cloud of the jet fire, and calculating thermal radiation at any position based on the surface area of ​​the jet fire; The determining, based on the injection angle of the corresponding jet fire, the segmentation window size of the visible light image of any jet fire includes: The segmentation window size of any jet fire visible light image is determined based on the following formula: in, is the segmentation window size of any jet fire visible light image, is the minimum window size, is the maximum window size, is the jet angle of the jet fire, It is the maximum spray angle of jet fire.

2. The method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline according to claim 1 is characterized in that: The image segmentation model is constructed based on an encoder-decoder model; Inputting the initial segmented image and the any jet fire visible light image into an image segmentation model to obtain a segmented image of the any jet fire visible light image output by the image segmentation model, comprising: performing feature extraction on the initial segmented image and any one of the jet fire visible light images based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map; performing feature extraction on the infrared image registered with any jet fire visible light image based on multiple downsampling layers of different scales of the second encoder in the image segmentation model to obtain an infrared feature map; Based on the feature fusion module in the image segmentation model, the visible light feature map and the infrared feature map are fused to obtain a jet fire fusion feature map; The jet fire fusion feature map is decoded based on multiple upsampling layers of different scales in a decoder in the image segmentation model to obtain the segmented image.

3. The method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline according to claim 2 is characterized in that: Performing feature extraction on the initial segmented image and any jet fire visible light image based on multiple downsampling layers of different scales of the first encoder in the image segmentation model to obtain a visible light feature map, including: Performing feature extraction on the initial segmented image and any one of the jet fire visible light images based on the first downsampling layer to obtain an initial segmentation feature map and an initial visible light feature map; Based on the feature value of each feature point in the initial segmentation feature map, weighting the feature value of the corresponding feature point in the initial visible light feature map to obtain a fused visible light feature map; Feature extraction is performed on the fused visible light feature map based on the remaining downsampling layers to obtain the visible light feature map.

4. The method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline according to claim 2 is characterized in that: The downsampling layer of the second encoder includes multiple convolutional layers and a fusion layer, wherein the multiple convolutional layers use convolution kernels in different directions; the fusion layer is used to perform weighted fusion on the feature maps output by the multiple convolutional layers based on the weights corresponding to the multiple convolutional layers; wherein the weights corresponding to the multiple convolutional layers are determined based on the injection angle of the corresponding jet fire.

5. The method for calculating heat radiation of multi-angle jet fire in a natural gas pipeline according to claim 2 is characterized in that: Any upsampling layer is jump-connected to a downsampling layer of the same scale in the first encoder and a downsampling layer of the same scale in the second encoder.

6. The method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline according to claim 1 is characterized in that: The segmentation threshold of any visible light sub-image is the average value of the pixel value of each pixel point in the any visible light sub-image.

7. A natural gas pipeline multi-angle jet fire heat radiation calculation system, characterized in that: include: an image acquisition unit, configured to acquire visible light images of the jet fire at different angles at the current moment, and determine the jet angle of the jet fire in each visible light image of the jet fire; a segmentation unit, configured to determine, for any jet fire visible light image, a segmentation window size of the jet fire visible light image based on a jet angle of the corresponding jet fire, and segment the jet fire visible light image based on the segmentation window size to obtain a plurality of visible light sub-images; A binarization unit is used to determine the segmentation threshold of each visible light sub-image respectively, and binarize each visible light sub-image based on the corresponding segmentation threshold to obtain an initial segmented image; a segmentation unit, configured to input the initial segmented image and the any jet fire visible light image into an image segmentation model, and obtain a segmented image of the any jet fire visible light image output by the image segmentation model; a 3D reconstruction unit, configured to perform multi-view stereo matching based on the segmented images of the visible light images of each jet fire to obtain a depth map of the jet fire, and perform 3D reconstruction based on the depth map of the jet fire to obtain a 3D point cloud of the jet fire; a thermal radiation calculation unit, configured to determine a surface area of ​​the jet fire based on a three-dimensional point cloud of the jet fire, and calculate thermal radiation at any position based on the surface area of ​​the jet fire; The determining, based on the injection angle of the corresponding jet fire, the segmentation window size of the visible light image of any jet fire includes: The segmentation window size of any jet fire visible light image is determined based on the following formula: in, is the segmentation window size of any jet fire visible light image, is the minimum window size, is the maximum window size, is the jet angle of the jet fire, It is the maximum spray angle of jet fire.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline as claimed in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for calculating thermal radiation of multi-angle jet fire in a natural gas pipeline as claimed in any one of claims 1 to 6 is implemented.

Images