Concrete panel dam face crack identification system based on time-space correlation

Abstract

The invention relates to the technical field of dam body safety monitoring, in particular to a concrete face dam face crack recognition system based on time-space correlation, which tracks the evolution of a crack along with time by constructing a time sequence model, and quantifies the change degree of the crack by comparing and analyzing current and historical image data. And the environment adaptation module adjusts a detection threshold value according to real-time environment conditions, so that high accuracy can be kept under different weather and illumination conditions. And then a decision-making module integrates all data and intelligently judges whether maintenance or further inspection is needed or not, so that early discovery and preventive maintenance of the cracks of the concrete face dam face are realized. In conclusion, according to the scheme, the system can sensitively capture the tiny change of the crack through time sequence modeling and a change detection mechanism, it is ensured that the system can still operate stably under changeable natural conditions, the monitoring applicability is improved, and meanwhile the manual analysis burden is relieved through an automatic decision support function.

Description

Technical Field

[0001] The invention relates to the technical field of dam body safety monitoring, and in particular to a concrete panel dam surface crack identification system based on time-space correlation. Background Art

[0002] As an important structure in water conservancy facilities, the health status of the concrete panel dam is directly related to the safety and service life of the dam. In recent years, with the development of computer vision and image processing technology, crack detection methods based on image recognition have gradually become an important means of crack detection in concrete panel dams. These methods collect multi-source imaging data of the panel (such as visible light images, infrared images or laser scanning data, etc.) and use advanced image recognition technology to automatically identify cracks on the panel. Compared with traditional manual inspection methods, crack detection through image recognition technology has the advantages of more efficiency, wider coverage and higher degree of automation, which greatly improves the efficiency of crack detection.

[0003] However, in actual applications, existing image recognition technology cannot accurately distinguish between real cracks and environmental interference due to factors such as weather conditions, lighting changes or temperature fluctuations, which can affect image quality. For example, when there is a significant difference in light and shadow, shadows may be mistaken for cracks; when temperature changes cause the material to expand and contract, causing the surface of the material to deform, it may also be mistaken for a crack.

[0004] Such misjudgments not only trigger unnecessary maintenance activities and increase operating costs, but may also mask crack problems that really need attention, delay necessary repair measures, and pose potential dangers to the dam. Therefore, it is urgent to develop a more intelligent and reliable crack identification system to ensure that high-precision detection capabilities can be maintained in complex environments and significantly reduce the occurrence of false alarms. Summary of the invention

[0005] The technical problem solved by the present invention is to provide a concrete panel dam surface crack recognition system based on time-space correlation to solve the technical problem that the existing image recognition technology is difficult to accurately distinguish between real cracks and environmental interference due to factors such as weather conditions, light changes or temperature fluctuations.

[0006] The basic solution provided by the present invention is: a concrete panel dam surface crack identification system based on time-space association, comprising a data acquisition terminal and a server; the data acquisition terminal comprises an image acquisition module and a communication module, the image acquisition module is used to acquire panel images, and the communication module is used to establish a wireless connection between the data acquisition terminal and the server;

[0007] The server includes a crack identification module, a time series modeling module, a change detection module, a multi-period verification module, an environmental adaptation module and a decision-making module;

[0008] The crack recognition module is used to identify the crack area on the panel and extract the crack features in the image;

[0009] The time series modeling module is used to build a time series model for each monitoring point and record the state changes of the panel in different periods of time;

[0010] The change detection module quantifies the degree of change of the cracks by comparing the differences between the current image and the historical image:

[0011]

[0012] In the formula, C t is the degree of crack change, I t (i) and I t-1 (i) represents the grayscale value of the i-th pixel in the current image and the historical image respectively; N is the total number of pixels in the image; by calculating the pixel difference between the two images, a difference map is generated, and C t To quantify the degree of crack change;

[0013] The multi-period detection module is used to continuously monitor the panel area where the crack change degree reaches the preset value T for multiple periods, and conduct spatiotemporal correlation analysis based on the monitoring results of multiple periods:

[0014] Y t =g(C t ,C t-1 ,…,C t-n )

[0015] Where Y t is the crack state at the tth moment, C t represents the change at the tth moment, g is a nonlinear function used to capture the change trend of the crack. t Conduct analysis to determine the expansion trend of cracks;

[0016] The environmental adaptation module is used to dynamically adjust the preset value T according to the environmental parameters to improve the detection accuracy in different environments. The mapping relationship between the environmental parameters and the preset values ​​is:

[0017] T=h(E)

[0018] In the formula, E is the environmental parameter, and h is a nonlinear function;

[0019] The decision module is used to determine whether to perform maintenance based on the output results of the change detection module, the multi-period detection module, and the environmental adaptation module:

[0020] D=k(T,C t ,E)

[0021] Where D is the final judgment result, and k is a nonlinear function used to extract the output results from each module.

[0022] Furthermore, the image acquisition module is provided with image quality control rules:

[0023]

[0024] Where R is the required ground sampling distance, w is the camera sensor width, f is the camera focal length, and the camera parameters are adjusted according to the image acquisition height of the data acquisition terminal and its camera focal length so that the image resolution meets the requirements. H is the flight altitude required for the UAV to fly.

[0025] Further, the crack identification module includes an image preprocessing module and a crack feature extraction module;

[0026] The image preprocessing module is used to preprocess the collected images;

[0027] The crack feature extraction module is based on an image recognition algorithm and uses an edge detection algorithm to identify crack features, locate the identified cracks and measure their length, average width and maximum width.

[0028] Furthermore, the image preprocessing module uses Gaussian filtering to perform denoising and linear stretching to perform contrast enhancement:

[0029]

[0030] Where I(X,Y) is the original image, X and Y are the coordinate positions of the pixels; σ is the standard deviation, which controls the smoothness of the filter; I′(X,Y) is the enhanced image, min(I) and max(I) are the minimum and maximum grayscale values ​​of the image, and L is the number of grayscale levels.

[0031] Furthermore, the crack feature extraction module includes a coordinate positioning module and a feature calculation module.

[0032] The coordinate positioning module is used to determine the coordinate information of the crack location based on geographic information system technology:

[0033]

[0034] In the formula, G x and G y is the gradient of the image in the x and y directions; G is the gradient magnitude; θ is the gradient direction;

[0035] The feature calculation module is used to measure the length, average width and maximum width of the crack:

[0036]

[0037] Where Lc is the length of the crack; for each point (x i ,y i ), find the points (x) on both sides of it i1 ,y i1 ) and (x i2 ,y i2 ), calculate the vertical distance and get the width of the points on both sides:

[0038]

[0039] w max =max(w 1 ,w 2 ,…,w n )

[0040] In the formula, (x i ,y i ) is the coordinate of the i-th edge point, w i is the width of the i-th node; w avg Average crack width; w max Maximum crack width; crack direction vector v = (v x ,v y ), v x is the component of the crack in the x-axis direction; v y is the component of the crack in the y-axis direction.

[0041] Furthermore, the time series modeling module is used to control the data acquisition terminal to periodically acquire image data of the concrete panel dam panel, and synchronously record the timestamp and environmental parameters, and construct a complete time series data set and a time series model based on the data acquired each time:

[0042] Based on historical image data, a time series model of each monitoring point is constructed using a long short-term memory network (LSTM) to record the state changes of the panel at different time periods:

[0043] h t =LSTM(X t-1 ,h t-1 )

[0044] In the formula, h t is the hidden state of LSTM, indicating the state of the panel at time t; X t-1 is the external input at time t-1, h t-1 is the hidden state at time t-1.

[0045] Furthermore, the data acquisition terminal uses a drone carrying an image acquisition device to acquire panel images.

[0046] Furthermore, the environmental parameters include temperature, humidity and light.

[0047] The principles and advantages of the present invention are as follows: This solution tracks the evolution of cracks over time by constructing a time series model, and quantifies the degree of change of cracks by comparing and analyzing current and historical image data. The environmental adaptation module adjusts the detection threshold according to real-time environmental conditions to ensure high accuracy under different weather and lighting conditions. The decision-making module then integrates all data and intelligently determines whether repairs or further inspections are needed, thereby achieving early detection and preventive maintenance of cracks on the concrete panel dam surface. The system can keenly capture tiny changes in cracks through time series modeling and change detection mechanisms, providing a scientific basis for predicting crack development trends. While ensuring that the system can still operate stably under changing natural conditions and improving the flexibility and applicability of monitoring, it also reduces the burden of manual analysis and accelerates maintenance response speed through automated decision support functions, which is of great significance for ensuring the safety of dams and extending their service life. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 The present invention is a logic block diagram of an embodiment of a concrete panel dam surface crack identification system based on time-space correlation. DETAILED DESCRIPTION

[0049] The following is further described in detail through specific implementation methods:

[0050] The specific implementation process is as follows:

[0051] Embodiment 1

[0052] Embodiment 1 is as attached Figure 1 As shown, a concrete panel dam surface crack identification system based on time-space correlation includes a data acquisition terminal and a server; the data acquisition terminal includes an image acquisition module and a communication module, the image acquisition module is used to acquire panel images, and the communication module is used to establish a wireless connection between the data acquisition terminal and the server.

[0053] Specifically, the data acquisition terminal uses a drone carrying a camera to collect panel images. The image acquisition module has image quality control rules:

[0054]

[0055] Where R is the required ground sampling distance, w is the camera sensor width, f is the camera focal length, and the camera parameters are adjusted according to the image acquisition height of the data acquisition terminal and its camera focal length so that the image resolution meets the requirements. H is the flight altitude required for the UAV to fly.

[0056] The server includes a crack recognition module, a time series modeling module, a change detection module, a multi-period verification module, an environmental adaptation module, an environmental data acquisition module, and a decision-making module. The crack recognition module includes an image preprocessing module and a crack feature extraction module; the image preprocessing module is used to preprocess the collected images; the crack feature extraction module uses an edge detection algorithm based on an image recognition algorithm to identify crack features, locate the identified cracks, and measure their length, average width, and maximum width.

[0057] Specifically, the image preprocessing module uses Gaussian filtering to remove noise and linear stretching to enhance contrast:

[0058]

[0059] Where I(X,Y) is the original image, X and Y are the coordinate positions of the pixels; σ is the standard deviation, which controls the smoothness of the filter; I′(X,Y) is the enhanced image, min(I) and max(I) are the minimum and maximum grayscale values ​​of the image, and L is the number of grayscale levels.

[0060] The crack feature extraction module includes a coordinate positioning module and a feature calculation module.

[0061] The coordinate positioning module is used to determine the coordinate information of the crack location based on geographic information system technology:

[0062]

[0063] In the formula, G x and G y is the gradient of the image in the x and y directions; G is the gradient magnitude; θ is the gradient direction;

[0064] The feature calculation module is used to measure the length, average width and maximum width of the crack:

[0065]

[0066] Where Lc is the length of the crack; for each point (x i ,y i ), find the points (x) on both sides of it i1 ,y i1 ) and (x i2 ,y i2 ), calculate the vertical distance and get the width of the points on both sides:

[0067]

[0068]

[0069] w max =max(w1 , w 2 , ..., w n )

[0070] In the formula, (x i ,y i ) is the coordinate of the i-th edge point, w i is the width of the i-th node; w avg Average crack width; w max Maximum crack width; crack direction vector v = (v x , v y ), v x is the component of the crack in the x-axis direction; v y is the component of the crack in the y-axis direction.

[0071] The time series modeling module is used to control the data acquisition terminal to regularly collect image data of the concrete panel dam panel, and synchronously record the timestamp and environmental parameters. A complete time series data set is constructed based on the data collected each time, and a time series model is constructed:

[0072] Based on historical image data, a time series model of each monitoring point is constructed using a long short-term memory network (LSTM) to record the state changes of the panel at different time periods:

[0073] h t =LSTM(X t-1 ,h t-1 )

[0074] In the formula, h t is the hidden state of LSTM, indicating the state of the panel at time t; X t-1 is the external input at time t-1, h t-1 is the hidden state at time t-1.

[0075] The change detection module quantifies the degree of change of the cracks by comparing the differences between the current image and the historical image:

[0076]

[0077] In the formula, C t is the degree of crack change, I t (i) and I t-1 (i) represents the grayscale value of the i-th pixel in the current image and the historical image respectively; N is the total number of pixels in the image; by calculating the pixel difference between the two images, a difference map is generated, and C t To quantify the degree of crack change;

[0078] The multi-period detection module is used to continuously monitor the panel area where the crack change degree reaches the preset value T for multiple periods, and conduct spatiotemporal correlation analysis based on the monitoring results of multiple periods:

[0079] Y t =g(C t , C t-1 , ..., C t-n )

[0080] Where Y t is the crack state at the tth moment, C t represents the change at the tth moment, g is a nonlinear function used to capture the change trend of the crack. t Conduct analysis to determine the expansion trend of cracks;

[0081] The environmental data acquisition module is used to collect real-time environmental parameters through temperature sensors, humidity sensors and light sensors. Environmental parameters include temperature, humidity and light. The environmental adaptation module is used to dynamically adjust the preset value T according to the environmental parameters to improve the detection accuracy in different environments. The mapping relationship between environmental parameters and preset values ​​is:

[0082] T=h(E)

[0083] In the formula, E is the environmental parameter, and h is a nonlinear function;

[0084] The decision module is used to determine whether to perform maintenance based on the output results of the change detection module, the multi-period detection module, and the environmental adaptation module:

[0085] D=k(T,C t ,E)

[0086] Where D is the final judgment result, and k is a nonlinear function used to extract the output results from each module.

[0087] In summary, this solution combines the crack identification module based on the deep learning algorithm with multi-period verification and spatiotemporal correlation analysis, effectively reducing the false alarm rate, capturing early cracks, and ensuring timely discovery of potential risks. At the same time, this solution dynamically adjusts the detection threshold according to environmental parameters, reduces false alarms caused by environmental interference, and improves the accuracy of the system under different weather conditions, which not only improves the safety and maintenance efficiency of the dam, but also reduces operating costs, with significant economic and social benefits.

[0088] The above are only embodiments of the present invention. Common knowledge such as the known specific structures and characteristics in the scheme is not described in detail here. Ordinary technicians in the relevant field know all the common technical knowledge in the technical field to which the invention belongs before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply conventional experimental means before that date. Ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A concrete panel dam surface crack identification system based on time-space association, comprising a data acquisition terminal and a server; the data acquisition terminal comprises an image acquisition module and a communication module, the image acquisition module is used to acquire panel images, and the communication module is used to establish a wireless connection between the data acquisition terminal and the server; characterized in that: The server includes a crack identification module, a time series modeling module, a change detection module, a multi-period verification module, an environmental adaptation module and a decision-making module; The crack recognition module is used to identify the crack area on the panel and extract the crack features in the image; The time series modeling module is used to build a time series model for each monitoring point and record the state changes of the panel in different periods of time; The change detection module quantifies the degree of change of the cracks by comparing the differences between the current image and the historical image: In the formula, C t is the degree of crack change, I t (i) and I t-1 (i) represents the grayscale value of the i-th pixel in the current image and the historical image respectively; N is the total number of pixels in the image; by calculating the pixel difference between the two images, a difference map is generated, and C t To quantify the degree of crack change; The multi-period detection module is used to continuously monitor the panel area where the crack change degree reaches the preset value T for multiple periods, and conduct spatiotemporal correlation analysis based on the monitoring results of multiple periods: Y t =g(C t ,C t-1 ,…,C t-n ) Where Y t is the crack state at the tth moment, C t represents the change at the tth moment, g is a nonlinear function used to capture the change trend of the crack. t Conduct analysis to determine the expansion trend of cracks; The environmental adaptation module is used to dynamically adjust the preset value T according to the environmental parameters to improve the detection accuracy in different environments. The mapping relationship between the environmental parameters and the preset values ​​is: T=h(E) In the formula, E is the environmental parameter, and h is a nonlinear function; The decision module is used to determine whether to perform maintenance based on the output results of the change detection module, the multi-period detection module, and the environmental adaptation module: D=k(T,C t ,E) Where D is the final judgment result, and k is a nonlinear function used to extract the output results from each module.

2. The concrete panel dam surface crack identification system based on time-space correlation according to claim 1 is characterized by: The image acquisition module is provided with image quality control rules: Where R is the required ground sampling distance, w is the camera sensor width, f is the camera focal length, and the camera parameters are adjusted according to the image acquisition height of the data acquisition terminal and its camera focal length so that the image resolution meets the requirements. H is the flight altitude required for the UAV to fly.

3. The concrete panel dam surface crack identification system based on time-space correlation according to claim 2 is characterized by: The crack identification module includes an image preprocessing module and a crack feature extraction module; The image preprocessing module is used to preprocess the collected images; The crack feature extraction module is based on an image recognition algorithm and uses an edge detection algorithm to identify crack features, locate the identified cracks and measure their length, average width and maximum width.

4. The concrete panel dam surface crack identification system based on time-space correlation according to claim 3 is characterized by: The image preprocessing module uses Gaussian filtering to remove noise and linear stretching to enhance contrast: Where I(X,Y) is the original image, X and Y are the coordinate positions of the pixels; σ is the standard deviation, which controls the smoothness of the filter; I′(X,Y) is the enhanced image, min(I) and max(I) are the minimum and maximum grayscale values ​​of the image, and L is the number of grayscale levels.

5. The concrete panel dam surface crack identification system based on time-space correlation according to claim 4 is characterized by: The crack feature extraction module includes a coordinate positioning module and a feature calculation module The coordinate positioning module is used to determine the coordinate information of the crack location based on geographic information system technology: In the formula, G x and G y is the gradient of the image in the x and y directions; G is the gradient magnitude; θ is the gradient direction; The feature calculation module is used to measure the length, average width and maximum width of the crack: Where Lc is the length of the crack; for each point (x i ,y i ), find the points (x) on both sides of it i1 ,y i1 ) and (x i2 ,y i2 ), calculate the vertical distance and get the width of the points on both sides: w max =max(w1,w2,…,w n ) In the formula, (x i ,y i ) is the coordinate of the i-th edge point, w i is the width of the i-th node; w avg Average crack width; w max Maximum crack width; crack direction vector v = (v x ,v y ), v x is the component of the crack in the x-axis direction; v y is the component of the crack in the y-axis direction.

6. The concrete panel dam surface crack identification system based on time-space correlation according to claim 5 is characterized by: The time series modeling module is used to control the data acquisition terminal to regularly collect image data of the concrete panel dam panel, and synchronously record the timestamp and environmental parameters, and construct a complete time series data set and a time series model based on the data collected each time: Based on historical image data, a time series model of each monitoring point is constructed using a long short-term memory network (LSTM) to record the state changes of the panel at different time periods: h t =LSTM(X t-1 ,h t-1 ) In the formula, h t is the hidden state of LSTM, indicating the state of the panel at time t; X t-1 is the external input at time t-1, h t-1 is the hidden state at time t-1.

7. The concrete panel dam surface crack identification system based on time-space correlation according to claim 6 is characterized by: The data acquisition terminal uses a drone carrying an image acquisition device to acquire panel images.

8. The concrete panel dam surface crack identification system based on time-space correlation according to claim 7 is characterized by: The environmental parameters include temperature, humidity and light.

Images