Hydraulic ring geological fracture measurement data accuracy proofreading method

Abstract

The invention discloses a hydraulic ring geological fracture measurement data accuracy proofreading method, and the method comprises the steps: extracting manually inputted hydraulic ring geological fracture measurement data from a database, and obtaining a hydraulic ring geological fracture image collected by an unmanned plane; and crack target detection and feature extraction are performed on the hydraulic ring geological crack image by adopting an image recognition and analysis technology based on computer vision, so that an image semantic segmentation result is automatically obtained according to representation obtained by performing spatial soft constraint enhancement on hydraulic ring geological crack shallow layer features. And generating hydraulic ring geological fracture measurement intelligent generation data. And then the intelligently generated data is compared with the manually input measurement data to intelligently judge whether the data measurement accuracy is abnormal or not. Therefore, the precision of crack measurement is improved, and the accuracy of hydraulic ring geological crack measurement data is automatically corrected.

Description

Technical Field

[0001] The present application relates to the field of intelligent perception, and more specifically, to a method for calibrating the accuracy of hydraulic engineering and geological fracture measurement data. Background Art

[0002] The presence of geological cracks poses a potential threat to the safety of water conservancy projects, bridges, roads and other infrastructure. In the field of hydraulic engineering and environmental protection (covering water conservancy projects, environmental engineering and geological engineering), accurate measurement of key attributes such as the location, length and width of geological cracks is essential for assessing the safety of these structures. Accurate crack measurement data can better judge the overall health status of the structure, so as to take timely and effective maintenance measures.

[0003] Chinese patent CN115496721A proposes a wall crack measurement method and a portable wall crack measurement device, which constructs a crack image training set, uses the YOLOv5 model to train crack recognition, and optimizes the anchor box size to achieve automatic annotation of crack images and combine with an improved U-net model for crack segmentation and length measurement.

[0004] In the above patent, crack identification and measurement are mainly completed by using the YOLOv5 model to detect cracks, and then using the improved U-net model to perform image segmentation of cracks. However, since YOLOv5 mainly focuses on target detection and the U-net model refines the crack boundaries, it does not fully consider the semantic associations between pixels inside the cracks and their contextual interactions with the surrounding environment, which may affect the fineness and integrity of the crack boundaries in some cases. In other words, if the shape of the crack is complex or the crack area is slightly different from other structures, this method may result in inaccurate definition of the crack boundary or errors in the segmentation of the crack edge area, thereby affecting the accuracy of the measurement.

[0005] Therefore, an optimized accuracy calibration scheme for hydraulic ring geological fracture measurement data is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a method for calibrating the accuracy of hydraulic geotechnical fracture measurement data, which extracts the hydraulic geotechnical fracture measurement data input manually from the database, obtains the hydraulic geotechnical fracture image collected by the drone, and uses the image recognition and analysis technology based on computer vision to perform fracture target detection and feature extraction on the hydraulic geotechnical fracture image, so as to automatically obtain the image semantic segmentation result according to the representation obtained by spatial soft constraint enhancement of the shallow characteristics of the hydraulic geotechnical fracture, and generate the intelligent generation data of the hydraulic geotechnical fracture measurement. Then the intelligently generated data is compared with the manually input measurement data to intelligently determine whether there is an abnormality in the data measurement accuracy. In this way, the fracture features can be detected and extracted more accurately, and the semantic relationship between the fracture and other structures can be better understood, thereby improving the accuracy of the fracture measurement. And by comparing and analyzing with the manually measured data, the consistency and reliability of the fracture measurement results can be effectively verified, and the automatic calibration of the accuracy of the hydraulic geotechnical fracture measurement data can be finally achieved.

[0007] According to one aspect of the present application, a method for calibrating the accuracy of hydraulic ring geological fracture measurement data is provided, which comprises:

[0008] Extracting the hydraulic ring geological fracture measurement data input manually from the database;

[0009] Obtain images of hydraulic ring geological fractures collected by drones;

[0010] Performing crack target detection on the hydraulic ring geological crack image to obtain a hydraulic ring geological crack anchor frame image;

[0011] Extracting crack features from the hydraulic annular geological crack anchor frame image to obtain a shallow characteristic map of the hydraulic annular geological crack;

[0012] The shallow feature map of the hydraulic annular geological fractures is enhanced to obtain the shallow enhanced features of the hydraulic annular geological fractures, including: feature dispersion of the shallow feature map of the hydraulic annular geological fractures along the channel dimension to obtain a set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures; semantic association and attenuation modulation based on spatial soft constraints on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures; based on the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures, convolution coding enhancement is performed on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain the shallow enhanced features of the hydraulic annular geological fractures;

[0013] Based on the shallow enhanced features of the hydraulic ring geological fractures, an image semantic segmentation result is obtained, and based on the image semantic segmentation result, intelligent generation data of hydraulic ring geological fracture measurement is obtained;

[0014] Based on the comparison between the intelligently generated data of the hydraulic engineering ring geological fracture measurement and the hydraulic engineering ring geological fracture measurement data, it is determined whether there is an abnormality in the data measurement accuracy.

[0015] Compared with the prior art, the present application provides a method for calibrating the accuracy of hydraulic geotechnical fracture measurement data, which extracts the hydraulic geotechnical fracture measurement data input manually from the database, obtains the hydraulic geotechnical fracture image collected by the drone, and uses the image recognition and analysis technology based on computer vision to detect the fracture target and extract the feature of the hydraulic geotechnical fracture image, so as to automatically obtain the image semantic segmentation result according to the representation obtained by spatial soft constraint enhancement of the shallow characteristics of the hydraulic geotechnical fracture, and generate the intelligent generation data of the hydraulic geotechnical fracture measurement. Then the intelligent generation data is compared with the manually input measurement data to intelligently determine whether there is an abnormality in the data measurement accuracy. In this way, the fracture features can be detected and extracted more accurately, and the semantic relationship between the fracture and other structures can be better understood, thereby improving the accuracy of the fracture measurement. And by comparing and analyzing with the manually measured data, the consistency and reliability of the fracture measurement results can be effectively verified, and the automatic calibration of the accuracy of the hydraulic geotechnical fracture measurement data can be finally achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0017] Figure 1 The present invention is a flowchart of a method for calibrating the accuracy of hydraulic ring geological fracture measurement data according to an embodiment of the present application.

[0018] Figure 2 The data flow diagram of the method for calibrating the accuracy of hydraulic ring geological fracture measurement data according to an embodiment of the present application is shown.

[0019] Figure 3 The present invention is a flow chart of enhancing the shallow characteristics of the hydraulic annular geological fractures in the hydraulic annular geological fracture measurement data accuracy calibration method according to an embodiment of the present application to obtain the shallow enhanced characteristics of the hydraulic annular geological fractures.

[0020] Figure 4The present invention is a flowchart of performing semantic association and attenuation modulation based on spatial soft constraints on a set of shallow pixel granularity feature vectors of hydraulic annular geological fractures in the accuracy correction method of hydraulic annular geological fracture measurement data according to an embodiment of the present application to obtain a shallow semantic association spatial soft constraint score topological matrix of hydraulic annular geological fractures.

[0021] Figure 5 A flowchart of calculating the semantic association score between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures in the accuracy correction method of hydraulic annular geological fracture measurement data according to an embodiment of the present application to obtain a shallow semantic association score topological matrix of hydraulic annular geological fractures. DETAILED DESCRIPTION

[0022] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, but rather these embodiments are provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0023] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0024] In the description of the embodiments of the present disclosure, the term "including" and similar terms should be understood as open inclusion, that is, "including but not limited to". The term "based on" should be understood as "based at least in part on". The term "one embodiment" or "the embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may also be included below.

[0025] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0026] The presence of geological cracks poses a potential threat to the safety of water conservancy projects, bridges, roads and other infrastructure. In the field of hydraulic engineering and environmental protection (covering water conservancy projects, environmental engineering and geological engineering), accurate measurement of key attributes such as the location, length and width of geological cracks is essential for assessing the safety of these structures. Accurate crack measurement data can better judge the overall health status of the structure, so as to take timely and effective maintenance measures.

[0027] Chinese patent CN115496721A proposes a wall crack measurement method and a portable wall crack measurement device, which constructs a crack image training set, uses the YOLOv5 model to train crack recognition, and optimizes the anchor box size to achieve automatic annotation of crack images and combine with an improved U-net model for crack segmentation and length measurement.

[0028] In the above patent, crack identification and measurement are mainly completed by using the YOLOv5 model to detect cracks, and then using the improved U-net model to perform image segmentation of cracks. However, since YOLOv5 mainly focuses on target detection and the U-net model refines the crack boundaries, it does not fully consider the semantic associations between pixels inside the cracks and their contextual interactions with the surrounding environment, which may affect the fineness and integrity of the crack boundaries in some cases. In other words, if the shape of the crack is complex or the crack area is slightly different from other structures, this method may result in inaccurate definition of the crack boundary or errors in the segmentation of the crack edge area, thereby affecting the accuracy of the measurement.

[0029] Therefore, in view of the above technical problems, the technical concept of the present application is to extract the hydraulic annular geological fracture measurement data input manually from the database, obtain the hydraulic annular geological fracture image collected by the drone, and use the image recognition and analysis technology based on computer vision to detect the fracture target and extract the feature of the hydraulic annular geological fracture image, so as to automatically obtain the image semantic segmentation result according to the representation obtained by spatial soft constraint enhancement of the shallow characteristics of the hydraulic annular geological fracture, and generate the intelligent generation data of the hydraulic annular geological fracture measurement. Then the intelligently generated data is compared with the manually input measurement data to intelligently determine whether there is an abnormality in the data measurement accuracy. In this way, the fracture features can be detected and extracted more accurately, and the semantic relationship between the fracture and other structures can be better understood, thereby improving the accuracy of the fracture measurement. And by comparing and analyzing with the manually measured data, the consistency and reliability of the fracture measurement results can be effectively verified, and the automatic calibration of the accuracy of the hydraulic annular geological fracture measurement data can be finally achieved.

[0030] Figure 1 The present invention is a flowchart of a method for calibrating the accuracy of hydraulic ring geological fracture measurement data according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the method for calibrating the accuracy of hydraulic ring geological fracture measurement data according to an embodiment of the present application. Figure 1 and Figure 2As shown, according to the accuracy calibration method of hydraulic annular geological fracture measurement data of the embodiment of the present application, the method includes: S110, extracting hydraulic annular geological fracture measurement data input manually from a database; S120, acquiring hydraulic annular geological fracture images collected by drones; S130, performing fracture target detection on the hydraulic annular geological fracture images to obtain hydraulic annular geological fracture anchor frame images; S140, performing fracture feature extraction on the hydraulic annular geological fracture anchor frame images to obtain hydraulic annular geological fracture shallow layer feature maps; S150, performing fracture shallow layer feature enhancement on the hydraulic annular geological fracture shallow layer feature maps to obtain hydraulic annular geological fracture shallow layer enhancement features; S160, obtaining image semantic segmentation results based on the hydraulic annular geological fracture shallow layer enhancement features, and obtaining hydraulic annular geological fracture measurement intelligent generation data based on the image semantic segmentation results; S170, determining whether there is data measurement accuracy abnormality based on a comparison between the hydraulic annular geological fracture measurement intelligent generation data and the hydraulic annular geological fracture measurement data.

[0031] In step S110, the hydraulic annular geological fracture measurement data manually input is extracted from the database. In step S120, the hydraulic annular geological fracture image collected by the drone is obtained. It should be understood that the hydraulic annular geological fracture measurement data manually input includes the specific location of the fracture, the length of the fracture, etc., and is obtained by manual detection. The hydraulic annular geological fracture image collected by the drone can observe the details of the fracture in detail. Based on this, in order to verify the accuracy of the fracture measurement data more carefully, in the technical solution of the present application, the hydraulic annular geological fracture measurement data manually input is extracted from the database, and the hydraulic annular geological fracture image collected by the drone is obtained, so as to achieve high precision, high efficiency and high reliability of fracture measurement by combining the manually input fracture measurement data and the fracture image collected by the drone.

[0032] In step S130, the hydraulic annular geological fracture image is subjected to fracture target detection to obtain a hydraulic annular geological fracture anchor frame image. Specifically, in an embodiment of the present application, the hydraulic annular geological fracture image is subjected to fracture target detection to obtain a hydraulic annular geological fracture anchor frame image, including: inputting the hydraulic annular geological fracture image into a fracture target detection network based on a YOLOv5 recognition model to obtain the hydraulic annular geological fracture anchor frame image. Accordingly, considering that the hydraulic annular geological fracture image contains key parts of the fracture itself, such as the width, length, shape, etc. of the fracture, but there are also some background environments such as trees and weeds in the background. The YOLOv5 model shows a high accuracy in target detection tasks, especially in processing small target detection in complex backgrounds. Cracks are usually small and easily masked by complex geological backgrounds. The YOLOv5 model can effectively detect these small cracks and improve the accuracy of crack detection. Based on this, in order to accurately determine the location of the cracks and eliminate the influence of irrelevant parts in the image to focus on the information of the cracks themselves, in the technical solution of the present application, the hydraulic ring geological crack image is input into the crack target detection network based on the YOLOv5 recognition model to obtain the hydraulic ring geological crack anchor frame image.

[0033] In step S140, the hydraulic annular geological fracture anchor frame image is subjected to fracture feature extraction to obtain a shallow characteristic map of the hydraulic annular geological fracture. Specifically, in an embodiment of the present application, the hydraulic annular geological fracture anchor frame image is subjected to fracture feature extraction to obtain a shallow characteristic map of the hydraulic annular geological fracture, including: inputting the hydraulic annular geological fracture anchor frame image into a fracture feature extractor based on a void convolution layer to obtain the shallow characteristic map of the hydraulic annular geological fracture. It should be understood that in order to further capture and mine the key fracture feature information in the hydraulic annular geological fracture anchor frame image, such as the fine features and texture information of the fracture, in the technical solution of the present application, the hydraulic annular geological fracture anchor frame image is input into a fracture feature extractor based on a void convolution layer to obtain a shallow characteristic map of the hydraulic annular geological fracture. That is, void convolution helps to retain more detailed information in the hydraulic annular geological fracture anchor frame image, which is particularly important for the extraction of fracture features, because fractures often appear as subtle changes in the image.

[0034] In step S150, the shallow characteristic map of the hydraulic annular geological fracture is enhanced to obtain the shallow enhanced characteristics of the hydraulic annular geological fracture. Further, considering that the shallow characteristic map of the hydraulic annular geological fracture contains shallow local details and textures about the fracture, such as tiny branches of the fracture, the roughness of the edge of the fracture, etc. Although the hole convolution can effectively expand the receptive field and maintain the spatial resolution in many cases, its jump sampling method may lead to the loss of some detail information, especially when dealing with features such as geological fractures that are rich in details and complex in morphology. Therefore, in order to further highlight the key features of the fracture on the basis of maintaining the integrity of the fracture characteristics and suppress unnecessary background noise and interference, in the technical solution of the present application, the shallow characteristic map of the hydraulic annular geological fracture is enhanced to obtain the shallow enhanced characteristic map of the hydraulic annular geological fracture as the shallow enhanced characteristics of the hydraulic annular geological fracture.

[0035] In detail, first, the shallow feature map of the hydraulic annular geological fracture is feature-dispersed along the channel dimension to convert its shape into pixel-granular features to obtain a set of shallow pixel-granular feature vectors of the hydraulic annular geological fracture. Then, the semantic association score between any two shallow pixel-granular feature vectors of the hydraulic annular geological fracture is calculated to obtain a shallow semantic association score topological matrix of the hydraulic annular geological fracture. In this way, the semantic relationship between each pixel in the fracture feature map can be better understood, which helps to identify which pixels are more likely to belong to the same fracture or part of a fracture. Specifically, first, the set of shallow pixel-granular feature vectors of the hydraulic annular geological fracture is mapped to the Poincare space. It can be understood that the Poincare space is a hyperbolic geometric model suitable for processing data with an intrinsic hierarchical structure, that is, the fracture features are usually hierarchical, for example, a fracture may be composed of multiple sub-fractures, and the sub-fractures may be composed of smaller fracture units. The Poincare space can capture this hierarchical relationship well, thereby better representing the hierarchical connection between fracture features. After that, the Poincare distance between any two mapped pixel-granular feature vectors is calculated as the semantic association score between any two mapped pixel-granular feature vectors. That is to say, in the hyperbolic geometric space, the Poincare distance can better capture the nonlinear relationship between features. Compared with the distance metric in the Euclidean space, the Poincare distance can not only more accurately reflect the actual similarity between features, but also imply the semantic association. In particular, in the hyperbolic space, as the distance between two points increases, the actual distance grows exponentially. This means that in the area close to the origin, many similar node features can be placed closely, while in the area far from the origin, different node features can be distributed relatively sparsely. In this way, even small feature differences can be mapped to farther positions, which helps to enhance the model's ability to distinguish features of different categories. Next, considering that the correlation between adjacent regions is usually higher than that between distant regions, this means that when processing the score matrix, a higher weight should be given to the interaction between adjacent regions, and as the distance between regions increases, the weight of the interaction between them should gradually decrease. Therefore, the semantic association score topology matrix is ​​further modulated by spatial spatial attention attenuation based on the spatial distance between any two vectors in the set of pixel-sized feature vectors to obtain the shallow semantic association space soft constraint score topology matrix of hydraulic ring geological fractures. Subsequently, the shallow semantic association space soft constraint score topology matrix of hydraulic ring geological fractures is subjected to dilated convolution coding to capture and extract the complex semantic association relationship between each pixel-sized feature vector to obtain the shallow semantic association space soft constraint score topology feature matrix of hydraulic ring geological fractures.Next, in order to better capture the global contextual semantic association relationship between fracture features, the set of spatial soft constraint score topological feature matrix and pixel granularity feature vectors are input into the graph convolutional coding module to enhance the feature representation ability and context understanding ability, and obtain the set of shallow context pixel granularity feature vectors of hydraulic ring geological fractures.

[0036] Figure 3 The present invention is a flow chart of enhancing the shallow characteristics of the shallow characteristic map of the hydraulic annular geological fractures to obtain the shallow enhanced characteristics of the hydraulic annular geological fractures in the accuracy calibration method of the hydraulic annular geological fracture measurement data according to the embodiment of the present application. Specifically, in the embodiment of the present application, Figure 3 As shown, the shallow feature map of the hydraulic annular geological fractures is enhanced to obtain the shallow enhanced features of the hydraulic annular geological fractures, including: S210, feature dispersion of the shallow feature map of the hydraulic annular geological fractures along the channel dimension to obtain a set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures; S220, semantic association and attenuation modulation based on spatial soft constraints on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures; S230, based on the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures, convolution coding enhancement is performed on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain the shallow enhanced features of the hydraulic annular geological fractures.

[0037] Specifically, in step S220, the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is subjected to semantic association and attenuation modulation based on spatial soft constraints to obtain a shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures. Figure 4 The present invention is a flowchart of performing semantic association and attenuation modulation based on spatial soft constraints on the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures in the accuracy calibration method of hydraulic annular geological fracture measurement data according to the embodiment of the present application to obtain a shallow semantic association spatial soft constraint score topological matrix of hydraulic annular geological fractures. More specifically, in the embodiment of the present application, Figure 4As shown, the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is subjected to semantic association and attenuation modulation based on spatial soft constraints to obtain a shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures, including: S310, calculating the semantic association score between any two shallow pixel granularity feature vectors of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a shallow semantic association score topological matrix of the hydraulic annular geological fractures; S320, based on the spatial distance between any two shallow pixel granularity feature vectors of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures, performing spatial attention attenuation modulation on the shallow semantic association score topological matrix of the hydraulic annular geological fractures to obtain a shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures.

[0038] Figure 5 The present invention is a flowchart of calculating the semantic association score between any two shallow pixel granularity feature vectors of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures in the accuracy calibration method of the hydraulic annular geological fractures measurement data according to the embodiment of the present application to obtain the shallow semantic association score topological matrix of the hydraulic annular geological fractures. More specifically, in the embodiment of the present application, Figure 5 As shown, the semantic association score between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures is calculated to obtain a shallow semantic association score topological matrix of hydraulic annular geological fractures, including: S410, mapping each shallow pixel granularity feature vector of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures to the Poincare space to obtain a set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures; S420, calculating the semantic association score between any two mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures to obtain the shallow semantic association score topological matrix of hydraulic annular geological fractures composed of multiple shallow semantic association score values ​​of hydraulic annular geological fractures.

[0039] More specifically, in an embodiment of the present application, each shallow pixel granularity feature vector of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is mapped to the Poincare space to obtain a set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures after mapping, including: after matrix-vector multiplication of a first weight matrix and the shallow pixel granularity feature vector of the hydraulic annular geological fractures, vector-matrix multiplication of the obtained shallow pixel granularity weight vector of the hydraulic annular geological fractures and a second weight matrix to obtain the shallow pixel granularity feature vector of the hydraulic annular geological fractures after mapping.

[0040] More specifically, in the embodiment of the present application, the semantic association score between any two mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures is calculated to obtain the shallow semantic association score topological matrix of hydraulic annular geological fractures composed of multiple shallow semantic association score values ​​of hydraulic annular geological fractures, including: extracting the i-th mapped shallow pixel granularity feature vector of hydraulic annular geological fractures and the j-th mapped shallow pixel granularity feature vector from the set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures. The shallow pixel granularity feature vector of the hydraulic annular geological fracture is obtained by calculating the position difference between the i-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture and the j-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture to obtain the shallow pixel granularity difference vector of the hydraulic annular geological fracture after mapping, and calculating the square of the one-norm of the shallow pixel granularity difference vector of the hydraulic annular geological fracture after mapping to obtain the shallow characteristic difference norm value of the hydraulic annular geological fracture after mapping; calculating the constant one and the shallow pixel granularity feature vector of the i-th mapped hydraulic annular geological fracture The difference between the square of the one-norm feature vector of the eigenvector is used to obtain the shallow characteristic difference of the i-th mapped hydraulic annular geological fracture; the difference between the constant one and the square of the one-norm feature vector of the j-th mapped hydraulic annular geological fracture shallow pixel granularity feature vector is calculated to obtain the j-th mapped hydraulic annular geological fracture shallow characteristic difference; the product between the i-th mapped hydraulic annular geological fracture shallow characteristic difference and the j-th mapped hydraulic annular geological fracture shallow characteristic difference is calculated to obtain the shallow characteristic product value of the mapped hydraulic annular geological fracture; the mapped hydraulic annular geological fracture shallow characteristic difference is calculated. After dividing the differential norm value of the shallow characteristics of the hydraulic annular geological fractures by the product value of the shallow characteristics of the mapped hydraulic annular geological fractures, the obtained value is first multiplied by a constant two and then added to a constant one to obtain a semantic metric value of the shallow characteristics of the hydraulic annular geological fractures after mapping; the semantic metric value of the shallow characteristics of the hydraulic annular geological fractures after mapping is input into the arccosh function to obtain the shallow semantic association score value of the hydraulic annular geological fractures between the i-th shallow pixel granularity feature vector of the hydraulic annular geological fractures after mapping and the j-th shallow pixel granularity feature vector of the hydraulic annular geological fractures after mapping.

[0041] Specifically, in step S320, based on the spatial distance between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures, the shallow semantic association score topological matrix of the hydraulic annular geological fractures is subjected to spatial attention attenuation modulation to obtain the shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures, including: calculating the spatial distance between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures to obtain a set of spatial distance values ​​of hydraulic annular geological fractures; calculating the spatial distance between each water conservancy unit in the set of spatial distance values ​​of hydraulic annular geological fractures with the natural constant e as the base; The spatial distance value of the hydraulic annular geological fracture is an exponential function of the exponent to obtain a set of spatial distance index values ​​of the hydraulic annular geological fracture; the set of spatial distance index values ​​of the hydraulic annular geological fracture is multiplied by the attenuation modulation value by position to obtain a set of spatial distance weighted values ​​of the hydraulic annular geological fracture; each shallow semantic association score value of the hydraulic annular geological fracture in the shallow semantic association score topological matrix of the hydraulic annular geological fracture is divided by the corresponding spatial distance weighted value of the hydraulic annular geological fracture to obtain the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fracture composed of multiple shallow semantic association space soft constraint score values ​​of the hydraulic annular geological fracture.

[0042] Specifically, in S230, based on the shallow semantic association space soft constraint score topology matrix of the hydraulic annular geological fractures, the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is convolutionally coded and enhanced to obtain the shallow enhancement features of the hydraulic annular geological fractures, including: performing void convolution coding on the shallow semantic association space soft constraint score topology matrix of the hydraulic annular geological fractures to obtain the shallow semantic association space soft constraint score topology feature matrix of the hydraulic annular geological fractures; inputting the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures and the shallow semantic association space soft constraint score topology feature matrix of the hydraulic annular geological fractures into a graph convolution coding module to obtain a set of shallow context pixel granularity feature vectors of the hydraulic annular geological fractures; performing feature shape reshaping on the set of shallow context pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a shallow enhancement feature map of the hydraulic annular geological fractures as the shallow enhancement features of the hydraulic annular geological fractures.

[0043] In the embodiment of the present application, specifically, the shallow characteristic map of the hydraulic ring geological fracture is enhanced with shallow characteristic of the fracture, which can be expressed by the formula:

[0044]

[0045] h i =W1v i W2

[0046] H = {h1, h2, ..., h i, ..., h j , ..., h n}

[0047]

[0048] T = AtrousConv(S′)

[0049] V′=GCN(V,T)

[0050] F′=reshape[F, (H, W, C)]

[0051] Where F is the shallow characteristic map of the hydraulic ring geological fractures, and F∈R H×W×C , R is a real number set, W and H represent the width and height of each characteristic matrix of the shallow characteristic map of the hydraulic ring geological fractures along the channel dimension, C represents the number of channels of the shallow characteristic map of the hydraulic ring geological fractures, reshape[·|] is to reshape the characteristic map, v i is the i-th pixel granularity feature vector of the shallow layer of the hydraulic ring geological fracture in the set of pixel granularity feature vectors of the shallow layer of the hydraulic ring geological fracture, and v i ∈R C , V is the set of shallow pixel granularity feature vectors of the hydraulic ring geological fractures, W1 and W2 are the first weight matrix and the second weight matrix respectively, H is the set of shallow pixel granularity feature vectors of the hydraulic ring geological fractures after mapping, h1, h2, ..., h i , ..., h j , ..., h n are respectively the first, second, ..., ith, ..., jth, ..., and nth pixel granularity feature vectors of shallow layer of hydraulic engineering ring geological fractures after mapping in the set of the shallow layer pixel granularity feature vectors of hydraulic engineering ring geological fractures after mapping, ||·|| 2 is the square of the one-norm of the vector, arccosh(·) is the inverse hyperbolic cosine function, d P (h i ,h j ) is used to calculate h i and h j The shallow semantic association score between the hydraulic ring geological fractures, S i j is the eigenvalue of each position in the shallow semantic association score topological matrix of hydraulic ring geological fractures, d(v i , v j ) is v i and v j The spatial distance between the hydraulic ring geological fractures, α is the attenuation modulation value, S ij′ is the eigenvalue of each position in the soft constraint score topology matrix of the shallow semantic association space of hydraulic and geological fractures, S′ is the soft constraint score topology matrix of the shallow semantic association space of hydraulic and geological fractures, AtrousConv(·) is the void convolution coding operation, T is the soft constraint score topology feature matrix of the shallow semantic association space of hydraulic and geological fractures, GCN(·t·) is the graph convolution coding operation, V′ is the set of pixel granularity feature vectors of the shallow context of hydraulic and geological fractures, and F′ is the shallow enhanced feature map of the hydraulic and geological fractures.

[0052] In step S160, based on the shallow enhancement features of the hydraulic annular geological fractures, an image semantic segmentation result is obtained, and based on the image semantic segmentation result, the intelligent generation data of the hydraulic annular geological fracture measurement is obtained. Specifically, in an embodiment of the present application, based on the shallow enhancement features of the hydraulic annular geological fractures, an image semantic segmentation result is obtained, and based on the image semantic segmentation result, the intelligent generation data of the hydraulic annular geological fracture measurement is obtained, including: performing image semantic segmentation on the shallow enhancement feature map of the hydraulic annular geological fractures to obtain the image semantic segmentation result, and obtaining the intelligent generation data of the hydraulic annular geological fracture measurement based on the image semantic segmentation result. Correspondingly, the image semantic segmentation is performed on the shallow enhancement feature map of the hydraulic annular geological fractures to obtain the image semantic segmentation result, and the intelligent generation data of the hydraulic annular geological fracture measurement is obtained based on the image semantic segmentation result. That is, semantic segmentation is performed on the shallow enhancement features of the hydraulic annular geological fractures obtained by feature enhancement using the shallow enhancement features of the hydraulic annular geological fractures, so as to automatically obtain the image semantic segmentation result and generate the intelligent generation data of the hydraulic annular geological fracture measurement. In particular, in a specific implementation of the present application, image semantic segmentation is performed on the shallow enhanced feature map of the hydraulic annular geological fractures to obtain the image semantic segmentation result, and based on the image semantic segmentation result, the hydraulic annular geological fracture measurement intelligent generation data is obtained, which can be processed by using an image segmentation model based on the improved U-net method. The specific processing method is explained in detail in patent CN115496721A. Since this processing method is not the focus of this solution, it will not be described in detail.

[0053] In step S170, based on the comparison between the intelligently generated data for the hydraulic ring geological fracture measurement and the measured data for the hydraulic ring geological fracture, it is determined whether there is an abnormality in the data measurement accuracy. It should be understood that in order to more accurately perform measurement accuracy calibration, so as to effectively verify the consistency of the measurement results, in the technical solution of the present application, based on the comparison between the intelligently generated data for the hydraulic ring geological fracture measurement and the measured data for the hydraulic ring geological fracture, it is determined whether there is an abnormality in the data measurement accuracy. In this way, the crack features can be detected and extracted more accurately, and the semantic relationship between the crack and other structures can be better understood, thereby improving the accuracy of crack measurement. And by comparing and analyzing with the manual measurement data, the consistency and reliability of the crack measurement results can be effectively verified, and the efficiency and intelligence of the crack measurement work can be finally realized.

[0054] It should be understood that, considering that the shallow characteristic map of the hydraulic annular geological fractures represents the shallow characteristics of the image of the anchor frame image of the hydraulic annular geological fractures, when the shallow characteristics of the fractures are enhanced, the characteristic information of each position in the shallow characteristic map of the hydraulic annular geological fractures will be enhanced to varying degrees of spatial significance, which makes the shallow enhanced characteristic map of the hydraulic annular geological fractures have a characteristic spatial distribution structure difference. Therefore, it is expected to improve the convergence and generalization effect of the complex spatial structure discriminant regression of the shallow enhanced characteristic map of the hydraulic annular geological fractures.

[0055] Preferably, in an example of the present application, performing image semantic segmentation on the shallow enhanced feature map of the hydraulic annular geological fractures to obtain an image semantic segmentation result includes: calculating the sum of the absolute values ​​of each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fractures to obtain a first shallow enhanced spatial structure value of the hydraulic annular geological fractures, and calculating the square root of the sum of the squares of each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fractures to obtain a second shallow enhanced spatial structure value of the hydraulic annular geological fractures; multiplying each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fractures with the first shallow enhanced spatial structure value of the hydraulic annular geological fractures and the second shallow enhanced spatial structure value of the hydraulic annular geological fractures to obtain a first shallow enhanced structural reference value and a second shallow enhanced structural reference value of the hydraulic annular geological fractures corresponding to each eigenvalue; multiplying each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fractures with the scale of the shallow enhanced feature map of the hydraulic annular geological fractures and the square root of the scale to obtain a first shallow enhanced structural reference value and a second shallow enhanced structural reference value of the hydraulic annular geological fractures corresponding to each eigenvalue. The first hydraulic annular geological fracture shallow enhancement scale transformation value and the second hydraulic annular geological fracture shallow enhancement scale transformation value corresponding to the eigenvalues; the first hydraulic annular geological fracture shallow enhancement structure reference value is divided by the difference between the first hydraulic annular geological fracture shallow enhancement spatial structure value and the first hydraulic annular geological fracture shallow enhancement scale transformation value to obtain the first hydraulic annular geological fracture shallow enhancement transformation adjustment value; the second hydraulic annular geological fracture shallow enhancement structure reference value is divided by the difference between the second hydraulic annular geological fracture shallow enhancement spatial structure value and the second hydraulic annular geological fracture shallow enhancement scale transformation value to obtain the second hydraulic annular geological fracture shallow enhancement transformation adjustment value; the weighted sum of the first hydraulic annular geological fracture shallow enhancement transformation adjustment value and the second hydraulic annular geological fracture shallow enhancement transformation adjustment value is calculated to obtain each eigenvalue of the optimized hydraulic annular geological fracture shallow enhancement feature map; the optimized hydraulic annular geological fracture shallow enhancement feature map is subjected to image semantic segmentation to obtain the image semantic segmentation result.

[0056]

[0057] f 1i =(α×f i ) / (α-n×f i )

[0058]

[0059] n=W×H×C

[0060] Wherein, F is the shallow layer enhancement characteristic map of the hydraulic annular geological fractures, W and H represent the width and height of each characteristic matrix of the shallow layer enhancement characteristic map of the hydraulic annular geological fractures along the channel dimension, C represents the number of channels of the shallow layer enhancement characteristic map of the hydraulic annular geological fractures, n represents the scale of the shallow layer enhancement characteristic map of the hydraulic annular geological fractures, and f i represents the characteristic value of each position in the shallow layer enhancement characteristic map of the hydraulic annular geological fracture, α represents the shallow layer enhancement spatial structure value of the first hydraulic annular geological fracture, β represents the shallow layer enhancement spatial structure value of the second hydraulic annular geological fracture, f 1i represents the shallow enhancement transformation adjustment value of the first hydraulic ring geological fracture at each position in the shallow enhancement transformation adjustment map of the first hydraulic ring geological fracture, f 2i represents the shallow enhancement transformation adjustment value of the second hydraulic annular geological fracture at each position in the shallow enhancement transformation adjustment map of the second hydraulic annular geological fracture, F1 represents the shallow enhancement transformation adjustment map of the first hydraulic annular geological fracture, F2 represents the shallow enhancement transformation adjustment map of the second hydraulic annular geological fracture, represents point-by-point addition, ⊙ represents point-by-point multiplication, ω represents a weighted hyperparameter, and F′ represents the optimized shallow-layer enhanced characteristic map of hydraulic ring geological fractures.

[0061] Therefore, in the preferred example, for the spatial structure information of the feature set of the shallow enhanced feature map of the hydraulic annular geological fracture in the high-dimensional space, the scale-based frame transformation of each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fracture is performed by using the class norm space structured representation of the shallow enhanced feature map of the hydraulic annular geological fracture as the reference window, and the frame attention weight adjustment based on the spatial structure of each eigenvalue of the shallow enhanced feature map of the hydraulic annular geological fracture is realized to ensure the spatial transformation (translation, scaling and rotation) invariance of the shallow enhanced feature map of the hydraulic annular geological fracture under the interaction of the feature space, thereby improving the convergence and generalization effect of the discriminant regression of the feature set of the shallow enhanced feature map of the hydraulic annular geological fracture under the complex spatial structure representation, and improving the fine-grained segmentation accuracy of the image semantic segmentation. In this way, the crack features can be detected and extracted more accurately, and the semantic relationship between the crack and other structures can be better understood, thereby improving the accuracy of the crack measurement. And by comparing and analyzing with the manual measurement data, the consistency and reliability of the crack measurement results can be effectively verified, and the efficiency and intelligence of the crack measurement work can be finally realized.

[0062] In summary, the accuracy calibration method of the hydraulic ring geological fracture measurement data based on the embodiment of the present application is explained, which extracts the hydraulic ring geological fracture measurement data input manually from the database, obtains the hydraulic ring geological fracture image collected by the drone, and uses the image recognition and analysis technology based on computer vision to perform fracture target detection and feature extraction on the hydraulic ring geological fracture image, so as to automatically obtain the image semantic segmentation result according to the spatial soft constraint enhancement of the shallow characteristics of the hydraulic ring geological fracture, and generate the intelligent generation data of the hydraulic ring geological fracture measurement. Then the intelligently generated data is compared with the manually input measurement data to intelligently determine whether there is an abnormality in the data measurement accuracy. In this way, the fracture features can be detected and extracted more accurately, and the semantic relationship between the fracture and other structures can be better understood, thereby improving the accuracy of the fracture measurement. And by comparing and analyzing with the manually measured data, the consistency and reliability of the fracture measurement results can be effectively verified, and the automatic calibration of the accuracy of the hydraulic ring geological fracture measurement data can be finally achieved.

[0063] The above is merely an example of the principles of the present disclosure, and various modifications may be made by those skilled in the art without departing from the scope of the present disclosure. The above embodiments are presented for the purpose of illustration rather than limitation. The present disclosure may also take many forms other than those explicitly described herein.

Claims

1. A method for checking the accuracy of hydraulic ring geological fracture measurement data, characterized in that: include: Extracting the hydraulic ring geological fracture measurement data input manually from the database; Obtain images of hydraulic ring geological fractures collected by drones; Performing crack target detection on the hydraulic ring geological crack image to obtain a hydraulic ring geological crack anchor frame image; Extracting crack features from the hydraulic annular geological crack anchor frame image to obtain a shallow characteristic map of the hydraulic annular geological crack; The shallow feature map of the hydraulic annular geological fractures is enhanced to obtain the shallow enhanced features of the hydraulic annular geological fractures, including: feature dispersion of the shallow feature map of the hydraulic annular geological fractures along the channel dimension to obtain a set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures; semantic association and attenuation modulation based on spatial soft constraints on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures; based on the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures, convolution coding enhancement is performed on the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain the shallow enhanced features of the hydraulic annular geological fractures; Based on the shallow enhanced features of the hydraulic ring geological fractures, an image semantic segmentation result is obtained, and based on the image semantic segmentation result, intelligent generation data of hydraulic ring geological fracture measurement is obtained; Based on the comparison between the intelligently generated data of the hydraulic engineering ring geological fracture measurement and the hydraulic engineering ring geological fracture measurement data, it is determined whether there is an abnormality in the data measurement accuracy.

2. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 1 is characterized in that: Performing crack target detection on the hydraulic annular geological crack image to obtain a hydraulic annular geological crack anchor frame image, including: inputting the hydraulic annular geological crack image into a crack target detection network based on a YOLOv5 recognition model to obtain the hydraulic annular geological crack anchor frame image.

3. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 2 is characterized in that: The method comprises: extracting crack features from the hydraulic annular geological crack anchor frame image to obtain a shallow characteristic map of the hydraulic annular geological crack, comprising: inputting the hydraulic annular geological crack anchor frame image into a crack feature extractor based on a void convolution layer to obtain the shallow characteristic map of the hydraulic annular geological crack.

4. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 3 is characterized in that: The set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is subjected to semantic association and attenuation modulation based on spatial soft constraints to obtain a shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures, including: Calculating the semantic association score between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures to obtain a shallow semantic association score topological matrix of hydraulic annular geological fractures; Based on the spatial distance between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures, the shallow semantic association score topological matrix of the hydraulic annular geological fractures is spatially attended to and attenuated modulated to obtain the shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures.

5. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 4 is characterized in that: Calculating the semantic association score between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures to obtain a shallow semantic association score topological matrix of hydraulic annular geological fractures, including: Mapping each of the shallow pixel granularity feature vectors of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures to the Poincare space to obtain a set of mapped shallow pixel granularity feature vectors of the hydraulic annular geological fractures; The semantic association score between any two mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures is calculated to obtain the shallow semantic association score topological matrix of hydraulic annular geological fractures composed of multiple shallow semantic association score values ​​of hydraulic annular geological fractures.

6. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 5 is characterized in that: Each shallow pixel granularity feature vector of the hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures is mapped to the Poincare space to obtain a set of shallow pixel granularity feature vectors of the hydraulic annular geological fractures after mapping, including: after matrix-vector multiplication of a first weight matrix and the shallow pixel granularity feature vector of the hydraulic annular geological fractures, vector-matrix multiplication of the obtained shallow pixel granularity weight vector of the hydraulic annular geological fractures and a second weight matrix to obtain the shallow pixel granularity feature vector of the hydraulic annular geological fractures after mapping.

7. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 6 is characterized in that: Calculating the semantic association score between any two mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of mapped shallow pixel granularity feature vectors of hydraulic annular geological fractures to obtain the shallow semantic association score topological matrix of hydraulic annular geological fractures composed of multiple shallow semantic association score values ​​of hydraulic annular geological fractures, including: Extracting the i-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture and the j-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture from the set of the mapped shallow pixel granularity feature vectors of the hydraulic annular geological fracture; Calculate the position difference between the i-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture and the j-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture to obtain the mapped shallow pixel granularity difference vector of the hydraulic annular geological fracture, and calculate the square of the norm of the mapped shallow pixel granularity difference vector of the hydraulic annular geological fracture to obtain the mapped shallow characteristic difference norm value of the hydraulic annular geological fracture; Calculate the difference between the constant one and the square of the one-norm of the shallow pixel granularity feature vector of the i-th mapped hydraulic annular geological fracture to obtain the shallow characteristic difference of the i-th mapped hydraulic annular geological fracture; Calculate the difference between the constant one and the square of the one-norm characteristic vector of the shallow pixel granularity of the j-th mapped hydraulic annular geological fracture to obtain the shallow characteristic difference of the j-th mapped hydraulic annular geological fracture; Calculate the product of the shallow characteristic difference value of the hydraulic annular geological fracture after the i-th mapping and the shallow characteristic difference value of the hydraulic annular geological fracture after the j-th mapping to obtain the shallow characteristic product value of the hydraulic annular geological fracture after the mapping; After dividing the mapped shallow-layer characteristic differential norm value of the hydraulic annular geological fracture by the mapped shallow-layer characteristic product value of the hydraulic annular geological fracture, the obtained value is first multiplied by a constant two and then added to a constant one to obtain a semantic metric value of the mapped shallow-layer characteristic of the hydraulic annular geological fracture; The mapped shallow feature semantic metric value of the hydraulic annular geological fracture is input into the arccosh function to obtain the shallow semantic association score value of the hydraulic annular geological fracture between the i-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture and the j-th mapped shallow pixel granularity feature vector of the hydraulic annular geological fracture.

8. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 7 is characterized in that: Based on the spatial distance between any two shallow pixel granularity feature vectors of hydraulic annular geological fractures in the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures, the shallow semantic association score topological matrix of the hydraulic annular geological fractures is subjected to spatial attention attenuation modulation to obtain the shallow semantic association spatial soft constraint score topological matrix of the hydraulic annular geological fractures, including: Calculating the spatial distance between any two shallow pixel granularity feature vectors of the hydraulic annular geological fractures to obtain a set of spatial distance values ​​of the hydraulic annular geological fractures; Calculate an exponential function with the natural constant e as the base, with each hydraulic annular geological fracture spatial distance value in the set of hydraulic annular geological fracture spatial distance values ​​as the exponent, to obtain a set of hydraulic annular geological fracture spatial distance index values; The set of the hydraulic annular geological fracture spatial distance index values ​​is multiplied by the attenuation modulation value according to the position to obtain a set of hydraulic annular geological fracture spatial distance weighted values; Each shallow semantic association score value of the hydraulic annular geological fractures in the shallow semantic association score topological matrix of the hydraulic annular geological fractures is divided by the corresponding weighted value of the spatial distance of the hydraulic annular geological fractures to obtain the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fractures composed of multiple shallow semantic association space soft constraint score values ​​of the hydraulic annular geological fractures.

9. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 8, characterized in that: Based on the shallow semantic association space soft constraint score topological matrix of the hydraulic annular geological fracture, a set of shallow pixel granularity feature vectors of the hydraulic annular geological fracture is enhanced by convolution coding to obtain the shallow enhanced features of the hydraulic annular geological fracture, including: Performing dilated convolution coding on the shallow semantic association space soft constraint score topology matrix of the hydraulic ring geological fractures to obtain the shallow semantic association space soft constraint score topology feature matrix of the hydraulic ring geological fractures; Inputting the set of shallow pixel granularity feature vectors of hydraulic annular geological fractures and the shallow semantic association space soft constraint score topological feature matrix of hydraulic annular geological fractures into a graph convolutional coding module to obtain a set of shallow context pixel granularity feature vectors of hydraulic annular geological fractures; The set of shallow context pixel granularity feature vectors of the hydraulic annular geological fractures is reshaped to obtain a shallow enhancement feature map of the hydraulic annular geological fractures as the shallow enhancement feature of the hydraulic annular geological fractures.

10. The method for checking the accuracy of hydraulic ring geological fracture measurement data according to claim 9, characterized in that: Based on the shallow enhanced features of the hydraulic annular geological fractures, an image semantic segmentation result is obtained, and based on the image semantic segmentation result, hydraulic annular geological fracture measurement intelligent generation data is obtained, including: performing image semantic segmentation on the shallow enhanced feature map of the hydraulic annular geological fractures to obtain the image semantic segmentation result, and obtaining the hydraulic annular geological fracture measurement intelligent generation data based on the image semantic segmentation result.

Images