Hydraulic engineering construction quality monitoring method and system based on image recognition

Abstract

The invention relates to the technical field of image analysis, in particular to a hydraulic engineering construction quality monitoring method and system based on image recognition, and the method comprises the following steps: analyzing the pixel intensity and gray gradient based on the image data of a construction region, building the adjacency characteristics of a local region, and generating an adjacency matrix through the spatial connection characteristics. According to the method, the pixel intensity and the gray gradient in the image data are analyzed, the spatial correlation characteristic weight distribution is established, the spatial characteristic analysis capability of the crack and the slope protection deformation is improved, the distribution characteristics of the crack and the deformation are optimized through path connection and edge weight extraction, the method is more accurate, the reliability of crack edge characteristic identification is enhanced, and the identification accuracy of the crack and the slope protection deformation is improved. The crack propagation direction and the slope protection deformation trend are comprehensively mastered by calculating node dynamic offset and connectivity characteristics, the prediction capacity of construction dynamic change is improved, local and global structure change characteristics are integrated, and quality monitoring from crack microscopic characteristics to the overall state of the slope protection is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and particularly to a method and system for monitoring the construction quality of water conservancy projects based on image recognition. Background Art

[0002] The technical field of image analysis is an important branch of computer vision, mainly studying how to extract meaningful information from images through algorithms and computational means. Its core lies in detecting, recognizing, segmenting, and classifying objects, features, patterns, etc. in images, and realizing functions such as object tracking, change detection, and feature matching according to specific requirements.

[0003] Among them, the method for monitoring the construction quality of water conservancy projects based on image recognition refers to using image analysis technology to perform specific quality monitoring and evaluation on the image data generated during the construction process of water conservancy projects. Its main purpose is to discover quality problems existing in construction in real time through image feature extraction and pattern analysis, improve the intelligence and refinement level of project management, and ensure the construction quality and safety of the project.

[0004] The existing technology has insufficient analysis of the spatial correlation between cracks and slope deformation, resulting in the inability to comprehensively identify the crack distribution and deformation characteristics in complex construction scenarios. The sensitivity of crack edge recognition to gradient fluctuations is low, which easily causes recognition errors and affects the accuracy of crack distribution. The ability to analyze dynamic offset and connectivity features is insufficient, making it difficult to accurately predict the crack propagation direction and dynamic changes in deformation, affecting the timeliness of construction quality monitoring. The lack of integration ability for local and global structural change characteristics fails to provide a comprehensive assessment of the quality change trend in the construction environment, which may lead to the failure to detect or handle engineering safety hazards in a timely manner. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art, and to propose a method and system for monitoring the construction quality of water conservancy projects based on image recognition.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solutions: A method for monitoring the construction quality of water conservancy projects based on image recognition, including the following steps: S1: Based on the image data of the construction area, analyze the pixel intensity and gray gradient, establish the adjacency characteristics of the local area, generate an adjacency matrix through spatial connection characteristics, establish the spatial correlation characteristics weight distribution for the dam crack distribution and slope deformation, and generate a crack topology weight matrix; S2: Use the crack topology weight matrix to extract the path connection between nodes and the edge weight, analyze the central offset of the slope deformation area and the gradient fluctuation of the crack edge, and obtain the optimized crack and deformation distribution characteristic values through hierarchical screening and adjusting the crack edge weight distribution; S3: Based on the optimized crack and deformation distribution characteristic values, analyze the node deformation and crack integrity, calculate the characteristic offset value and the change of the edge expansion direction within the region, integrate the local and global structure change characteristics, and generate the prediction values of the slope protection and crack abnormal characteristics; S4: Based on the prediction values of the slope protection and crack abnormal characteristics, analyze the displacement characteristics of the structural settlement and the crack expansion direction, extract the dynamic offset and connectivity characteristics of the nodes, calculate the crack length and the deformation characteristic distribution, and generate the dynamic prediction characteristic results of the cracks and slope protection; S5: Based on the dynamic prediction characteristic results of the cracks and slope protection, analyze the concrete surface stress and the structural settlement form, calculate the spatial characteristics of the crack length and the slope protection deformation, conduct classification and clustering analysis, and generate the risk warning indicators of the hydraulic structure.

[0007] The crack topological weight matrix is specifically an adjacency characteristic matrix, a spatial connection characteristic weight distribution, and a crack weight distribution matrix. The optimized crack and deformation distribution characteristic values include path connection characteristic values, edge weight distribution characteristic values, center offset characteristic values, and edge gradient fluctuation characteristic values. The prediction values of the slope protection and crack abnormal characteristics include node deformation characteristics, crack integrity characteristics, characteristic offset values, edge expansion direction change characteristics, and local and global structure change characteristics. The dynamic prediction characteristic results of the cracks and slope protection include crack length distribution, slope protection deformation distribution, node dynamic offset characteristics, connectivity characteristics, structural settlement displacement characteristics, and crack expansion direction characteristics. The risk warning indicators of the hydraulic structure include concrete surface stress characteristics, structural settlement form characteristics, crack length spatial characteristics, slope protection deformation spatial characteristics, and classification and clustering analysis results.

[0008] As a further solution of the present invention, the steps for obtaining the crack topological weight matrix are specifically as follows: S111: Extract the pixel intensity and gray gradient features of the construction area image data, and generate the adjacency characteristic distribution matrix of the local area by statistically analyzing the intensity difference and gradient change between adjacent pixels within the local area; S112: Based on the adjacency characteristic distribution matrix of the local area, calculate the spatial connection weights between pixels, and establish the spatial connection characteristic matrix between pixels by screening the significance of the connection weights and eliminating the connection relationships with low weight values; S113: Use the spatial connection characteristic matrix between pixels, combined with the spatial distribution of crack pixels, to calculate the spatial correlation weight value of each crack pixel, using the formula: ; Generate the crack topological weight matrix; Among them, represents the weight between the pixels in the crack topological weight matrix, represents the gray - scale gradient value of the th pixel, represents the gray - scale gradient value of the th pixel, is the spatial distance between pixel and pixel , represents the pixel intensity value of the th pixel, represents the pixel intensity value of the th pixel.

[0009] As a further solution of the present invention, the steps for obtaining the optimized crack and deformation distribution characteristic values are specifically as follows: S211: Extract the node - to - node path connections and edge weights in the crack topology weight matrix, calculate the central offset value of each slope protection deformation area, and generate the central offset characteristic distribution of the slope protection deformation area by statistically averaging the path lengths of all nodes in the area and comparing with the geometric center of the area; S212: Based on the central offset characteristic distribution of the slope protection deformation area, combined with the edge weights of the crack topology weight matrix, screen the crack paths with edge gradient fluctuations in the area, and establish a crack path gradient characteristic distribution matrix by calculating the gradient change rate and fluctuation range of each path; S213: Combine the crack path gradient characteristic distribution matrix with the central offset characteristic distribution, optimize the weight distribution of the crack edge, and by readjusting the gradient fluctuation amplitude of the crack edge, using the formula: ; Calculate the optimized crack edge weight distribution to obtain the optimized crack and deformation distribution characteristic values; Among them, is the optimized weight distribution value of the crack edge, is the gradient change value between node and , represents the central offset characteristic value of node in the crack path, represents the central offset characteristic value of node in the crack path, is the edge weight value on path , is the total number of nodes on the path, is the total number of edges in the path, is the base value of the natural logarithm, is the absolute difference of the central offset values between node and node , and the square - root term in the denominator represents the square root of the cumulative value of all edge weights in the path.

[0010] As a further solution of the present invention, the steps for obtaining the prediction value of the slope protection and crack abnormal characteristics are specifically as follows: S311: Based on the optimized crack and deformation distribution characteristic values, extract the node deformation characteristics, calculate the offset of the deformation amount of multiple nodes from the initial state, and generate the node deformation characteristic analysis result through differential analysis; S312: Screen the crack integrity data from the node deformation characteristic analysis result, evaluate the crack state of each node, and establish a crack integrity characteristic matrix through the width and length data of the cracks; S313: Combine the crack integrity characteristic matrix and the node deformation characteristic analysis result, analyze the characteristic offset value and the edge expansion direction change in the area, and use the formula: ; Calculate the change index of the regional characteristic offset and the edge expansion direction to obtain the prediction value of the slope protection and crack abnormal characteristics; Wherein, represents the change index of the regional characteristic offset and the edge expansion direction, is the deformation amount of node , is the influence factor of the crack at node , is the angle of node with respect to the regional center, is the weight value of node , is the total number of nodes, is the total number of cracks.

[0011] As a further solution of the present invention, the steps for obtaining the result of the dynamic prediction characteristics of the crack and the slope protection are specifically as follows: S411: Based on the prediction value of the slope protection and crack abnormal characteristics, extract the displacement characteristics of the structural settlement, calculate the displacement amount of each structural node, and generate the displacement characteristic analysis result of the structural settlement through comprehensive node dynamic offset data; S412: Use the displacement characteristic analysis result of the structural settlement to evaluate the crack propagation direction, and generate the crack propagation direction analysis result by calculating the vector direction difference between the starting and ending nodes of the crack; S413: Combine the crack propagation direction analysis result and the displacement characteristic analysis result of the structural settlement, extract the dynamic offset and connectivity characteristics of the nodes, and calculate the distance change between each node and its adjacent nodes by using the formula: ; Obtain the crack length and deformation characteristic distribution; Wherein, Indicates the distribution of crack length and deformation characteristics is the displacement of the node is the average value of the displacements of all nodes is the node crack length is the node and the node distance is the total number of nodes is the node number of connected nodes; S414: Utilize the distribution of crack length and deformation characteristics, integrate local and global structural change characteristics, and generate the results of dynamic prediction characteristics of cracks and slope protection by comparing and analyzing the correlation between local deformation and global deformation.

[0012] As a further solution of the present invention, the steps for obtaining the risk warning index of the water conservancy structure are specifically as follows: S511: Analyze the results of dynamic prediction characteristics of cracks and slope protection, measure the stress distribution on the concrete surface, calculate the settlement form of structural nodes, and generate the analysis result of structural settlement form by comparing the displacement data of different nodes; S512: Utilize the analysis result of structural settlement form, apply multi-dimensional space analysis technology, calculate the spatial characteristics of crack length and slope protection deformation, and generate the description of spatial characteristics of crack and slope protection deformation through distance and direction analysis between nodes; S513: Input the description of spatial characteristics of crack and slope protection deformation into classification and clustering analysis, stratify the influence degree of cracks by multivariate statistical methods, and adopt the formula: ; Calculate and generate the risk warning index of the water conservancy structure; Among them, represents the risk warning index of the water conservancy structure, is the spatial characteristics of crack and slope protection deformation in the region, is the influence level of regional structural settlement, is the weight index of the region, reflecting the key nature of regional risk, is the adjustment coefficient, referring to the influence of external environment and engineering measures, is the total number of analysis regions,

[0013] ​A water conservancy project construction quality monitoring system based on image recognition, which is used to execute the above-mentioned water conservancy project construction quality monitoring method based on image recognition. The system includes: The crack topology matrix generation module extracts the image pixel intensity value and the gray gradient value based on the construction area image data, calculates the connection relationship between adjacent pixels and the spatial region segmentation value, calculates the local region adjacency weight value and the spatial correlation value, and generates a crack topology weight matrix; The crack and deformation characteristic optimization module extracts the node path connectivity value and the edge weight value based on the crack topology weight matrix, calculates the slope deformation center offset value and the crack edge gradient fluctuation value, adjusts the edge node weight distribution and recalculates the crack characteristics, and generates an optimized crack and deformation distribution characteristic value; The abnormal characteristic analysis module extracts the node deformation value and the crack integrity value based on the optimized crack and deformation distribution characteristic value, calculates the characteristic offset value and the edge expansion direction change value, integrates the local and global structure change values to calculate the abnormal distribution characteristic, and generates a slope and crack abnormal characteristic prediction value; The dynamic prediction characteristic generation module extracts the settlement displacement characteristic value and the crack expansion direction value based on the slope and crack abnormal characteristic prediction value, calculates the dynamic offset value and the connectivity characteristic value, analyzes the crack length change value and the deformation characteristic distribution value, and generates a crack and slope dynamic prediction characteristic result; The risk warning index calculation module extracts the concrete surface stress value and the settlement form value based on the crack and slope dynamic prediction characteristic result, calculates the crack length distribution value and the slope deformation spatial characteristic value, classifies and calculates and analyzes the risk distribution characteristic value, and generates a water conservancy structure risk warning index.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by analyzing the pixel intensity and gray gradient in the image data, the spatial correlation characteristic weight distribution is established, which improves the spatial characteristic analysis ability of cracks and slope deformations. The path connection and edge weight extraction optimize the distribution characteristics of cracks and deformations, making them more accurate and enhancing the reliability of crack edge characteristic recognition. By calculating the node dynamic offset and connectivity characteristics, the crack expansion direction and slope deformation trend are comprehensively grasped, improving the prediction ability of construction dynamic changes. Integrating the local and global structure change characteristics realizes the quality monitoring from the microscopic characteristics of cracks to the overall state of slopes, constructs a risk warning mechanism based on dynamic characteristic analysis, and ensures the safety and quality stability of the construction process. Brief Description of the Drawings

[0015] Figure 1 It is a schematic diagram of the working process of the present invention; Figure 2Flow chart of the steps for obtaining the crack topology weight matrix of the present invention; Figure 3 Flow chart of the steps for obtaining the optimized crack and deformation distribution characteristic values of the present invention; Figure 4 Flow chart of the steps for obtaining the predicted values of slope protection and crack abnormal characteristics of the present invention; Figure 5 Flow chart of the steps for obtaining the results of dynamic prediction characteristics of cracks and slope protection of the present invention; Figure 6 Flow chart of the steps for obtaining the risk warning indicators of the water conservancy structure of the present invention. Detailed implementation manners

[0016] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0018] Embodiment 1: Please refer to Figure 1 , the present invention provides a technical solution: a method for monitoring the construction quality of water conservancy projects based on image recognition, including the following steps: S1: Based on the image data of the construction area, analyze the pixel intensity and gray scale gradient, establish the adjacency characteristics of the local area, generate an adjacency matrix through the spatial connection characteristics, establish the spatial correlation characteristic weight distribution for the dam crack distribution and slope protection deformation, and generate the crack topology weight matrix; S2: Utilize the crack topology weight matrix, extract the path connection between nodes and the edge weights, analyze the central offset of the slope protection deformation area and the crack edge gradient fluctuation, and obtain the optimized crack and deformation distribution characteristic values through hierarchical screening and adjustment of the crack edge weight distribution; S3: Based on the optimized crack and deformation distribution characteristic values, analyze the node deformation and crack integrity, calculate the characteristic offset value within the area and the change in the edge expansion direction, integrate the local and global structure change characteristics, and generate the predicted values of slope protection and crack abnormal characteristics; S4: Based on the predicted values of the abnormal characteristics of the slope protection and cracks, analyze the displacement characteristics of the structural settlement and the crack propagation direction, extract the dynamic offset and connectivity characteristics of the nodes, calculate the crack length and the distribution of deformation characteristics, and generate the results of the dynamic prediction characteristics of cracks and slope protection; S5: Based on the results of the dynamic prediction characteristics of cracks and slope protection, analyze the surface stress of the concrete and the structural settlement form, calculate the spatial characteristics of the crack length and the slope protection deformation, conduct classification and clustering analysis, and generate the risk warning indicators for the hydraulic structure.

[0019] The crack topological weight matrix specifically includes the adjacency characteristic matrix, the spatial connection characteristic weight distribution, and the crack weight distribution matrix. The optimized crack and deformation distribution characteristic values include the path connection characteristic value, the edge weight distribution characteristic value, the central offset characteristic value, and the edge gradient fluctuation characteristic value. The predicted values of the abnormal characteristics of the slope protection and cracks include the node deformation characteristics, the crack integrity characteristics, the characteristic offset value, the change characteristic of the edge expansion direction, and the local and global structure change characteristics. The results of the dynamic prediction characteristics of cracks and slope protection include the crack length distribution, the slope protection deformation distribution, the dynamic offset characteristics of the nodes, the connectivity characteristics, the structural settlement displacement characteristics, and the crack propagation direction characteristics. The risk warning indicators for the hydraulic structure include the surface stress characteristics of the concrete, the structural settlement form characteristics, the spatial characteristics of the crack length, the spatial characteristics of the slope protection deformation, and the classification and clustering analysis results.

[0020] Please refer to Figure 2 , and the steps for obtaining the crack topological weight matrix are specifically as follows: S111: Extract the pixel intensity and gray-scale gradient features of the image data in the construction area. By statistically analyzing the intensity differences and gradient changes between adjacent pixels in the local area, generate the adjacency characteristic distribution matrix of the local area; First, collect the image data of the construction area, and extract the intensity information of each pixel point from it. At the same time, calculate the gray-scale gradient of these pixel points. The gray-scale gradient is obtained by calculating the intensity difference between the pixel point and its neighboring pixels. The key to this step is to be able to reveal the texture and edge information in the image. Further, these data are used to analyze the similarity and difference between pixel points. In order to process these data more effectively, statistical methods are used to calculate the intensity differences and gradient changes between adjacent pixels in the local area. These calculation results will be used to generate the adjacency characteristic distribution matrix of the local area, which details the adjacency relationship between each pixel point and is the basis for understanding and analyzing the characteristics of the construction area.

[0021] S112: Based on the adjacency characteristic distribution matrix of the local area, calculate the spatial connection weights between pixels. By screening the significance of the connection weights and eliminating the connection relationships with low weight values, establish the spatial connection characteristic matrix between pixels; This calculation process not only considers the physical distance between pixels, but also takes into account their intensity and gray-scale gradient differences. The calculation of spatial connection weights is achieved through a specialized formula that considers the product of the reciprocal of the distance between pixels and the gradient difference, thereby reflecting the spatial correlation between pixels. By screening out highly significant connections and eliminating those with low weight values, an accurate spatial connection characteristic matrix is established. This matrix is a key tool for analyzing and judging the internal structure and potential problems of image regions, such as cracks and deformations.

[0022] S113: Using the spatial connection characteristic matrix between pixels and combining with the spatial distribution of crack pixels, calculate the spatial correlation weight value of each crack pixel, using the formula: ; Generate a crack topology weight matrix; Among them, represents the weight between the th pixel in the crack topology weight matrix, represents the th gray-scale gradient value of a pixel, represents the th gray-scale gradient value of a pixel, is the spatial distance between pixel and pixel , represents the th pixel intensity value of a pixel, represents the th pixel intensity value of a pixel.

[0023] Formula: ; The advantage of the formula is that by combining pixel intensity and distance factors, it can accurately reflect the spatial correlation between pixels in the image. Especially when dealing with cracks and deformations in the image, it can effectively identify and analyze the formation pattern of cracks and their propagation paths, enhancing the accuracy and practicality of image analysis.

[0024] Detailed explanation of the formula and the derivation process of formula calculation: Consider two pixel points i and j, where and represent their gray-scale gradient values respectively, and these values can be directly calculated through image processing software. If the gray-scale gradient of pixel i is 120 and the gray-scale gradient of pixel j is 100, then the gray-scale gradient difference is , represents the spatial distance between these two pixels. For example, if they are separated by 1 pixel, then , and are the intensity values of these two pixels, assumed to be 200 and 180 respectively. Substituting them into the formula gives: ; This result shows that the weight value is 0.368, indicating that these two pixels are relatively close in space and there is a certain gray-scale gradient difference. This calculation method provides a quantitative method to evaluate the spatial correlation between pixels.

[0025] Please refer to Figure 3 , and the specific steps for optimizing the acquisition of the crack and deformation distribution characteristic values are as follows: S211: Extract the node-to-node path connections and edge weights in the crack topology weight matrix, calculate the central offset value of each slope protection deformation area, and generate the central offset characteristic distribution of the slope protection deformation area by statistically averaging the path lengths of all nodes within the area and comparing with the geometric center of the area; First, extract the connection status of each node according to the image data, and then calculate the edge weights. This process takes into account the pixel intensity difference and gray-scale gradient, and the spatial distance between pixels is also used as an important factor in calculating the edge weights. Through these data, a characteristic distribution model describing the central offset of each area can be constructed. The connection strength of each node and the edge weights together reflect the spatial distribution of cracks and deformations. This data processing step is based on the crack topology weight matrix and the actual measurement data of the slope protection deformation area, ensuring the accuracy and practicality of the analysis results, and thus obtaining the central offset characteristic distribution of the slope protection deformation area.

[0026] S212: Based on the central offset characteristic distribution of the slope protection deformation area, combined with the edge weights of the crack topology weight matrix, screen the crack paths with edge gradient fluctuations within the area, and establish a crack path gradient characteristic distribution matrix by calculating the gradient change rate and fluctuation range of each path; First, screen out the paths with significant gradient fluctuations from the crack topology weight matrix. This screening is based on the distribution of edge weights and the spatial connection characteristics of cracks. The gradient fluctuation of each path is evaluated by calculating the gray-scale gradient change on the path, taking into account the relative positions and connection strengths of the nodes within the path. Through these analyses, a crack path gradient characteristic distribution matrix is established. This not only involves calculations and data processing but also includes logical analysis and practical applications of the data, making the obtained gradient characteristic distribution more operable and applicable.

[0027] S213: Combine the crack path gradient characteristic distribution matrix with the central offset characteristic distribution to optimize the weight distribution of the crack edges. By readjusting the gradient fluctuation amplitude of the crack edges, using the formula: ; Calculate the optimized crack edge weight distribution to obtain the optimized crack and deformation distribution characteristic values; Among them, is the optimized weight distribution value of the crack edge, is the node and The gradient change value between them, represents the central offset characteristic value of the node in the crack path, represents the central offset characteristic value of the node in the crack path, is the edge weight value on the path is the total number of nodes on the path, is the total number of edges in the path, is the base value of the natural logarithm, is the node and the node The absolute difference of the central offset values. The square root term in the denominator represents the square root of the sum of all edge weights in the path.

[0028] Formula: ; The benefit of the formula is that it allows dynamic adjustment of the weights of the crack edges by combining the crack path gradient characteristic distribution matrix and the central offset characteristic distribution, and optimizing according to the actual changes in the spatial position and gradient of the crack. This approach enhances the adaptability of the model to crack behavior under different geographical and environmental conditions.

[0029] Detailed explanation of the formula and the process of formula calculation and derivation: Assume that there are three nodes in a specific crack area, and the gradient change between the nodes is [2, 3, 4], and the central offset and are [1, 2, 3] and [1, 1, 2] respectively, and the edge weight is [5, 5, 5]. Then, according to the formula, the contribution of each pair of nodes is calculated as: ; The total weight distribution is calculated as: ; The result shows that the optimized crack edge weight distribution value is approximately 1.13, which indicates that by considering the gradient change and central offset between nodes, the weight of the crack edge is appropriately enhanced, helping to more accurately describe the distribution characteristics of cracks and deformations, and thus guiding actual engineering adjustment and optimization measures.

[0030] Please refer to​Figure 4 , the steps for obtaining the predicted values of the abnormal characteristics of the slope protection and cracks are specifically as follows: S311: Based on the optimized crack and deformation distribution characteristics values, extract the node deformation characteristics, calculate the offset of the deformation amount of multiple nodes from the initial state, and generate the node deformation characteristics analysis result through differential analysis; First, extract the position and shape of each node from the initial measurement data of the nodes. These data are collected immediately after the initial construction by the site monitoring equipment. Then, calculate the real-time deformation amount of each node, which requires comparing the current node state with the initial state. Differential analysis relies on high-precision terrain change monitoring technologies such as differential global positioning system (DGPS) and laser scanning. These technologies can provide millimeter-level accuracy to ensure the accuracy of the analysis results. The calculated deformation amount will reflect the physical changes of the structure after being stressed and provide basic data for the subsequent crack integrity assessment. The data processing technologies used in this process include signal filtering and data fusion to ensure the extraction of effective signals from environmental noise and generate the node deformation characteristics analysis result.

[0031] S312: From the node deformation characteristics analysis result, screen the crack integrity data, evaluate the crack state of each node, and establish a crack integrity characteristics matrix through the width and length data of the cracks; This process involves the precise measurement of the width and length of the cracks in each node. These data are obtained through crack monitoring sensors. These sensors can monitor the development state of the cracks in real time and automatically record the data. The evaluation of the cracks is carried out through image analysis software. This software can process and analyze a large number of crack images, automatically identify the crack edges and calculate their sizes. The established crack integrity characteristics matrix reflects the health state of the structure. Through the statistical analysis of the data, predict the formation and development trend of the cracks. The prediction algorithm is based on data modeling to speculate on the future crack behavior. This process ensures the timeliness and effectiveness of the crack management measures.

[0032] S313: Combine the crack integrity characteristics matrix and the node deformation characteristics analysis result, analyze the characteristic offset value and the change of the edge expansion direction in the area, and use the formula: ; Calculate the change index of the regional characteristic offset and the edge expansion direction to obtain the predicted values of the abnormal characteristics of the slope protection and cracks; Among them, represents the change index of the regional characteristic offset and the edge expansion direction, is the deformation amount of node , is the influence factor of the crack at node , is node For the angle of the regional center, is the node weight value, is the total number of nodes, is the total number of cracks.

[0033] Formula: ; The advantage of the formula is that it can comprehensively consider the deformation of nodes and the spatial distribution characteristics of cracks. By exponentially decaying the influencing factors, it emphasizes the importance of cracks near the regional center and enhances the model's ability to predict the development trend of cracks.

[0034] Detailed explanation of the formula and the derivation process of formula calculation: Set the total number of crack nodes n = 5, the total number of edge nodes m = 3, the node deformation = [0.2, 0.5, 0.1, 0.3, 0.4] (unit: mm), the influencing factor λ = [1, 0.5, 2, 1.5, 1], the node angle θ = [45°, 30°, 60°, 90°, 120°] (unit: degree), the node weight w = [1, 2, 1]. Fill these parameters with actual monitoring data and calculate: ; The result shows that the change index of the regional characteristic deviation and the edge expansion direction is 0.245, which characterizes the degree of overall structural change. This value reflects the comprehensive influence of cracks and deformations on structural stability and is used to evaluate the health status of the structure and predict future deformation trends.

[0035] Please refer to Figure 5 , and the steps to obtain the results of the dynamic prediction characteristics of cracks and slope protection are specifically as follows: S411: Based on the predicted values of the abnormal characteristics of slope protection and cracks, extract the displacement characteristics of structural settlement, calculate the displacement of each structural node, and generate the analysis result of the displacement characteristics of structural settlement by comprehensively considering the dynamic offset data of nodes; First, it involves the analysis of the displacement characteristics of structural settlement. This process requires monitoring the position changes of each node within a specific time. Through geological scanning equipment and displacement sensors, the displacement data of each node are collected. These data are processed by spatial analysis software to obtain the displacement of each node. Then, statistical analysis is performed on these displacements to calculate the offset between the node and its initial position. Through a high-precision geographic information system, these displacement data can be compared with the original geographical location of the node to ensure the accurate calculation of the displacement and generate the analysis result of the displacement characteristics of structural settlement. This result will be directly used for the subsequent evaluation of the crack propagation direction.

[0036] S412: Analyze the crack propagation direction using the analysis results of the displacement characteristics of structural settlement. By calculating the vector direction difference between the starting and ending nodes of the crack, generate the analysis results of the crack propagation direction. In this process, it is first necessary to determine the positions of the starting and ending points of the crack, which can be obtained through crack monitoring instruments such as crack width gauges and depth detectors. The collected data needs to be processed by crack analysis software to calculate the direction vector of the crack. This calculation involves vector analysis. The direction vector is calculated based on the coordinate differences between the two endpoints of the crack. The differences between these vectors will show the crack propagation trend and generate the analysis results of the crack propagation direction. The key to this result is to reveal the development dynamics and possible propagation trends of the crack.

[0037] S413: Combine the analysis results of the crack propagation direction and the analysis results of the displacement characteristics of structural settlement, extract the dynamic offset and connectivity characteristics of the nodes, and calculate the distance change between each node and its neighboring nodes using the formula: ; Obtain the crack length and deformation characteristic distribution. Among them, represents the crack length and deformation characteristic distribution, is the displacement of node , is the average value of the displacements of all nodes, is the crack length of node , is the distance between node and node , is the total number of nodes, is the number of nodes connected to node ; Formula: ; The advantage of the formula is that by introducing the distance change between nodes and the absolute deviation of node displacements, it can more accurately quantify the dynamic changes of cracks and their impact on the surrounding structure, which helps to make more detailed predictions and management of the dynamic behavior of cracks.

[0038] Detailed explanation of the formula and the derivation process of formula calculation: Suppose there are 5 nodes, and the displacements are 2 cm, 3 cm, 4 cm, 5 cm, and 6 cm respectively, and the average displacement is 4 cm. The crack length of each node is assumed to be 10 cm, and the distance is assumed to be 1 m. The calculation process is as follows: ; ; The result shows that the comprehensive influence value of the dynamic offset of the node and the crack length is 12. This value represents the comprehensive quantitative index of the crack length and the distribution of deformation characteristics, which is used to evaluate the dynamic changes of the crack and the possible impact on the structure. This result directly reflects the dynamic behavior of the crack and the potential threat to the structural safety.

[0039] S414: Utilize the crack length and the distribution of deformation characteristics, integrate the local and global structural change characteristics, and generate the results of the dynamic prediction characteristics of the crack and the slope protection by comparing and analyzing the correlation between the local deformation and the global deformation.

[0040] First, it is required to conduct a multi-dimensional evaluation of the crack length and deformation characteristic data. Through data fusion by structural monitoring software and combined with GIS (Geographic Information System) technology, the deformation conditions of each part in the structure and the precise positions of the cracks are accurately mapped. In addition, statistical analysis methods are needed to compare the local deformation data with the global deformation data to find out the correlations and differences. This step is achieved by calculating the correlation coefficient between the local deformation data and the global data. The level of the correlation coefficient will directly affect the accuracy of the prediction characteristic results. The generated results of the dynamic prediction characteristics of the crack and the slope protection will provide a comprehensive perspective to observe and predict the future safety status of the structure.

[0041] Please refer to Figure 6 , and the specific steps for obtaining the risk warning indicators of the hydraulic structure are as follows: S511: Analyze the results of the dynamic prediction characteristics of the crack and the slope protection, measure the stress distribution on the concrete surface, calculate the settlement form of the structural nodes, and generate the analysis result of the structural settlement form by comparing the displacement data of the differential nodes; First, it is necessary to extract the stress distribution on the concrete surface from the existing dataset, measure the stress values at specific points using stress sensors, and collect the displacement data of the corresponding nodes, which are obtained after displacement sensors are installed at different positions of the structure body. Each sensor will record the displacement changes of the corresponding nodes in the time series, so as to be used to calculate the displacement differences between the nodes. Next, by comparing and analyzing these displacement data, calculate the displacement difference between node A and node B, so as to understand the form and trend of the structural settlement. Use data analysis software, such as Matlab or R, to conduct statistical analysis on these displacement difference data, and predict the settlement trend of the structure by establishing a model to generate the analysis result of the structural settlement form. This result will directly affect the subsequent engineering decisions and the setting of the warning system.

[0042] S512: Utilize the analysis result of the structural settlement form, apply multi-dimensional space analysis technology, calculate the spatial characteristics of the crack length and the slope protection deformation, and generate the description of the spatial characteristics of the crack and the slope protection deformation through the analysis of the distance and directionality between the nodes. In this process, specific data within the crack area need to be collected, including the start and end positions, length, and width of the cracks. These data are usually obtained through on-site measurements or aerial photography using drones. Subsequently, using GIS software, these data are input for spatial analysis to analyze the distribution pattern of the cracks and their relationship with the deformation of the slope protection. By establishing a spatial analysis model, the impact of different crack parameters on the stability of the slope protection and the potential direction of crack propagation are simulated, thereby generating a description of the spatial characteristics of the cracks and slope protection deformation. This description will be used to evaluate and predict the stability of the slope protection and potential risk points.

[0043] S513: Input the description of the spatial characteristics of the cracks and slope protection deformation into the classification and clustering analysis. Stratify the degree of crack influence through multivariate statistical methods, using the formula: ; Calculate and generate the risk warning index for the hydraulic structure; Among them, represents the risk warning index for the hydraulic structure, is the spatial characteristics of the cracks and slope protection deformation in the th area, is the structural settlement influence level of the area, is the weight index of the area, reflecting the key nature of the area risk, is the adjustment coefficient, referring to the influence of the external environment and engineering measures, is the total number of analysis areas, is the number of types of adjustment coefficients.

[0044] Formula: ; The advantage of the formula is that it takes into account the combined effect of the spatial characteristics of the cracks and slope protection deformation and the structural settlement influence level, and adjusts the degree of attention to the risk of each area through the weight index , while considering the adjustment factor of the external environment and engineering measures.

[0045] Detailed explanation of the formula and the derivation process of the formula calculation: Suppose there are three areas, among which , , respectively represent the spatial characteristics of the cracks and slope protection deformation in the three areas, , , is the structural settlement influence level, the weight index , , , the adjustment factor , , : ; ; ; ; ; The result shows that the overall risk rating of the structure is 3.191, which characterizes the comprehensive risk assessment based on the spatial characteristics of cracks and slope protection deformation and the structure settlement within the current analysis area, and is further used to determine the monitoring frequency and the priority of reinforcement measures.

[0046] A water conservancy project construction quality monitoring system based on image recognition, which is used to execute the above-mentioned water conservancy project construction quality monitoring method based on image recognition. The system includes: The crack topology matrix generation module extracts the image pixel intensity value and the gray scale gradient value based on the construction area image data, calculates the connection relationship between adjacent pixels and the spatial region segmentation value, calculates the local region adjacency weight value and the spatial association value, and generates a crack topology weight matrix; The crack and deformation characteristic optimization module extracts the node path connectivity value and the edge weight value based on the crack topology weight matrix, calculates the slope protection deformation center offset value and the crack edge gradient fluctuation value, adjusts the edge node weight distribution and recalculates the crack characteristics, and generates an optimized crack and deformation distribution characteristic value; The abnormal characteristic analysis module extracts the node deformation value and the crack integrity value based on the optimized crack and deformation distribution characteristic value, calculates the characteristic offset value and the edge extension direction change value, integrates the local and global structure change values to calculate the abnormal distribution characteristic, and generates a slope protection and crack abnormal characteristic prediction value; The dynamic prediction characteristic generation module extracts the settlement displacement characteristic value and the crack propagation direction value based on the slope protection and crack abnormal characteristic prediction value, calculates the dynamic offset value and the connectivity characteristic value, analyzes the crack length change value and the deformation characteristic distribution value, and generates a crack and slope protection dynamic prediction characteristic result; The risk warning index calculation module extracts the concrete surface stress value and the settlement form value based on the crack and slope protection dynamic prediction characteristic result, calculates the crack length distribution value and the slope protection deformation spatial characteristic value, classifies and calculates and analyzes the risk distribution characteristic value, and generates a water conservancy structure risk warning index.

[0047] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the relevant art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as it does not depart from the technical solution content of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for monitoring the construction quality of water conservancy projects based on image recognition, characterized in that, It includes the following steps: S1: Based on the image data of the construction area, analyze the pixel intensity and gray gradient, establish the adjacency characteristics of the local area, generate an adjacency matrix through the spatial connection characteristics, establish the spatial correlation characteristic weights for the distribution of dam cracks and slope protection deformation, and generate a crack topology weight matrix; S2: Utilize the crack topology weight matrix, extract the path connections between nodes and edge weights, analyze the central offset of the slope protection deformation area and the gradient fluctuation of the crack edge, and obtain the optimized crack and deformation distribution characteristic values through hierarchical screening and adjustment of the crack edge weight distribution; S3: Based on the optimized crack and deformation distribution characteristic values, analyze the node deformation and crack integrity, calculate the characteristic offset value and the change in the edge expansion direction within the area, integrate the local and global structure change characteristics, and generate the prediction values of the abnormal characteristics of the slope protection and cracks; S4: Based on the prediction values of the abnormal characteristics of the slope protection and cracks, analyze the displacement characteristics of the structural settlement and the crack expansion direction, extract the dynamic offset and connectivity characteristics of the nodes, calculate the crack length and deformation characteristic distribution, and generate the dynamic prediction characteristic results of the cracks and slope protection; S5: Based on the dynamic prediction characteristic results of the cracks and slope protection, analyze the surface stress of the concrete and the structural settlement form, calculate the spatial characteristics of the crack length and slope protection deformation, conduct classification and clustering analysis, and generate the risk warning indicators for the water conservancy structure.

2. The method for monitoring the construction quality of hydraulic engineering based on image recognition according to claim 1, characterized in that The crack topology weight matrix specifically includes an adjacency characteristic matrix, a spatial connection characteristic weight distribution, and a crack weight distribution matrix. The optimized crack and deformation distribution characteristic values include path connection characteristic values, edge weight distribution characteristic values, central offset characteristic values, and edge gradient fluctuation characteristic values. The prediction values of the abnormal characteristics of the slope protection and cracks include node deformation characteristics, crack integrity characteristics, characteristic offset values, edge expansion direction change characteristics, and local and global structure change characteristics. The dynamic prediction characteristic results of the cracks and slope protection include crack length distribution, slope protection deformation distribution, node dynamic offset characteristics, connectivity characteristics, structural settlement displacement characteristics, and crack expansion direction characteristics. The risk warning indicators for the water conservancy structure include concrete surface stress characteristics, structural settlement form characteristics, crack length spatial characteristics, slope protection deformation spatial characteristics, and classification and clustering analysis results.

3. The method for monitoring the construction quality of water conservancy projects based on image recognition according to claim 2, characterized in that, The specific steps for obtaining the crack topology weight matrix are as follows: S111: Extract the pixel intensity and gray gradient features of the image data of the construction area, and generate an adjacency characteristic distribution matrix of the local area by statistically analyzing the intensity difference and gradient change between adjacent pixels within the local area; S112: Based on the adjacency characteristic distribution matrix of the local area, calculate the spatial connection weights between pixels, and establish a spatial connection characteristic matrix between pixels by screening the significance of the connection weights and eliminating the connection relationships with low weight values; S113: Adopt the spatial connection characteristic matrix between pixels, combine with the spatial distribution of crack pixels, calculate the spatial correlation weight value of each crack pixel, using the formula: ； Generate a crack topology weight matrix; Among them, represents the weight between the pixels in the crack topology weight matrix, represents the gray gradient value of the th pixel, represents the gray gradient value of the th pixel, is the spatial distance between pixel and pixel represents the pixel intensity value of the th pixel, represents the pixel intensity value of the th pixel.

4. The method for monitoring the construction quality of hydraulic engineering based on image recognition according to claim 3, wherein The specific steps for obtaining the optimized crack and deformation distribution characteristic values are as follows: S211: Extract the path connections between nodes and edge weights in the crack topology weight matrix, calculate the central offset value of each slope protection deformation area, and generate the central offset characteristic distribution of the slope protection deformation area by statistically averaging the path lengths of all nodes in the area and comparing with the geometric center of the area; S212: Based on the central offset characteristic distribution of the slope protection deformation area and combined with the edge weights of the crack topology weight matrix, screen the crack paths with edge gradient fluctuations in the area, and establish a crack path gradient characteristic distribution matrix by calculating the gradient change rate and fluctuation range of each path; S213: Combine the crack path gradient characteristic distribution matrix with the central offset characteristic distribution to optimize the weight distribution of the crack edge. By readjusting the gradient fluctuation amplitude of the crack edge, use the formula: ； Calculate the optimized crack edge weight distribution to obtain the optimized crack and deformation distribution characteristic values; Among them, is the optimized weight distribution value at the crack edge, is the node and is the gradient change value between them, represents the central offset characteristic value of the node in the crack path, represents the central offset characteristic value of the node in the crack path, is the edge weight value on the path , is the total number of nodes on the path, is the total number of edges in the path, is the base value of the natural logarithm, is the node and the node is the absolute difference of the central offset values. The square root term in the denominator represents the square root of the cumulative value of all edge weights in the path.

5. The method for monitoring the construction quality of water conservancy projects based on image recognition according to claim 4, wherein, The steps for obtaining the prediction value of the slope protection and crack anomaly characteristics are specifically as follows: S311: Based on the optimized crack and deformation distribution characteristic values, extract the node deformation characteristics, calculate the deformation amount of multiple nodes and the offset from the initial state, and generate the analysis result of node deformation characteristics through differential analysis; S312: Screen the crack integrity data from the analysis result of node deformation characteristics, evaluate the crack state of each node, and establish a crack integrity characteristic matrix through the width and length data of the cracks; S313: Combine the crack integrity characteristic matrix with the analysis result of node deformation characteristics, analyze the characteristic offset value and edge expansion direction change in the area, and use the formula: ； Calculate the change index of the area characteristic offset and edge expansion direction to obtain the prediction value of the slope protection and crack anomaly characteristics; Among them, represents the change index of the regional characteristic offset and the edge expansion direction, is the deformation amount of the node , is the influence factor of the crack at the node , is the angle of the node with respect to the regional center, is the weight value of the node , is the total number of nodes, is the total number of cracks.

6. The method for monitoring the construction quality of water conservancy projects based on image recognition according to claim 5, characterized in that, The steps for obtaining the result of the dynamic prediction characteristics of the crack and slope protection are specifically as follows: S411: Based on the prediction value of the slope protection and crack anomaly characteristics, extract the displacement characteristics of the structural settlement, calculate the displacement of each structural node, and generate the analysis result of the displacement characteristics of the structural settlement by comprehensively analyzing the dynamic offset data of the nodes; S412: Use the analysis result of the displacement characteristics of the structural settlement to evaluate the crack propagation direction, and generate the analysis result of the crack propagation direction by calculating the vector direction difference between the starting and ending nodes of the crack; S413: Combine the analysis result of the crack propagation direction and the analysis result of the displacement characteristics of the structural settlement, extract the dynamic offset and connectivity characteristics of the nodes, and calculate the distance change between each node and its neighboring nodes by using the formula: ； Obtain the crack length and deformation characteristic distribution; Among them, represents the crack length and deformation characteristic distribution, is the displacement of node , is the average value of the displacements of all nodes, is the crack length of node , is the distance between node and node , is the total number of nodes, is the number of nodes connected to node ; S414: Use the crack length and deformation characteristic distribution to integrate the local and global structural change characteristics, and generate the result of the dynamic prediction characteristics of the crack and slope protection by comparing and analyzing the correlation between local and global deformations.

7. The method for monitoring the construction quality of hydraulic engineering based on image recognition according to claim 6, wherein The steps for obtaining the risk warning index of the hydraulic structure are specifically as follows: S511: Analyze the result of the dynamic prediction characteristics of the crack and slope protection, measure the stress distribution on the concrete surface, calculate the settlement form of the structural nodes, and generate the analysis result of the structural settlement form by comparing the displacement data of different nodes; S512: Using the results of the structural settlement pattern analysis, apply multi-dimensional space analysis techniques to calculate the spatial characteristics of the crack length and slope protection deformation. Through the analysis of the distance and directionality between nodes, generate a description of the spatial characteristics of the cracks and slope protection deformation; S513: Input the description of the spatial characteristics of the cracks and slope protection deformation into the classification and clustering analysis. Stratify the degree of influence of the cracks through multivariate statistical methods, and use the formula: ； Calculate and generate the risk warning index for the hydraulic structure; Among them, represents the risk warning index of the water conservancy structure, is the spatial characteristics of cracks and slope protection deformation in the th area, is the influence level of structural settlement in the area, is the weight index of the area, reflecting the key nature of the area risk, is the adjustment coefficient, referring to the influence of the external environment and engineering measures, is the total number of analyzed areas, is the number of types of adjustment coefficients.

8. A water conservancy project construction quality monitoring system based on image recognition, characterized in that, According to the method for monitoring the construction quality of hydraulic engineering based on image recognition according to any one of claims 1-7, the system includes: The crack topology matrix generation module extracts the image pixel intensity value and gray-scale gradient value based on the image data of the construction area, calculates the connection relationship between adjacent pixels and the spatial region segmentation value, calculates the local area adjacency weight value and the spatial correlation value, and generates a crack topology weight matrix; The crack and deformation characteristic optimization module extracts the node path connectivity value and edge weight value based on the crack topology weight matrix, calculates the offset value of the slope protection deformation center and the crack edge gradient fluctuation value, adjusts the weight distribution of the edge nodes and recalculates the crack characteristics, and generates an optimized crack and deformation distribution characteristic value; The abnormal characteristic analysis module extracts the node deformation value and crack integrity value based on the optimized crack and deformation distribution characteristic value, calculates the characteristic offset value and the edge expansion direction change value, integrates the local and global structure change values to calculate the abnormal distribution characteristic, and generates a prediction value for the abnormal characteristics of the slope protection and cracks; The dynamic prediction characteristic generation module extracts the settlement displacement characteristic value and crack expansion direction value based on the prediction value of the abnormal characteristics of the slope protection and cracks, calculates the dynamic offset value and connectivity characteristic value, analyzes the crack length change value and deformation characteristic distribution value, and generates the result of the dynamic prediction characteristics of the cracks and slope protection; The risk warning index calculation module extracts the concrete surface stress value and settlement pattern value based on the result of the dynamic prediction characteristics of the cracks and slope protection, calculates the crack length distribution value and the spatial characteristics value of the slope protection deformation, classifies and calculates and analyzes the risk distribution characteristic value, and generates the risk warning index for the hydraulic structure.

Images