Water conservancy project risk identification method and system based on unmanned aerial vehicle AI image identification

Abstract

The invention discloses a water conservancy project risk identification method and system based on unmanned aerial vehicle AI image identification, and the method comprises the steps: obtaining an image of a water conservancy project region, and carrying out the preprocessing of the image to obtain a quality optimization image; features are extracted, regions are divided to generate feature vectors, leakage is calculated, and a leakage distribution diagram is obtained; acquiring initial three-dimensional data, generating an initial structure model, calculating a structure deformation amount, and when a preset deformation amount threshold value is exceeded, judging that the region is a high-risk region, and generating a structure deformation graph; the crack width, the leakage amount and the deformation degree are extracted, and a comprehensive risk value is calculated; preprocessing the image data, adopting a neural network algorithm to extract features and classify the features, and performing clustering analysis on a classification result to generate structured data; and uploading the comprehensive risk value and the structured data to a unified platform, performing data analysis, predicting a future risk change trend, and obtaining a risk early warning index. The method can realize accurate assessment and intelligent early warning of the risk amount of the water conservancy project.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy project risk identification, and particularly to a method and system for water conservancy project risk identification based on UAV AI image recognition. Background Art

[0002] At present, water conservancy projects are key infrastructure for ensuring important functions such as water resource allocation, flood control, and irrigation. However, water conservancy projects face various risks during operation, including structural damage, leakage, cracks, deformation, etc. These risks have a serious impact on the safe operation of the project and the surrounding environment. Therefore, timely and accurate identification and assessment of the risks of water conservancy projects are of great significance for ensuring their safe operation and reducing potential losses.

[0003] In an existing technology, this method decomposes complex decision-making problems into multiple levels and multiple factors by establishing a hierarchical structure model, which can be divided into an objective layer, a criterion layer, and a scheme layer, and determines the weights of each factor through expert scoring, and finally calculates the comprehensive risk value. This technology relies on the experience and subjective judgment of experts, lacks objective data support, is difficult to adapt to the dynamic changes of risks, and lacks the ability to process real-time data. There are problems in the prior art that the risks of water conservancy projects cannot be accurately evaluated and predicted. Summary of the Invention

[0004] The present invention provides a method and system for water conservancy project risk identification based on UAV AI image recognition to achieve precise assessment and intelligent early warning of water conservancy project risks.

[0005] In a first aspect, to solve the above technical problems, the present invention provides a method for water conservancy project risk identification, including: Obtaining high-resolution images of the water conservancy project area by using a UAV, and performing denoising, dehazing, and de-raining processing by using an image quality optimization algorithm to obtain a quality-optimized image; Extracting crack features by using an edge detection algorithm according to the quality-optimized image to obtain a crack distribution map; Performing region division according to the quality-optimized image, extracting color features and texture features, generating feature vectors, calculating the leakage volume according to the feature vectors, and obtaining a leakage distribution map; Obtaining the initial three-dimensional data of the water conservancy project structure, generating an initial structure model by using a three-dimensional reconstruction algorithm, calculating the structural deformation amount by using a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation threshold, determining that the area is a high-risk area and generating a structural deformation map; Extracting the crack width, leakage volume, and deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and calculating the comprehensive risk value by using a weighted algorithm; Use a distributed computing framework to process the quality-optimized image in parallel, extract features and classify using a neural network algorithm, and perform clustering analysis on the classification results to generate structured data; Upload the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof for data analysis, predict future risk change trends, and obtain a risk warning index.

[0006] In an alternative implementation, the obtaining of a high-resolution image of the water conservancy project area by a drone and the use of an image quality optimization algorithm for denoising, defogging, and de-raining to obtain a quality-optimized image include: Obtain a high-resolution image of the water conservancy project area by a drone, perform denoising processing using a denoising algorithm based on wavelet transform, perform defogging processing using a defogging algorithm based on dark channel prior, and perform de-raining processing using a de-raining algorithm based on convolutional neural network to obtain a second image; According to the second image, use an image quality assessment algorithm to calculate the image sharpness and contrast to obtain a quality assessment score. When the quality assessment score is lower than a preset quality threshold, re-adjust the parameters of the denoising, defogging, and de-raining algorithms to generate a quality-optimized image.

[0007] In an alternative implementation, the extracting of crack features using an edge detection algorithm based on the quality-optimized image to obtain a crack distribution map includes: Extract crack features using an edge detection algorithm based on the quality-optimized image to obtain a crack feature map; Perform region positioning and segmentation using a region growth algorithm based on the crack feature map to obtain a crack region map; Calculate crack width parameters using a morphological algorithm based on the crack region map to obtain width data. When the width data exceeds a preset width threshold, determine that the region is a high-risk region to obtain a high-risk region map; Overlay the crack region map and the high-risk region map to obtain a crack distribution map.

[0008] In an alternative implementation, the performing of region division based on the quality-optimized image and the extracting of color features and texture features to generate a feature vector, and the calculating of the leakage amount based on the feature vector to obtain a leakage distribution map include: Perform region division on the quality-optimized image using an image segmentation algorithm, extract color features and texture features from the segmented image to generate a feature vector; Calculate the leakage amount based on the feature vector. When the leakage amount is greater than a preset leakage threshold, determine that the region is a high-risk region, combine the high-risk region and the image coordinate information to generate a leakage distribution map.

[0009] In an alternative embodiment, the method for obtaining the initial three-dimensional data of the hydraulic engineering structure, generating an initial structure model using a three-dimensional reconstruction algorithm, calculating the structural deformation amount through a deformation analysis algorithm, and determining the area as a high-risk area when the structural deformation amount exceeds a preset deformation amount threshold, and generating a structural deformation map, includes: Obtain the original three-dimensional data of the key structures of the hydraulic engineering, and generate an initial structure model using a three-dimensional reconstruction algorithm; According to the initial structure model, calculate the deformation amount of each structural unit through a deformation analysis algorithm to obtain deformation distribution data; According to the deformation distribution data, when the deformation amount exceeds the preset deformation amount threshold, determine the area as a high-risk area; use a spatial interpolation algorithm to optimize the boundary of the high-risk area and generate an accurate risk area distribution map; Overlay the risk area distribution map with the initial structure model to generate a three-dimensional structural deformation map with risk marks; According to the three-dimensional structural deformation map, extract the structural parameters of the high-risk area and calculate the structural safety factor; when the structural safety factor is lower than the preset safety standard, generate a final structural deformation map.

[0010] In an alternative embodiment, the method for extracting the crack width, the leakage amount, and the deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and calculating the comprehensive risk value using a weighted algorithm, includes: According to the crack distribution map, the leakage distribution map, and the structural deformation map, extract the crack width, the leakage amount, and the deformation degree data; input the crack width, the leakage amount, and the deformation degree data into a pre-established risk assessment system, and calculate a first risk value using a weighted algorithm; According to the first risk value, determine the risk level. When the first risk value is greater than the preset risk threshold, determine it as a high-risk level, otherwise as a medium-low risk level; perform a clustering analysis on the crack width, the leakage amount, and the deformation degree data through a machine learning algorithm, identify the high-risk areas, and use a regression algorithm to predict the future change trends of the crack width, the leakage amount, and the deformation degree to generate a trend analysis result; Perform a weighted average calculation on the risk level, the high-risk areas, and the trend analysis result to obtain the comprehensive risk value.

[0011] In an alternative embodiment, the method for performing parallel processing on the quality-optimized image using a distributed computing framework, extracting features and classifying using a neural network algorithm, and performing a clustering analysis on the classification results to generate structured data, includes: Use a distributed computing framework to perform parallel processing on the quality-optimized image, divide the image data into multiple subsets, and allocate each subset to a different computing node; on each computing node, perform denoising, grayscaling, and normalization operations on the image data to obtain the processed image; According to the processed image, use a convolutional neural network algorithm to extract edge, texture, and shape features, generate a feature vector; input the extracted feature vector into a support vector machine algorithm for classification to determine the category to which the image data belongs and obtain a classification result; According to the classification result, use a clustering algorithm to perform clustering analysis on the image data, group the image data with similar features into the same category, and generate a clustering result; Match the clustering result with a preset category label to determine the category label of each cluster and generate structured data.

[0012] In a second aspect, the present invention provides a water conservancy project risk identification system, including: An image processing module, configured to obtain a high-resolution image of the water conservancy project area through a drone, and perform denoising, defogging, and de-raining processing using an image quality optimization algorithm to obtain a quality-optimized image; A crack extraction module, configured to extract crack features using an edge detection algorithm according to the quality-optimized image to obtain a crack distribution map; A leakage calculation module, configured to perform region division on the quality-optimized image, extract color features and texture features, generate a feature vector, and calculate the leakage amount according to the feature vector to obtain a leakage distribution map; A deformation analysis module, configured to obtain the initial three-dimensional data of the water conservancy project structure, generate an initial structure model using a three-dimensional reconstruction algorithm, calculate the structural deformation amount through a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation threshold, determine that the area is a high-risk area and generate a structural deformation map; A risk calculation module, configured to extract the crack width, leakage amount, and deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and calculate a comprehensive risk value using a weighted algorithm; A structuring module, configured to perform parallel processing on the quality-optimized image using a distributed computing framework, extract features and classify them using a neural network algorithm, and perform clustering analysis on the classification result to generate structured data; A risk warning module, configured to upload the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof, perform data analysis, predict the future risk change trend, and obtain a risk warning index.

[0013] In a third aspect, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the method for identifying risks in a water conservancy project described in any one of the above is implemented.

[0014] In a fourth aspect, the present invention also provides a computer-readable storage medium, which includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the method for identifying risks in a water conservancy project described in any one of the above.

[0015] Compared with the prior art, the present invention has the following beneficial effects: The method includes obtaining a high-resolution image of the water conservancy project area by using a drone, performing denoising, defogging, and de-raining processing by using an image quality optimization algorithm to obtain a quality-optimized image; extracting crack features by using an edge detection algorithm based on the quality-optimized image to obtain a crack distribution map; performing region division based on the quality-optimized image, extracting color features and texture features, generating a feature vector, calculating the leakage amount according to the feature vector to obtain a leakage distribution map; obtaining initial three-dimensional data of the water conservancy project structure, generating an initial structure model by using a three-dimensional reconstruction algorithm, calculating the structural deformation amount by using a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation amount threshold, determining that the area is a high-risk area and generating a structural deformation map; extracting the crack width, leakage amount, and deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and calculating a comprehensive risk value by using a weighted algorithm; performing parallel processing on the quality-optimized image by using a distributed computing framework, extracting features and classifying them by using a neural network algorithm, performing clustering analysis on the classification results to generate structured data; uploading the comprehensive risk value and the structured data to a unified platform by using data sharing and zero-knowledge proof for data analysis, predicting the future risk change trend, and obtaining a risk warning index.

[0016] The present invention obtains a high-resolution image by using a drone, processes it by using an image quality optimization algorithm, and extracts key features such as cracks, leakage, and structural deformation. By using feature analysis and image segmentation techniques, the risk area is accurately located and a distribution map is generated. The three-dimensional reconstruction technology is used to analyze the degree of structural deformation. Multiple risk factors are comprehensively evaluated to obtain the risk level. The present invention also uses a distributed computing framework to process massive data and uses zero-knowledge proof technology to securely share the evaluation results. By analyzing historical data, the future risk trend is predicted and a warning report is generated. The accurate assessment, efficient processing, and intelligent warning of water conservancy project risks are realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1It is a schematic flowchart of a water conservancy project risk identification method based on UAV AI image recognition provided by the first embodiment of the present invention; Figure 2 It is a schematic structural diagram of a water conservancy project risk identification method system based on UAV AI image recognition provided by the second embodiment of the present invention. Specific embodiments

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0019] Refer to Figure 1 , the first embodiment of the present invention provides a water conservancy project risk identification method, including the following steps: S11, obtain a high-resolution image of the water conservancy project area through a UAV, and perform denoising, dehazing, and de-raining processing using an image quality optimization algorithm to obtain a quality-optimized image; S12, according to the quality-optimized image, extract crack features using an edge detection algorithm to obtain a crack distribution map; S13, according to the quality-optimized image, perform region division and extract color features and texture features to generate feature vectors, calculate the leakage volume according to the feature vectors, and obtain a leakage distribution map; S14, obtain the initial three-dimensional data of the water conservancy project structure, generate an initial structure model using a three-dimensional reconstruction algorithm, calculate the structural deformation amount through a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation threshold, determine that the area is a high-risk area and generate a structural deformation map; S15, according to the crack distribution map, the leakage distribution map, and the structural deformation map, extract the crack width, leakage volume, and deformation degree, and calculate the comprehensive risk value using a weighted algorithm; S16, perform parallel processing on the quality-optimized image using a distributed computing framework, extract features and classify using a neural network algorithm, and perform clustering analysis on the classification results to generate structured data; S17, upload the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof, perform data analysis, predict the future risk change trend, and obtain a risk warning index.

[0020] In step S11, a high-resolution image of the water conservancy project area is obtained through a UAV, and denoising, dehazing, and de-raining processing are performed using an image quality optimization algorithm to obtain a quality-optimized image.

[0021] In one implementation, a high-resolution image of a water conservancy project area is obtained by a drone, and denoising processing is performed using a denoising algorithm based on wavelet transform, defogging processing is performed using a defogging algorithm based on dark channel prior, and de-raining processing is performed using a de-raining algorithm based on convolutional neural network to obtain a second image; According to the second image, an image quality assessment algorithm is used to calculate the image sharpness and contrast to obtain a quality assessment score. When the quality assessment score is lower than a preset quality threshold, the parameters of the denoising, defogging, and de-raining algorithms are readjusted to generate a quality-optimized image.

[0022] It should be noted that wavelet transform denoising is a signal processing technology based on wavelet transform. By decomposing a signal into sub-signals of different frequencies and using a threshold method to perform denoising processing on these sub-signals. Its basic principle is to use wavelet transform to decompose a noisy signal into different scales, then process the wavelet coefficients, and finally reconstruct the signal through inverse wavelet transform to remove noise. The dark channel prior defogging algorithm is a defogging method based on image brightness information. Its core idea is to use the atmospheric scattering model and dark channel prior knowledge to estimate the atmospheric light value and transmittance, so as to achieve defogging. The image quality assessment algorithm is used to measure the quality of an image. The objective assessment method evaluates the image quality by calculating certain eigenvalue of the image. The image quality is evaluated by comparing the pixel value differences between the original image and the processed image. The higher the PSNR value, the better the image quality. The similarity of the image can also be evaluated based on the brightness, contrast, and structural information of the image. The closer the SSIM value is to 1, the better the image quality.

[0023] Exemplarily, in a reservoir dam construction project, a six-rotor drone equipped with a 4K camera flew at an altitude of 500 meters at a speed of 30 meters per second, obtaining an orthophoto covering an area of 20 square kilometers with a resolution of 2 centimeters. The sym4 wavelet basis was selected for denoising processing, and the image was decomposed into 3 layers. The soft threshold method was used for coefficient shrinkage in the high-frequency subbands, and finally wavelet reconstruction was performed to obtain the denoised image. This method not only preserves the edge details of the image but also effectively suppresses noise. Under foggy conditions, the image will be affected by fog. The dehazing algorithm based on dark channel prior can effectively restore the scene details obscured by fog by estimating the atmospheric light and transmittance. By calculating the dark channel map, the estimated atmospheric light value is 220, and the guided filter is used to optimize the transmittance map, and finally a clear dehazed image is obtained, making the originally blurred river channel contour and surrounding vegetation clearly visible. In the images taken on rainy days, raindrops will seriously affect the image quality. The rain removal algorithm based on convolutional neural network can automatically learn the raindrop characteristics and remove them. For example, a multi-scale residual network is adopted, which contains 20 convolutional layers, uses 64 3x3 convolutional kernels, and is trained for 100 rounds on 10,000 rainy images, which can effectively remove raindrops of different sizes and densities and restore the image details. Image quality assessment is a key step to ensure the processing effect. Metrics such as structural similarity and peak signal-to-noise ratio can be used. Set the SSIM threshold to 0.85 and the PSNR threshold to 30 dB. If the evaluation result is lower than the threshold, the aforementioned algorithm parameters need to be adjusted. Increase the wavelet decomposition layer number of the denoising algorithm, increase the minimum filtering window size of the dehazing algorithm, or increase the training rounds of the rain removal network to obtain better processing effects. Through this series of processes, the finally obtained high-quality images can be used for precise analysis of water conservancy projects.

[0024] In step S12, according to the quality-optimized image, an edge detection algorithm is used to extract crack features, and a crack distribution map is obtained.

[0025] In one implementation, according to the quality-optimized image, an edge detection algorithm is used to extract crack features, and a crack feature map is obtained; According to the crack feature map, a region growth algorithm is used for region positioning and segmentation, and a crack region map is obtained; According to the crack region map, a morphological algorithm is used to calculate the crack width parameter, and width data is obtained. When the width data exceeds the preset width threshold, it is determined that the region is a high-risk region, and a high-risk region map is obtained; The crack region map and the high-risk region map are superimposed to obtain a crack distribution map.

[0026] It should be noted that edge detection is an important technique in image processing, which is used to detect regions with significant changes in grayscale values in an image and can be used to extract the contours or features of the image. Edge detection algorithms include the Sobel algorithm, Canny algorithm, Prewitt algorithm, and LoG algorithm. The region growing algorithm is an image segmentation method based on pixel similarity. Its basic idea is to start from one or more seed points and gradually merge adjacent pixels with similar features into the same region. The advantages of the region growing algorithm are simplicity, intuitiveness, easy implementation, and a certain degree of robustness to noise.

[0027] Exemplarily, after obtaining the image with optimized quality, the Canny algorithm is used. It effectively extracts the edge information in the image through steps such as Gaussian filtering, gradient calculation, non-maximum suppression, and double-threshold detection, forming a crack feature map. This method can accurately capture the contours of the cracks, laying a foundation for subsequent analysis. Next, the region growing algorithm is used to achieve precise positioning and segmentation of the crack region. The region growing algorithm starts from a seed point and gradually merges similar neighboring pixels, finally forming a complete crack region. For example, a point with a lower grayscale value in the crack feature map can be selected as the seed point, and then according to a preset grayscale threshold, the surrounding qualified pixels are continuously incorporated into the region until no further expansion is possible. In this way, an accurate crack region map can be obtained, which is helpful for subsequent width calculation and risk assessment. After extracting the crack edge information from the crack region map, a morphological algorithm is used to calculate the crack width parameter. Common morphological operations include erosion and dilation. By performing an erosion operation on the crack region, the skeleton line of the crack can be obtained. Then, through a dilation operation, the original width of the crack can be restored. The pixel difference between the two operations is the crack width. For example, assume that the width of a certain crack is 1 pixel after erosion and 5 pixels after dilation, then the width of the crack at that location is 4 pixels. This method can quickly and accurately obtain the crack width data. According to the calculated width data and combined with a preset threshold, high-risk regions can be judged. For example, assume that the preset threshold is 5 mm, then the crack regions with a width exceeding 5 mm are marked as high-risk regions. These regions can be marked in red in the image to form a high-risk region map. This visualization method intuitively shows the regions that need to be focused on, helping engineers quickly identify potential dangerous points. The crack region map and the high-risk region map are superimposed to generate a comprehensive crack distribution map.

[0028] In step S13, according to the quality-optimized image, region division is performed and color features and texture features are extracted to generate a feature vector. The leakage amount is calculated based on the feature vector to obtain a leakage distribution map.

[0029] In one implementation, an image segmentation algorithm is used to perform region division on the quality-optimized image, and color features and texture features are extracted from the segmented image to generate a feature vector; According to the eigenvector, calculate the leakage volume. When the leakage volume is greater than the preset leakage threshold, determine that the area is a high-risk area, and combine the high-risk area with the image coordinate information to generate a leakage distribution map.

[0030] It should be noted that leakage image analysis is a key link in the maintenance of underground building structures. Image segmentation algorithms such as the watershed method or region growing method can effectively divide the image area and separate the leakage area from the non-leakage area. Color feature extraction can use the HSV color space to analyze the hue, saturation, and brightness values of the leakage area. For example, the leakage area may appear darker gray or brown with lower saturation. Texture features can be extracted through the gray-level co-occurrence matrix or wavelet transform to reflect the roughness and irregularity of the leakage area. The generation of the eigenvector combines color and texture information to form a multi-dimensional data representation. For example, a 5-dimensional eigenvector can contain 3 color components and 2 texture descriptors. Using machine learning algorithms such as support vector machines or random forests, the leakage area can be accurately identified based on the eigenvector. The leakage volume calculation considers the area and intensity of the leakage area. Suppose the area of a certain area is 10 square meters and the average leakage intensity is 2 (graded from 0 to 5), then the leakage volume can be estimated to be 20 units. If the preset threshold is 15 units, this area will be marked as high-risk. Combine the high-risk area with the image coordinate information to generate a leakage distribution map.

[0031] In step S14, obtain the initial three-dimensional data of the water conservancy project structure, use a three-dimensional reconstruction algorithm to generate an initial structure model, calculate the structural deformation amount through a deformation analysis algorithm. When the structural deformation amount exceeds the preset deformation threshold, determine that the area is a high-risk area and generate a structural deformation map.

[0032] In one implementation, obtain the original three-dimensional data of the key structures of the water conservancy project and use a three-dimensional reconstruction algorithm to generate an initial structure model.

[0033] Obtain the original three-dimensional data of the key structures of the water conservancy project and use a three-dimensional reconstruction algorithm to generate an initial structure model; According to the initial structure model, calculate the deformation amount of each structural unit through a deformation analysis algorithm to obtain deformation distribution data; According to the deformation distribution data, when the deformation amount exceeds the preset deformation threshold, determine that the area is a high-risk area; use a spatial interpolation algorithm to optimize the boundary of the high-risk area and generate an accurate risk area distribution map; Overlay the risk area distribution map with the initial structure model to generate a three-dimensional structural deformation map containing risk marks; According to the three-dimensional structural deformation map, extract the structural parameters of the high-risk area and calculate the structural safety factor; when the structural safety factor is lower than the preset safety standard, generate a final structural deformation map.

[0034] It should be noted that the 3D reconstruction of the key structures of water conservancy projects is an important link to ensure project safety. Taking a large reservoir dam as an example, the original 3D point cloud data of the dam needs to be obtained first. Through laser scanning technology, high-precision point cloud data can be obtained, including the geometric information of the dam surface. Using these data, a 3D reconstruction algorithm based on octree is adopted to convert the discrete point cloud into a continuous surface model and generate an initial structure model. For the initial structure model, the finite element method is used for deformation analysis. The dam model is divided into several elements, and factors such as water pressure and temperature change are considered to calculate the stress and deformation of each element. For example, the maximum deformation of a certain element is 5mm, exceeding the preset threshold of 3mm, so this area is marked as a high-risk area. In order to more accurately locate the high-risk area, the Kriging interpolation method can be used to optimize the boundary of the risk area. This method takes into account the spatial correlation and can make reasonable interpolation estimates between known points. Through interpolation optimization, a more continuous and smooth risk area distribution map can be obtained, which is beneficial to subsequent risk assessment and processing. The optimized risk area distribution map is superimposed on the initial structure model to obtain a 3D structure deformation map. Based on the 3D structure deformation map, the structural parameters of the high-risk area, such as stress distribution and crack width, can be extracted. Using these parameters, the structural safety factor can be calculated. Suppose the safety factor of a certain high-risk area is 1.2, which is lower than the preset standard of 1.5, indicating that there are potential safety hazards in this area. Finally, the final structure deformation map is generated.

[0035] In step S15, according to the crack distribution map, the leakage distribution map and the structure deformation map, the crack width, the leakage amount and the deformation degree are extracted, and a weighted algorithm is used to calculate the comprehensive risk value.

[0036] In one implementation, according to the crack distribution map, the leakage distribution map and the structure deformation map, the crack width, the leakage amount and the deformation degree data are extracted; the crack width, the leakage amount and the deformation degree data are input into a pre-established risk assessment system, and a weighted algorithm is used to calculate the first risk value; According to the first risk value, the risk level is judged. When the first risk value is greater than the preset risk threshold, it is determined as the high-risk level, otherwise it is the medium-low risk level; the crack width, the leakage amount and the deformation degree data are clustered and analyzed through a machine learning algorithm to identify the high-risk area and a regression algorithm is used to predict the future change trend of the crack width, the leakage amount and the deformation degree, and a trend analysis result is generated; The risk level, the high-risk area and the trend analysis result are weighted and averaged to calculate the comprehensive risk value.

[0037] It should be noted that the crack distribution map, leakage distribution map, and structural deformation map are obtained, and these image data reflect the actual condition of the engineering structure. For example, the crack distribution map of a certain dam shows that there is a crack with a length of 5 meters and a width of 2 millimeters in the middle of the dam body; the leakage distribution map indicates that there is a leakage of 100 liters per hour at the dam foundation; the structural deformation map reveals that the settlement deformation at the center of the dam crest reaches 10 millimeters. Key data such as crack width, leakage volume, and deformation degree are extracted from these images. These data are important indicators for risk assessment. The extracted data is input into a pre-established risk assessment system, and a weighted algorithm is used to calculate the comprehensive risk value. Different weights are assigned to different factors in this system. For example, the weight of crack width is 0.4, the weight of leakage volume is 0.3, and the weight of deformation degree is 0.3. If the comprehensive risk value calculated by weighted calculation exceeds the preset threshold of 0.7, it is determined to be a high-risk level; otherwise, it is a medium-low risk level. Next, machine learning algorithms are used to perform clustering analysis on the data to identify high-risk areas. For example, the K-means clustering algorithm is used to divide the data points into several clusters, and the clusters with higher risk indicators are identified as high-risk areas. This method can quickly locate the areas that need key attention. At the same time, regression algorithms are used to predict future change trends. For example, for the crack width, a linear regression model can be used to predict its change in the next year. If the prediction result shows that the crack width will increase to 3 millimeters, important early warning information will be provided. Finally, the risk level, high-risk areas, and trend analysis results are weighted and averaged to calculate the comprehensive risk value.

[0038] In step S16, a distributed computing framework is used to perform parallel processing on the quality-optimized image, a neural network algorithm is used to extract features and classify, and clustering analysis is performed on the classification results to generate structured data.

[0039] In one implementation, a distributed computing framework is used to perform parallel processing on the quality-optimized image, the image data is segmented into multiple subsets, and each subset is assigned to a different computing node; on each computing node, denoising, grayscale conversion, and normalization operations are performed on the image data to obtain the processed image; Based on the processed image, a convolutional neural network algorithm is used to extract edge, texture, and shape features to generate feature vectors; the extracted feature vectors are input into a support vector machine algorithm for classification to determine the category to which the image data belongs and obtain the classification results; Based on the classification results, a clustering algorithm is used to perform clustering analysis on the image data, and the image data with similar features is grouped into the same category to generate clustering results; The clustering results are matched with the preset category labels to determine the category labels of each clustering cluster and generate structured data.

[0040] It should be noted that the distributed computing framework has significant advantages in processing massive image data. Taking the example of a drone aerial survey of a forest fire, a large number of acquired image data can be segmented into multiple subsets and distributed to different computing nodes for parallel processing. Each node is responsible for processing images in a specific area or time period, improving the overall processing efficiency. Image preprocessing is a key step to improve the accuracy of subsequent analysis. For forest fire images, denoising can eliminate interference factors such as smoke, grayscale conversion helps to highlight the contrast between the fire and the surrounding environment, and normalization enables images taken under different lighting conditions to be comparable. These operations lay the foundation for feature extraction. Convolutional neural networks perform well in extracting image features. For forest fire images, the neural network will focus on extracting the edge features of the fire, the texture features of the smoke, and the shape change features after the trees burn. The feature vectors composed of these features can effectively describe various aspects of the fire. The support vector machine algorithm uses the extracted feature vectors for classification, and can divide the images into categories such as fire areas, potential fire hazard areas, and safe areas. Clustering analysis further refines the classification results. For example, the fire area can be clustered into sub-categories such as initial fire, full combustion, and ember stage. This detailed division helps to formulate more accurate fire extinguishing strategies. Matching the clustering results with preset labels generates structured data.

[0041] In summary, the present invention discloses a method for identifying risks in water conservancy projects based on drone AI image processing, including obtaining high-resolution images of the water conservancy project area through a drone, performing denoising, defogging, and de-raining processing using an image quality optimization algorithm to obtain a quality-optimized image; according to the quality-optimized image, using an edge detection algorithm to extract crack features to obtain a crack distribution map; according to the quality-optimized image, performing area division and extracting color features and texture features to generate a feature vector, calculating the leakage volume according to the feature vector to obtain a leakage distribution map; obtaining the initial three-dimensional data of the water conservancy project structure, using a three-dimensional reconstruction algorithm to generate an initial structure model, calculating the structural deformation amount through a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation threshold, determining that the area is a high-risk area and generating a structural deformation map; according to the crack distribution map, the leakage distribution map, and the structural deformation map, extracting the crack width, leakage volume, and deformation degree, and using a weighted algorithm to calculate the comprehensive risk value; using a distributed computing framework to perform parallel processing on the quality-optimized image, using a neural network algorithm to extract features and classify, performing clustering analysis on the classification results to generate structured data; uploading the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof for data analysis to predict the future risk change trend and obtain a risk warning index.

[0042] The present invention obtains high-resolution images through drones, processes them using image quality optimization algorithms, and extracts key features such as cracks, leaks, and structural deformations. Using feature analysis and image segmentation techniques, it accurately locates risk areas and generates distribution maps. Three-dimensional reconstruction technology is used to analyze the degree of structural deformation. Multiple risk factors are comprehensively evaluated to obtain a risk level. The present invention also uses a distributed computing framework to process massive data and uses zero-knowledge proof technology to securely share evaluation results. Through historical data analysis, it predicts future risk trends and generates warning reports. It realizes the accurate assessment, efficient processing, and intelligent warning of water conservancy project risks.

[0043] Referring to Figure 2 , the second embodiment of the present invention provides a water conservancy project risk identification system, including: An image processing module, configured to obtain high-resolution images of the water conservancy project area through drones, and perform denoising, dehazing, and de-raining processing using an image quality optimization algorithm to obtain a quality-optimized image; A crack extraction module, configured to extract crack features using an edge detection algorithm based on the quality-optimized image to obtain a crack distribution map; A leakage calculation module, configured to perform area division based on the quality-optimized image, extract color features and texture features, generate feature vectors, calculate the leakage volume according to the feature vectors, and obtain a leakage distribution map; A deformation analysis module, configured to obtain the initial three-dimensional data of the water conservancy project structure, generate an initial structure model using a three-dimensional reconstruction algorithm, calculate the structural deformation amount through a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation amount threshold, determine that the area is a high-risk area and generate a structural deformation map; A risk calculation module, configured to extract the crack width, leakage volume, and deformation degree based on the crack distribution map, the leakage distribution map, and the structural deformation map, and calculate a comprehensive risk value using a weighted algorithm; A structuring module, configured to perform parallel processing on the quality-optimized image using a distributed computing framework, extract features and classify them using a neural network algorithm, and perform clustering analysis on the classification results to generate structured data; A risk warning module, configured to upload the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof, perform data analysis, predict future risk change trends, and obtain a risk warning index.

[0044] It should be noted that a water conservancy project risk identification device provided in an embodiment of the present invention is used to execute all the process steps of a water conservancy project risk identification method in the above embodiment, and the working principles and beneficial effects of the two correspond one by one, so they will not be elaborated here.

[0045] An embodiment of the present invention also provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a water conservancy project risk identification program. When the processor executes the computer program, the steps in the above embodiments of various water conservancy project risk identification methods are implemented, such as Figure 1 the step S11 shown. Alternatively, when the processor executes the computer program, the functions of each module / unit in the above device embodiments are implemented, such as the water conservancy project risk identification module.

[0046] Exemplarily, the computer program may be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing specific functions, and these instruction segments are used to describe the execution process of the computer program in the electronic device.

[0047] The electronic device may be a computing device such as a desktop computer, a notebook, a palm computer, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above components are only examples of the electronic device and do not constitute a limitation on the electronic device. It may include more or fewer components than the above, or combine certain components, or different components. For example, the electronic device may further include input / output devices, network access devices, a bus, etc.

[0048] The so-called processor may be a central processing unit (CPU), or may also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the electronic device and connects various parts of the entire electronic device through various interfaces and lines.

[0049] The memory can be used to store the computer program and / or modules. By running or executing the computer program and / or modules stored in the memory, and by invoking the data stored in the memory, the processor can implement various functions of the electronic device. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store the operating system, application programs required for at least one function (such as the sound playback function, the image playback function, etc.); the data storage area can store the data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory can include high-speed random access memory, and can also include non-volatile memory, such as hard disks, memory, plug-in hard disks, smart media cards (SMC), secure digital (SD) cards, flash cards, at least one magnetic disk storage device, flash device, or other volatile solid-state storage devices.

[0050] Among them, if the modules / units integrated in the electronic device are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-mentioned embodiment methods of the present invention, it can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0051] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0052] The specific embodiments described above further elaborate on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. In particular, for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A risk identification method for water conservancy projects, characterized in that, Executed by a computer, including: Obtaining high-resolution images of the water conservancy project area by an unmanned aerial vehicle, and performing denoising, defogging, and de-raining processing using an image quality optimization algorithm to obtain quality-optimized images; Extracting crack features using an edge detection algorithm based on the quality-optimized images to obtain a crack distribution map; Performing region division on the quality-optimized images, extracting color features and texture features, generating feature vectors, calculating the leakage volume based on the feature vectors, and obtaining a leakage distribution map; Obtaining the initial three-dimensional data of the water conservancy project structure, generating an initial structure model using a three-dimensional reconstruction algorithm, calculating the structural deformation amount through a deformation analysis algorithm. When the structural deformation amount exceeds a preset deformation threshold, it is determined that the area is a high-risk area, and a structural deformation map is generated; Extracting the crack width, leakage volume, and deformation degree based on the crack distribution map, the leakage distribution map, and the structural deformation map, and calculating the comprehensive risk value using a weighted algorithm; Performing parallel processing on the quality-optimized images using a distributed computing framework, extracting features and classifying them using a neural network algorithm, and performing clustering analysis on the classification results to generate structured data; Uploading the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof, performing data analysis, predicting the future risk change trend, and obtaining a risk warning index.

2. The risk identification method for water conservancy projects according to claim 1, wherein The process of obtaining high-resolution images of the water conservancy project area by an unmanned aerial vehicle, and performing denoising, defogging, and de-raining processing using an image quality optimization algorithm to obtain quality-optimized images includes: Obtaining high-resolution images of the water conservancy project area by an unmanned aerial vehicle, performing denoising processing using a denoising algorithm based on wavelet transform, performing defogging processing using a defogging algorithm based on dark channel prior, and performing de-raining processing using a de-raining algorithm based on a convolutional neural network to obtain a second image; Based on the second image, using an image quality assessment algorithm to calculate the image sharpness and contrast, obtaining a quality assessment score. When the quality assessment score is lower than a preset quality threshold, the parameters of the denoising, defogging, and de-raining algorithms are readjusted to generate quality-optimized images.

3. The risk identification method for water conservancy projects according to claim 1, wherein The process of extracting crack features using an edge detection algorithm based on the quality-optimized images to obtain a crack distribution map includes: Extracting crack features using an edge detection algorithm based on the quality-optimized images to obtain a crack feature map; Performing region positioning and segmentation using a region growth algorithm based on the crack feature map to obtain a crack region map; Calculating the crack width parameter using a morphological algorithm based on the crack region map to obtain width data. When the width data exceeds a preset width threshold, it is determined that the area is a high-risk area, and a high-risk area map is obtained; Overlaying the crack region map and the high-risk area map to obtain a crack distribution map.

4. The risk identification method for water conservancy projects according to claim 1, wherein, Performing region division on the quality-optimized images and extracting color features and texture features to generate a feature vector, The process of calculating the leakage volume based on the feature vector to obtain a leakage distribution map includes: Performing region division on the quality-optimized images using an image segmentation algorithm, extracting color features and texture features from the segmented images, and generating feature vectors; According to the eigenvector, calculate the leakage volume. When the leakage volume is greater than the preset leakage threshold, determine that the area is a high-risk area, and combine the high-risk area with the image coordinate information to generate a leakage distribution map.

5. The risk identification method for water conservancy projects according to claim 1, characterized in that The acquisition of the initial three-dimensional data of the water conservancy project structure, the use of a three-dimensional reconstruction algorithm to generate an initial structure model, and the calculation of the structural deformation amount through a deformation analysis algorithm. When the structural deformation amount exceeds the preset deformation threshold, determine that the area is a high-risk area and generate a structural deformation map, including: Obtain the original three-dimensional data of the key structures of the water conservancy project and generate an initial structure model using a three-dimensional reconstruction algorithm; According to the initial structure model, calculate the deformation amount of each structural unit through a deformation analysis algorithm to obtain deformation distribution data; According to the deformation distribution data, when the deformation amount exceeds the preset deformation threshold, determine that the area is a high-risk area; use a spatial interpolation algorithm to optimize the boundary of the high-risk area to generate an accurate risk area distribution map; Overlay the risk area distribution map with the initial structure model to generate a three-dimensional structural deformation map containing risk marks; According to the three-dimensional structural deformation map, extract the structural parameters of the high-risk area and calculate the structural safety factor; when the structural safety factor is lower than the preset safety standard, generate a final structural deformation map.

6. The risk identification method for water conservancy projects according to claim 2, characterized in that The extraction of the crack width, leakage volume, and deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and the calculation of the comprehensive risk value using a weighted algorithm, including: Extract the crack width, leakage volume, and deformation degree data according to the crack distribution map, the leakage distribution map, and the structural deformation map; input the crack width, leakage volume, and deformation degree data into a pre-established risk assessment system and calculate the first risk value using a weighted algorithm; According to the first risk value, judge the risk level. When the first risk value is greater than the preset risk threshold, determine it as a high-risk level, otherwise it is a medium-low risk level; perform clustering analysis on the crack width, leakage volume, and deformation degree data through a machine learning algorithm, identify the high-risk area, and use a regression algorithm to predict the future change trend of the crack width, leakage volume, and deformation degree to generate a trend analysis result; Perform a weighted average calculation on the risk level, high-risk area, and trend analysis result to obtain the comprehensive risk value.

7. The risk identification method for hydraulic engineering according to claim 1, wherein The use of a distributed computing framework to perform parallel processing on the quality-optimized image, the use of a neural network algorithm to extract features and classify, and the performance of clustering analysis on the classification results to generate structured data, including: Use a distributed computing framework to perform parallel processing on the quality-optimized image, divide the image data into multiple subsets, and assign each subset to a different computing node; on each computing node, perform denoising, grayscale conversion, and normalization operations on the image data to obtain the processed image; According to the processed image, use a convolutional neural network algorithm to extract edge, texture, and shape features to generate a feature vector; input the extracted feature vector into a support vector machine algorithm for classification to judge the category to which the image data belongs and obtain the classification result; According to the classification results, a clustering algorithm is used to perform clustering analysis on the image data, and the image data with similar features are grouped into the same category to generate a clustering result; The clustering result is matched with a preset category label to determine the category label of each clustering cluster and generate structured data.

8. A water conservancy project risk identification system, characterized in that, It includes: An image processing module, which is used to obtain high-resolution images of the water conservancy project area through a drone, and perform denoising, defogging, and de-raining processing using an image quality optimization algorithm to obtain a quality-optimized image; A crack extraction module, which is used to extract crack features according to the quality-optimized image using an edge detection algorithm to obtain a crack distribution map; A leakage calculation module, which is used to perform regional division on the quality-optimized image, extract color features and texture features, generate feature vectors, calculate the leakage volume according to the feature vectors, and obtain a leakage distribution map; A deformation analysis module, which is used to obtain the initial three-dimensional data of the water conservancy project structure, generate an initial structure model using a three-dimensional reconstruction algorithm, calculate the structural deformation amount through a deformation analysis algorithm, and when the structural deformation amount exceeds a preset deformation threshold, it is determined that the area is a high-risk area and a structural deformation map is generated; A risk calculation module, which is used to extract the crack width, leakage volume, and deformation degree according to the crack distribution map, the leakage distribution map, and the structural deformation map, and calculate the comprehensive risk value using a weighted algorithm; A structuring module, which is used to perform parallel processing on the quality-optimized image using a distributed computing framework, extract features and classify them using a neural network algorithm, perform clustering analysis on the classification results, and generate structured data; A risk warning module, which is used to upload the comprehensive risk value and the structured data to a unified platform using data sharing and zero-knowledge proof, perform data analysis, predict the future risk change trend, and obtain a risk warning index.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the water conservancy project risk identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the water conservancy project risk identification method according to any one of claims 1 to 7.

Images