Dam deformation detection method and system based on radar

Abstract

The invention relates to the technical field of water conservancy construction, in particular to a radar-based dam deformation detection method and system. Firstly, dam radar image data and environment radar image data are acquired; then, the preprocessed dam radar image data and the preprocessed environment radar image data are recognized through a feature recognition model, and a first dam deformation detection value and a second dam deformation detection value are obtained through calculation according to dam features and environment features output by the model; thirdly, comprehensively evaluating the first dam deformation detection value and the second dam deformation detection value to obtain a comprehensive dam deformation detection value; and finally, comparing the comprehensive dam deformation detection value with a detection threshold value, and if the comprehensive dam deformation detection value exceeds the threshold value, informing related personnel to process. Otherwise, inputting the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model for parameter updating, and generating dam deformation early warning information according to a data result of the dam deformation prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy construction, and in particular to a radar-based dam deformation detection method and system. Background Art

[0002] As a key hub of water conservancy projects, dams play a significant role in regulating the spatial and temporal distribution of water resources. By storing and releasing water, dams not only effectively control flood flows but also ensure a stable water supply during droughts.

[0003] As an important hydraulic structure, the dam will be affected by external loads and changes in the surrounding geological structure during its long-term operation, which will cause the dam to deform to a certain extent during operation. In order to ensure the safety and stability of the dam, it is crucial to conduct regular deformation detection of the dam, so as to timely discover potential safety hazards and ensure that the dam continues to play its role in regulating water resources throughout its service life.

[0004] Currently, existing technologies for radar-based dam deformation detection still have shortcomings. Firstly, they fail to consider the impact of external environmental changes on dam deformation detection. Secondly, they fail to analyze the relationship between dam deformation and the environment in historical data to predict the changing trend of dam deformation, nor do they provide corresponding dam deformation early warning information. This reduces the reliability and accuracy of dam deformation detection.

[0005] Therefore, a radar-based dam deformation detection method and system are proposed. Summary of the Invention

[0006] The present invention aims to provide a radar-based dam deformation detection method and system. First, dam radar image data and environmental radar image data are acquired. A feature recognition model is then used to identify the pre-processed dam radar image data and the pre-processed environmental radar image data. A first dam deformation detection value and a second dam deformation detection value are calculated based on the dam features and environmental features output by the model. The first and second dam deformation detection values are then comprehensively evaluated to obtain a composite dam deformation detection value. Finally, the composite dam deformation detection value is compared with a detection threshold. If the threshold is exceeded, relevant personnel are notified to take appropriate action. Otherwise, the dam deformation prediction weight and real-time detection data are input into the dam deformation prediction model for parameter update. Dam deformation warning information is then generated based on the data output from the dam deformation prediction model.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A radar-based dam deformation detection method comprising:

[0009] Acquire dam radar image data and environmental radar image data collected by synthetic aperture radar;

[0010] Preprocessing the dam radar image data and the environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data;

[0011] Inputting the pre-processed dam radar image data into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; evaluating the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain a first dam deformation detection value;

[0012] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam's surrounding environment; evaluating the soil features, vegetation features, and water level features to obtain a second dam deformation detection value;

[0013] Performing a weighted evaluation on the first dam deformation detection value and the second dam deformation detection value to obtain a comprehensive dam deformation detection value;

[0014] The comprehensive dam deformation detection value is compared with the detection threshold. If it exceeds the detection threshold, relevant personnel are informed to handle it; otherwise, the dam deformation prediction weight and real-time detection data are input into the dam deformation prediction model for parameter update, and dam deformation early warning information is generated according to the data results of the dam deformation prediction model.

[0015] Furthermore, the dam radar image data and the environment radar image data are preprocessed to obtain preprocessed dam radar image data and preprocessed environment radar image data:

[0016] Acquire dam radar image data and environmental radar image data;

[0017] Further, using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data;

[0018] Further, histogram equalization is used to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data;

[0019] Furthermore, a normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

[0020] Furthermore, the pre-processed dam radar image data is input into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; and the displacement characteristics, permeability characteristics, and defect characteristics are evaluated to obtain a first dam deformation detection value. The specific implementation process includes:

[0021] The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained;

[0022] The displacement characteristics include horizontal displacement and vertical displacement; the permeability characteristics include pore water pressure and permeability flow; the defect characteristics include crack characteristics, depression characteristics and cavity characteristics;

[0023] Further, the horizontal displacement amount and the vertical displacement amount are compared with a horizontal displacement threshold and a vertical displacement threshold respectively to obtain a displacement detection value;

[0024] Further, the pore water pressure and the seepage flow rate are compared with a pore water pressure threshold and a seepage flow rate threshold, respectively, to obtain a seepage detection value;

[0025] Further, the crack characteristics, the depression characteristics, and the cavity characteristics are evaluated to obtain defect detection values;

[0026] Furthermore, a weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value.

[0027] Furthermore, the pre-processed environmental radar image data is input into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam surrounding environment; and the soil features, vegetation features, and water level features are evaluated to obtain a second dam deformation detection value. A specific implementation process includes:

[0028] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features;

[0029] According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients;

[0030] The soil characteristics, the vegetation characteristics and the water level characteristics are weightedly evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient and the water level characteristic coefficient respectively to obtain a second dam deformation detection value.

[0031] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model to update the parameters. The specific implementation process of generating dam deformation early warning information based on the data results of the dam deformation prediction model includes:

[0032] Obtain historical and real-time detection data;

[0033] The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data;

[0034] Furthermore, the historical detection data is input into a dam deformation prediction model for training to obtain a dam deformation prediction weight;

[0035] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model, the model parameters are updated, and the final model prediction weight is output; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight;

[0036] Furthermore, the real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation early warning information.

[0037] A radar-based dam deformation detection system comprises: a system control module, a data acquisition module, a data processing module, a dam deformation detection module and an output module; wherein the system control module is used to control the start, pause and stop of the system; the data acquisition module is used to obtain data samples collected by a synthetic aperture radar; the data processing module is used to perform preprocessing operations on the data samples; the dam deformation detection module is used to detect the degree of dam deformation from two aspects: the dam itself and the environmental impact, and obtain a first dam deformation detection value and a second dam deformation detection value; the output module is used to output the detection results, and comprises: a judgment unit, an early warning unit and a prompt unit.

[0038] Furthermore, the data processing module preprocesses the dam radar image data and the environment radar image data to obtain the preprocessed dam radar image data and the preprocessed environment radar image data. The specific implementation process includes:

[0039] Acquire dam radar image data and environmental radar image data;

[0040] Further, using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data;

[0041] Further, histogram equalization is used to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data;

[0042] Furthermore, a normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

[0043] Furthermore, the dam deformation detection module evaluates the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain the first dam deformation detection value. The specific implementation process includes:

[0044] The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained;

[0045] The displacement characteristics include horizontal displacement and vertical displacement; the permeability characteristics include pore water pressure and permeability flow; the defect characteristics include crack characteristics, depression characteristics and cavity characteristics;

[0046] Further, the horizontal displacement amount and the vertical displacement amount are compared with a horizontal displacement threshold and a vertical displacement threshold respectively to obtain a displacement detection value;

[0047] Further, the pore water pressure and the seepage flow rate are compared with a pore water pressure threshold and a seepage flow rate threshold, respectively, to obtain a seepage detection value;

[0048] Further, the crack characteristics, the depression characteristics, and the cavity characteristics are evaluated to obtain defect detection values;

[0049] Furthermore, a weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value.

[0050] Furthermore, the dam deformation detection module evaluates soil characteristics, vegetation characteristics, and water level characteristics to obtain a second dam deformation detection value. The specific implementation process includes:

[0051] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features;

[0052] According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients;

[0053] The soil characteristics, the vegetation characteristics and the water level characteristics are weightedly evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient and the water level characteristic coefficient respectively to obtain a second dam deformation detection value.

[0054] Furthermore, the early warning unit inputs the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model to update the parameters. The specific implementation process of generating the dam deformation early warning information according to the data results of the dam deformation prediction model includes:

[0055] Obtain historical and real-time detection data;

[0056] The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data;

[0057] Furthermore, the historical detection data is input into a dam deformation prediction model for training to obtain a dam deformation prediction weight;

[0058] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model, the model parameters are updated, and the final model prediction weight is output; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight;

[0059] Furthermore, the real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation early warning information.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] 1. The present invention proposes a dam deformation degree detection function for detecting the degree of deformation of the dam itself. This function evaluates the dam characteristics output by the dam feature recognition model and uses the calculated first dam deformation detection value to detect the degree of deformation from the dam itself. The first dam deformation detection value is calculated by comprehensively evaluating the displacement characteristics, permeability characteristics, and defect characteristics of the dam. This function can effectively improve the reliability and accuracy of dam deformation detection.

[0062] 2. The present invention proposes an environmental impact deformation detection function for detecting the degree of influence of the external environment on dam deformation. This function evaluates the environmental characteristics output by the environmental characteristic recognition model and uses the calculated second dam deformation detection value to detect the degree of dam deformation from the perspective of external environmental influence. The second dam deformation detection value is obtained by evaluating the degree of influence of soil characteristics, vegetation characteristics, and water level characteristics on dam deformation. This function can effectively improve the reliability and accuracy of dam deformation detection.

[0063] 3. The present invention proposes a dam deformation early warning function for providing real-time dam deformation early warning information based on the changing trends of the dam condition and environmental conditions; this function uses historical detection data to train a dam deformation prediction model to obtain dam deformation prediction weights; then, the dam deformation prediction model is updated with the dam deformation prediction weights and real-time detection data to obtain final model prediction weights; this function comprehensively evaluates the real-time detection data and the final model prediction weights to obtain dam deformation early warning information; the dam deformation early warning information is used to determine whether the degree of dam deformation in a future period of time exceeds a warning threshold; this function can effectively improve the reliability and accuracy of dam deformation detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 Schematic diagram of a flow chart of a radar-based dam deformation detection method of the present invention;

[0065] Figure 2 It is a structural schematic diagram of the dam feature recognition model of the present invention;

[0066] Figure 3 It is a structural schematic diagram of the dam deformation prediction model of the present invention;

[0067] Figure 4 The figure is a schematic structural diagram of a radar-based dam deformation detection system of the present invention. DETAILED DESCRIPTION

[0068] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0069] As a key hub of water conservancy projects, dams play a significant role in regulating the spatial and temporal distribution of water resources. By storing and releasing water, dams not only effectively control flood flows but also ensure a stable water supply during droughts.

[0070] As an important hydraulic structure, the dam will be affected by external loads and changes in the surrounding geological structure during its long-term operation, which will cause the dam to deform to a certain extent during operation. In order to ensure the safety and stability of the dam, it is crucial to conduct regular deformation detection of the dam, so as to timely discover potential safety hazards and ensure that the dam continues to play its role in regulating water resources throughout its service life.

[0071] Currently, existing technologies for radar-based dam deformation detection still have shortcomings. Firstly, existing technologies fail to consider the impact of external environmental changes on dam deformation detection, which reduces the reliability and accuracy of dam deformation detection. Secondly, existing technologies do not use historical data to predict dam deformation trends, nor do they provide corresponding dam deformation early warning information, which reduces the reliability and accuracy of dam deformation detection.

[0072] Example 1

[0073] The specific implementation process in the embodiment of the present application will be realized by a radar-based dam deformation detection method of the present invention, see Figure 1 The process of the method proposed in the present invention is described in the following; a radar-based dam deformation detection method includes:

[0074] S10. Acquire dam radar image data and environmental radar image data;

[0075] S20. Preprocess the dam radar image data and the environmental radar image data to obtain preprocessed dam radar image data and preprocessed environmental radar image data;

[0076] S30. Inputting the pre-processed dam radar image data into a dam feature recognition model for processing to obtain dam features; evaluating the dam features to obtain a first dam deformation detection value;

[0077] S40. Inputting the pre-processed environmental radar image data into an environmental feature recognition model for processing to obtain environmental features; evaluating the environmental features to obtain a second dam deformation detection value;

[0078] S50. Perform a weighted evaluation on the first dam deformation detection value and the second dam deformation detection value to obtain a comprehensive dam deformation detection value;

[0079] S60. Compare the comprehensive dam deformation detection value with the detection threshold. If it exceeds the detection threshold, inform relevant personnel to handle it; otherwise, generate dam deformation warning information based on the data results of the dam deformation prediction model.

[0080] Furthermore, the specific implementation process of a radar-based dam deformation detection method is as follows:

[0081] Acquire dam radar image data and environmental radar image data collected by synthetic aperture radar;

[0082] Preprocessing the dam radar image data and the environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data;

[0083] Inputting the pre-processed dam radar image data into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; evaluating the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain a first dam deformation detection value;

[0084] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam's surrounding environment; evaluating the soil features, vegetation features, and water level features to obtain a second dam deformation detection value;

[0085] Performing a weighted evaluation on the first dam deformation detection value and the second dam deformation detection value to obtain a comprehensive dam deformation detection value;

[0086] The comprehensive dam deformation detection value is compared with the detection threshold. If it exceeds the detection threshold, relevant personnel are informed to handle it; otherwise, the dam deformation prediction weight and real-time detection data are input into the dam deformation prediction model for parameter update, and dam deformation early warning information is generated according to the data results of the dam deformation prediction model.

[0087] In this embodiment, a radar-based dam deformation detection method is proposed. First, dam radar image data and environmental radar image data are obtained. Then, a feature recognition model is used to identify the pre-processed dam radar image data and the pre-processed environmental radar image data, respectively. A first dam deformation detection value and a second dam deformation detection value are calculated based on the dam features and environmental features output by the model. Next, the first dam deformation detection value and the second dam deformation detection value are comprehensively evaluated to obtain a comprehensive dam deformation detection value. Finally, the comprehensive dam deformation detection value is compared with a detection threshold. If the threshold is exceeded, relevant personnel are informed to handle the problem. Otherwise, the dam deformation prediction weight and real-time detection data are input into the dam deformation prediction model for parameter update, and dam deformation warning information is generated based on the data results of the dam deformation prediction model. This method can effectively improve the reliability and accuracy of dam deformation detection.

[0088] For the purpose of specific explanation, the present invention is described in conjunction with the following examples, as follows:

[0089] Acquire dam radar image data and environmental radar image data collected by synthetic aperture radar;

[0090] In this embodiment, the data acquisition device used is a synthetic aperture radar mounted on an unmanned aerial vehicle. The synthetic aperture radar is a high-resolution imaging radar that has strong anti-interference capabilities in meteorological environments with low visibility and can output high-resolution radar image data, thereby further improving the reliability and accuracy of dam deformation detection.

[0091] Furthermore, the dam radar image data and the environment radar image data are preprocessed to obtain preprocessed dam radar image data and preprocessed environment radar image data:

[0092] Acquire dam radar image data and environmental radar image data;

[0093] Further, using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data;

[0094] Further, histogram equalization is used to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data;

[0095] Furthermore, a normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

[0096] The preprocessing process of radar image data in this embodiment includes: Gaussian denoising, histogram equalization and normalization operations; the Gaussian denoising is used to reduce the environmental noise of the image data to improve the accuracy of the feature recognition model; the histogram equalization is used to enhance the contrast and brightness of the image to improve the detection and recognition capabilities of edges and sharp image features; the normalization operation is used to improve the computational efficiency of the feature recognition model; this process can effectively improve the accuracy and reliability of dam deformation detection.

[0097] Furthermore, the pre-processed dam radar image data is input into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; and the displacement characteristics, permeability characteristics, and defect characteristics are evaluated to obtain a first dam deformation detection value. The specific implementation process includes:

[0098] The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained;

[0099] The structure of the dam feature recognition model is as follows: Figure 2 As shown in the figure, it includes: input layer, multi-scale feature extraction layer, feature labeling layer, attention feature recognition layer and output layer; the specific implementation process of the model includes:

[0100] Inputting the radar image data into the input layer to obtain a dam feature map;

[0101] Furthermore, the dam feature map is input into the multi-scale feature extraction layer to obtain the multi-scale features of the dam; wherein the multi-scale feature extraction layer includes convolution kernels of 3×3, 5×5, 7×7 and 11×11 sizes for extracting multi-scale features;

[0102] Furthermore, the multi-scale features of the dam are input into the feature labeling layer to obtain an anchor frame; wherein the anchor frame is set to five different sizes;

[0103] Furthermore, the features in the anchor frame are extracted and input into the attention feature recognition layer, and the channel attention and spatial attention mechanisms are used to improve the model recognition performance to obtain the recognition result;

[0104] Furthermore, the recognition result is input into the output layer to obtain recognition feature data.

[0105] In this embodiment, a dam feature recognition model is proposed for identifying image features related to the degree of influence of dam deformation; the dam feature recognition model includes: an input layer, a multi-scale feature extraction layer, a feature labeling layer, an attention feature recognition layer and an output layer; wherein, the attention feature recognition layer combines spatial attention and channel attention mechanisms to improve the recognition ability of the model, so that the dam feature recognition model can output more accurate recognition results, further improving the reliability and accuracy of dam deformation detection.

[0106] Furthermore, the displacement characteristics include: horizontal displacement and vertical displacement; the permeability characteristics include: pore water pressure and permeability flow; the defect characteristics include: crack characteristics, depression characteristics and cavity characteristics;

[0107] Furthermore, the horizontal displacement amount and the vertical displacement amount are compared with the horizontal displacement threshold and the vertical displacement threshold respectively to obtain a displacement detection value; wherein the calculation formula of the displacement detection value is:

[0108]

[0109] Among them, wyjc i is represented by the displacement detection value of the i-th feature sample; arctan is represented by the hyperbolic tangent function; α1 is represented by the weight factor of the horizontal displacement ratio; hd i It is expressed as the horizontal displacement of the i-th sample; hd th It is represented as the horizontal displacement threshold; α2 is represented as the weight factor of the vertical displacement ratio; hd i Expressed as the vertical displacement of the i-th sample; hd th It is represented as the vertical displacement threshold.

[0110] In this embodiment, both α1 and α2 are set to 0.5; the setting of the horizontal displacement threshold and the vertical displacement threshold is related to factors such as the dam material, dam service life and geographical environment, and requires relevant staff to flexibly set them based on work experience and are not unique.

[0111] Furthermore, the pore water pressure and the seepage flow rate are compared with the pore water pressure threshold and the seepage flow rate threshold, respectively, to obtain a permeation detection value; wherein the calculation formula of the permeation detection value is:

[0112]

[0113] Among them, stjc i is represented by the permeability test value of the i-th characteristic sample; β1 is represented by the weight factor of the pore water pressure ratio; sy i Represented as the pore water pressure of the i-th sample; sy th is represented by the pore water pressure threshold; β2 is represented by the weight factor of the seepage flow ratio; ll i Expressed as the permeate flow rate of the i-th sample; th It is expressed as the permeation flow threshold.

[0114] In this embodiment, β1 and β2 are both set to 0.5; the setting of the pore water pressure threshold and the seepage flow threshold is related to factors such as the dam material, dam service life and geographical environment, and requires relevant staff to flexibly set them based on work experience and are not unique.

[0115] Furthermore, the crack characteristics, the depression characteristics, and the cavity characteristics are evaluated to obtain a defect detection value; wherein the calculation formula of the defect detection value is:

[0116]

[0117] Among them, qxjc i is represented by the defect detection value of the i-th feature sample; γ1 is represented by the crack weight factor; lf i It is expressed as the maximum number of cracks in the i-th sample; It is expressed as the length of the j1th crack of the i-th sample; It is represented by the depth of the j1th crack of the i-th sample; γ2 is represented by the concave weight factor; ax i It is expressed as the maximum number of depressions of the i-th sample; It is expressed as the area of the j2th concavity of the i-th sample; It is represented as the depth of the j2th concavity of the i-th sample; γ3 is the hole weight factor; kd i It is expressed as the maximum number of holes in the i-th sample; It is expressed as the area of the j3th hole in the i-th sample.

[0118] In this embodiment, γ1, γ2 and γ3 are set to 0.35, 0.3 and 0.35 respectively; the crack weight factor, the depression weight factor and the cavity weight factor can be flexibly set according to actual conditions and are not unique.

[0119] Furthermore, a weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value; wherein the calculation formula for the first dam deformation detection value is:

[0120]

[0121] ω wy +ω st +ω qx =1.0;

[0122] Where DYJC represents the first dam deformation detection value; N represents the number of detection samples; ω wy Expressed as displacement detection weight; wyjc i It is represented as the displacement detection value of the i-th sample; ω st Expressed as penetration test weight; stjc i Expressed as the penetration test value of the i-th sample; ω qx Expressed as defect detection weight; qxjc i Denoted as the defect detection value of the i-th sample.

[0123] In this embodiment, the settings of the displacement detection weight, the penetration detection weight, and the defect detection weight can be flexibly adjusted according to the degree of influence on the dam deformation and are not unique.

[0124] In this embodiment, a dam deformation degree detection function is proposed for detecting the degree of deformation of the dam itself. This function evaluates the dam characteristics output by the dam characteristic recognition model respectively, and uses the calculated first dam deformation detection value to detect the degree of deformation from the dam itself. The first dam deformation detection value is calculated by comprehensively evaluating the displacement characteristics, permeability characteristics, and defect characteristics of the dam. This function can effectively improve the reliability and accuracy of dam deformation detection.

[0125] Furthermore, the pre-processed environmental radar image data is input into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam surrounding environment; and the soil features, vegetation features, and water level features are evaluated to obtain a second dam deformation detection value. A specific implementation process includes:

[0126] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features;

[0127] According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients;

[0128] The soil characteristics, the vegetation characteristics, and the water level characteristics are weighted and evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient, and the water level characteristic coefficient, respectively, to obtain a second dam deformation detection value; wherein the calculation formula for the second dam deformation detection value is:

[0129]

[0130] λ tr +λ zb +λ sw =1.0;

[0131] Where DEJC is the deformation detection value of the second dam; M is the number of detection samples; λ tr Expressed as the soil characteristic coefficient; trcj i Expressed as the soil settlement value of the i-th sample; trcj th Expressed as soil settlement threshold; λ zb Expressed as the vegetation characteristic coefficient; A zb Expressed as vegetation coverage area; A jc Expressed as the environmental detection area; λ sw Expressed as the water level characteristic coefficient; swgd i Expressed as the upstream water level height of the i-th sample; swgd th Expressed as upstream water level height threshold.

[0132] In this embodiment, tr ,λ zb and λ sw They are set to 0.3, 0.3 and 0.4 respectively. At the same time, these characteristic coefficients can be adjusted according to actual conditions. The settings of soil subsidence threshold, environmental detection area and upstream water level height threshold will be affected by the geographical environment and will vary. Therefore, relevant personnel are required to set them according to the actual environment. They do not have unique values.

[0133] In this embodiment, an environmental impact deformation detection function is proposed for detecting the degree of influence of the external environment on dam deformation. This function evaluates the environmental characteristics output by the environmental characteristic recognition model and uses the calculated second dam deformation detection value to detect the degree of dam deformation from the perspective of external environmental influence. The second dam deformation detection value is obtained by evaluating the degree of influence of soil characteristics, vegetation characteristics, and water level characteristics on dam deformation. This function can effectively improve the reliability and accuracy of dam deformation detection.

[0134] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model to update the parameters. The specific implementation process of generating dam deformation early warning information based on the data results of the dam deformation prediction model includes:

[0135] Obtain historical and real-time detection data;

[0136] The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data;

[0137] Furthermore, the historical detection data is input into a dam deformation prediction model for training to obtain a dam deformation prediction weight;

[0138] In this embodiment, the dam deformation prediction model uses a hybrid network combining residual and LSTM. The structure of the model is as follows: Figure 3 As shown in the figure, the hybrid network adopts a dual-branch structure, inputting the input data into the residual branch and the LSTM branch for feature processing respectively, and finally fusing the output features of different networks to obtain the prediction results; the residual branch uses three residual blocks to extract convolution local features; the LSTM branch uses three LSTM layers to extract long-distance dependency features and learn the correlation between the dam deformation degree and the environment that changes over time.

[0139] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model, the model parameters are updated, and the final model prediction weight is output; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight;

[0140] Furthermore, the real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation warning information; wherein, the calculation formula of the dam deformation warning information is:

[0141]

[0142] Wherein, BXYJ represents the dam deformation early warning information used to predict and judge the dam deformation degree in the subsequent time period; P represents the number of data samples; INF is the real-time substrate temperature parameter of the i-th sample; (i,db) The real-time dam detection data represented as the i-th sample; INF is the final environmental prediction weight of the i-th sample; (i,hj) The real-time environment detection data represented as the i-th sample.

[0143] In this embodiment, a dam deformation early warning function is proposed for providing real-time dam deformation early warning information based on the changing trends of the dam condition and environmental conditions. This function uses historical detection data to train a dam deformation prediction model to obtain dam deformation prediction weights. Then, the dam deformation prediction model is updated with the dam deformation prediction weights and real-time detection data to obtain final model prediction weights. This function comprehensively evaluates the real-time detection data and the final model prediction weights to obtain dam deformation early warning information. The dam deformation early warning information is used to determine whether the degree of dam deformation in a future period of time exceeds a warning threshold. This function can effectively improve the reliability and accuracy of dam deformation detection.

[0144] Example 2

[0145] As an embodiment of the present invention, refer to Figure 4 ,A radar-based dam deformation detection system includes: a system control module, a data acquisition module, a data processing module, a dam deformation detection module and an output module;

[0146] Wherein, the system control module is used to control the start, pause and stop of the system;

[0147] The data acquisition module is used to obtain data samples collected by synthetic aperture radar;

[0148] The data processing module is used to perform preprocessing operations on the data samples;

[0149] Furthermore, the data processing module preprocesses the dam radar image data and the environment radar image data to obtain the preprocessed dam radar image data and the preprocessed environment radar image data. The specific implementation process includes:

[0150] Acquire dam radar image data and environmental radar image data;

[0151] Further, using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data;

[0152] Further, histogram equalization is used to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data;

[0153] Furthermore, a normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

[0154] The dam deformation detection module is used to detect the degree of dam deformation from two aspects: the dam itself and the environmental impact, and obtain a first dam deformation detection value and a second dam deformation detection value;

[0155] Furthermore, the dam deformation detection module evaluates the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain the first dam deformation detection value. The specific implementation process includes:

[0156] The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained;

[0157] The displacement characteristics include horizontal displacement and vertical displacement; the permeability characteristics include pore water pressure and permeability flow; the defect characteristics include crack characteristics, depression characteristics and cavity characteristics;

[0158] Furthermore, the horizontal displacement amount and the vertical displacement amount are compared with the horizontal displacement threshold and the vertical displacement threshold respectively to obtain a displacement detection value; wherein the calculation formula of the displacement detection value is:

[0159]

[0160] Among them, wyjc i is represented by the displacement detection value of the i-th feature sample; arctan is represented by the hyperbolic tangent function; α1 is represented by the weight factor of the horizontal displacement ratio; hd i It is expressed as the horizontal displacement of the i-th sample; hd th It is represented as the horizontal displacement threshold; α2 is represented as the weight factor of the vertical displacement ratio; hd i Expressed as the vertical displacement of the i-th sample; hd th It is represented as the vertical displacement threshold.

[0161] Furthermore, the pore water pressure and the seepage flow rate are compared with the pore water pressure threshold and the seepage flow rate threshold, respectively, to obtain a permeation detection value; wherein the calculation formula of the permeation detection value is:

[0162]

[0163] Among them, stjc i is represented by the permeability test value of the i-th characteristic sample; β1 is represented by the weight factor of the pore water pressure ratio; sy i Represented as the pore water pressure of the i-th sample; sy th is represented by the pore water pressure threshold; β2 is represented by the weight factor of the seepage flow ratio; ll i Expressed as the permeate flow rate of the i-th sample; th It is expressed as the permeation flow threshold.

[0164] Furthermore, the crack characteristics, the depression characteristics, and the cavity characteristics are evaluated to obtain a defect detection value; wherein the calculation formula of the defect detection value is:

[0165]

[0166] Among them, qxjc i is represented by the defect detection value of the i-th feature sample; γ1 is represented by the crack weight factor; lf i It is expressed as the maximum number of cracks in the i-th sample; It is expressed as the length of the j1th crack of the i-th sample; It is represented by the depth of the j1th crack of the i-th sample; γ2 is represented by the concave weight factor; ax i It is expressed as the maximum number of depressions of the i-th sample; It is expressed as the area of the j2th concavity of the i-th sample; It is represented as the depth of the j2th concavity of the i-th sample; γ3 is the hole weight factor; kd i It is expressed as the maximum number of holes in the i-th sample; It is expressed as the area of the j3th hole in the i-th sample.

[0167] Furthermore, a weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value; wherein the calculation formula for the first dam deformation detection value is:

[0168]

[0169] ω wy +ω st +ω qx =1.0;

[0170] Where DYJC represents the first dam deformation detection value; N represents the number of detection samples; ω wyExpressed as displacement detection weight; wyjc i It is represented as the displacement detection value of the i-th sample; ω st Expressed as penetration test weight; stjc i Expressed as the penetration test value of the i-th sample; ω qx Expressed as defect detection weight; qxjc i Denoted as the defect detection value of the i-th sample.

[0171] Furthermore, the dam deformation detection module evaluates soil characteristics, vegetation characteristics, and water level characteristics to obtain a second dam deformation detection value. The specific implementation process includes:

[0172] Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features;

[0173] According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients;

[0174] The soil characteristics, the vegetation characteristics, and the water level characteristics are weighted and evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient, and the water level characteristic coefficient, respectively, to obtain a second dam deformation detection value; wherein the calculation formula for the second dam deformation detection value is:

[0175]

[0176] λ tr +λ zb +λ sw =1.0;

[0177] Where DEJC is the deformation detection value of the second dam; M is the number of detection samples; λ tr Expressed as the soil characteristic coefficient; trcj i Expressed as the soil settlement value of the i-th sample; trcj th Expressed as soil settlement threshold; λ zb Expressed as the vegetation characteristic coefficient; A zb Expressed as vegetation coverage area; A jc Expressed as the environmental detection area; λ sw Expressed as the water level characteristic coefficient; swgd i Expressed as the upstream water level height of the i-th sample; swgd th Expressed as upstream water level height threshold.

[0178] The output module is used to output the detection results, and includes: a judgment unit, an early warning unit and a prompt unit.

[0179] The judgment unit is used to output the judgment results of the comprehensive dam deformation detection value and the detection threshold, and the dam deformation warning value and the warning threshold; if it exceeds the threshold range, a signal is sent to the prompt unit and the prompt unit is reminded to output relevant prompt information; the comprehensive dam deformation detection value is obtained by integrating the first dam deformation detection value and the second dam deformation detection value.

[0180] The early warning unit inputs the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model to update the parameters. The specific implementation process of generating the dam deformation early warning information according to the data results of the dam deformation prediction model includes:

[0181] Obtain historical and real-time detection data;

[0182] The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data;

[0183] Furthermore, the historical detection data is input into a dam deformation prediction model for training to obtain a dam deformation prediction weight;

[0184] Furthermore, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model, the model parameters are updated, and the final model prediction weight is output; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight;

[0185] Furthermore, the real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation early warning information.

[0186] The prompt unit is used to receive the signal from the judgment unit and send prompt information to relevant personnel in the form of voice or image.

[0187] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A radar-based dam deformation detection method, characterized in that: include: Acquire dam radar image data and environmental radar image data collected by synthetic aperture radar; Preprocessing the dam radar image data and the environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data; Inputting the pre-processed dam radar image data into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; evaluating the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain a first dam deformation detection value; Inputting the pre-processed environmental radar image data into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam's surrounding environment; evaluating the soil features, vegetation features, and water level features to obtain a second dam deformation detection value; Performing a weighted evaluation on the first dam deformation detection value and the second dam deformation detection value to obtain a comprehensive dam deformation detection value; Comparing the comprehensive dam deformation detection value with a detection threshold, and if it exceeds the detection threshold, notifying relevant personnel to handle the situation; Otherwise, the dam deformation prediction weight and the real-time detection data are input into the dam deformation prediction model to update the parameters, and dam deformation early warning information is generated according to the data results of the dam deformation prediction model.

2. The radar-based dam deformation detection method according to claim 1, characterized in that: The dam radar image data and the environment radar image data are preprocessed to obtain preprocessed dam radar image data and preprocessed environment radar image data: Acquire dam radar image data and environmental radar image data; Using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data; Using histogram equalization to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data; A normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

3. The radar-based dam deformation detection method according to claim 1, characterized in that: Inputting the pre-processed dam radar image data into a dam feature recognition model for processing to obtain displacement characteristics, permeability characteristics, and defect characteristics of the dam; The specific implementation process of evaluating the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain the first dam deformation detection value includes: The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained; The displacement characteristics include horizontal displacement and vertical displacement; the permeability characteristics include pore water pressure and permeability flow; the defect characteristics include crack characteristics, depression characteristics and cavity characteristics; Comparing the horizontal displacement amount and the vertical displacement amount with a horizontal displacement threshold and a vertical displacement threshold respectively to obtain a displacement detection value; Comparing the pore water pressure and the seepage flow rate with a pore water pressure threshold and a seepage flow rate threshold, respectively, to obtain a seepage detection value; Evaluating the crack characteristics, the depression characteristics, and the cavity characteristics to obtain a defect detection value; A weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value.

4. The radar-based dam deformation detection method according to claim 1, characterized in that: Inputting the pre-processed environmental radar image data into an environmental feature recognition model for processing to obtain soil features, vegetation features, and water level features of the dam's surrounding environment; The specific implementation process of evaluating the soil characteristics, the vegetation characteristics, and the water level characteristics to obtain the second dam deformation detection value includes: Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features; According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients; The soil characteristics, the vegetation characteristics and the water level characteristics are weightedly evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient and the water level characteristic coefficient respectively to obtain a second dam deformation detection value.

5. The radar-based dam deformation detection method according to claim 1, characterized in that: The specific implementation process of inputting the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model to update the parameters and generating the dam deformation early warning information according to the data results of the dam deformation prediction model includes: Obtain historical and real-time detection data; The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data; Inputting the historical detection data into a dam deformation prediction model for training to obtain a dam deformation prediction weight; Inputting the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model, updating the model parameters, and outputting a final model prediction weight; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight; The real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation early warning information.

6. A radar-based dam deformation detection system, characterized in that: include: A system control module, a data acquisition module, a data processing module, a dam deformation detection module and an output module; wherein the system control module is used to control the start, pause and stop of the system; the data acquisition module is used to obtain data samples collected by the synthetic aperture radar; the data processing module is used to perform preprocessing operations on the data samples; the dam deformation detection module is used to detect the degree of dam deformation from two aspects: the dam itself and the environmental impact, and obtain a first dam deformation detection value and a second dam deformation detection value; the output module is used to output the detection results, and includes: a judgment unit, an early warning unit and a prompt unit.

7. The radar-based dam deformation detection system according to claim 6, characterized in that: The data processing module preprocesses the dam radar image data and the environment radar image data to obtain the preprocessed dam radar image data and the preprocessed environment radar image data. The specific implementation process includes: Acquire dam radar image data and environmental radar image data; Using Gaussian filtering to remove noise from the dam radar image data and the environment radar image data to obtain denoised dam radar image data and denoised environment radar image data; Using histogram equalization to adjust the contrast and brightness of the denoised dam radar image data and the denoised environment radar image data to obtain enhanced dam radar image data and enhanced environment radar image data; A normalization operation is performed on the enhanced dam radar image data and the enhanced environment radar image data to obtain preprocessed dam radar image data and preprocessed environment radar image data.

8. The radar-based dam deformation detection system according to claim 6, characterized in that: The dam deformation detection module evaluates the displacement characteristics, the permeability characteristics, and the defect characteristics to obtain the first dam deformation detection value. The specific implementation process includes: The pre-processed dam radar image data is input into the dam feature recognition model for recognition, and the displacement characteristics, permeability characteristics and defect characteristics of the dam are obtained; The displacement characteristics include horizontal displacement and vertical displacement; the permeability characteristics include pore water pressure and permeability flow; the defect characteristics include crack characteristics, depression characteristics and cavity characteristics; Comparing the horizontal displacement amount and the vertical displacement amount with a horizontal displacement threshold and a vertical displacement threshold respectively to obtain a displacement detection value; Comparing the pore water pressure and the seepage flow rate with a pore water pressure threshold and a seepage flow rate threshold, respectively, to obtain a seepage detection value; Evaluating the crack characteristics, the depression characteristics, and the cavity characteristics to obtain a defect detection value; A weighted evaluation is performed on the displacement detection value, the penetration detection value, and the defect detection value to obtain a first dam deformation detection value.

9. The radar-based dam deformation detection system according to claim 6, characterized in that: The dam deformation detection module evaluates soil characteristics, vegetation characteristics, and water level characteristics to obtain a second dam deformation detection value. The specific implementation process includes: Inputting the pre-processed environmental radar image data into an environmental feature recognition model for recognition to obtain dam environmental features; wherein the dam environmental features include soil features, vegetation features, and water level features; According to the degree of influence on dam deformation, corresponding weight coefficients are set for different environmental characteristics to obtain soil characteristic coefficients, vegetation characteristic coefficients and water level characteristic coefficients; The soil characteristics, the vegetation characteristics and the water level characteristics are weightedly evaluated with the soil characteristic coefficient, the vegetation characteristic coefficient and the water level characteristic coefficient respectively to obtain a second dam deformation detection value.

10. The radar-based dam deformation detection system according to claim 6, characterized in that: The early warning unit inputs the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model to update the parameters. The specific implementation process of generating the dam deformation early warning information according to the data results of the dam deformation prediction model includes: Obtain historical and real-time detection data; The historical detection data includes historical dam detection data and historical environmental detection data; the real-time detection data includes real-time dam detection data and real-time environmental detection data; Inputting the historical detection data into a dam deformation prediction model for training to obtain a dam deformation prediction weight; Inputting the dam deformation prediction weight and the real-time detection data into the dam deformation prediction model, updating the model parameters, and outputting a final model prediction weight; wherein the final model prediction weight includes: a final dam prediction weight and a final environment prediction weight; The real-time detection data and the final model prediction weight are comprehensively evaluated to obtain dam deformation early warning information.

Images