An intelligent detection method for bridge pile pier defects in turbid waters based on fluorescence angiography

Abstract

The present invention discloses an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography, comprising obtaining an acoustic wave reflection signal to determine whether the bridge pile pier has a disease; constructing a prediction model for the injection parameters of a fluorescent injection device to determine the injection parameters of the fluorescent injection device; obtaining a fluorescence distribution image on the surface of the bridge pile pier, and extracting the regional contour of the diseased area in the fluorescence distribution image; constructing a disease diagnosis model for the diseased area to determine the disease type and disease severity of the diseased area of ​​the bridge pile pier; constructing a structural risk assessment model for the diseased area to assess the structural risk level of the diseased area; calculating the development rate of the diseased area and generating a preventive maintenance strategy; and continuously adjusting the warning threshold and monitoring strategy according to the maintenance results of the bridge pile pier. The present invention realizes efficient detection and maintenance optimization of bridge pile pier diseases by combining nondestructive testing technology, fluorescent injection technology and intelligent analysis.

Description

Technical Field

[0001] The invention relates to the field of bridge structure detection, and in particular to an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography. Background Art

[0002] At present, the health detection of underwater bridge structures is an important part of bridge safety monitoring, especially the damage detection of bridge piles and piers in underwater environments, which is crucial to the safety of the overall structure of the bridge. This includes the use of fluorescence imaging to detect bridge piles and piers. However, factors such as the complexity of the underwater environment and the optical interference of turbid waters have brought many challenges to the existing bridge pile and pier disease detection methods. First, the complexity of the underwater environment seriously limits the effectiveness of conventional detection methods. In a clear water environment, traditional optical cameras and acoustic wave detection equipment can obtain structural surface information more accurately. However, in waters with poor water quality and high turbidity, the acquisition effect of optical images is greatly reduced. Secondly, underwater detection equipment, such as sonar and underwater cameras, is limited by environmental factors such as underwater flow velocity, water temperature and water quality, and it is difficult to operate stably for a long time, and there are errors and blind spots in the detection results. In addition, the current detection of bridge pile pier diseases mostly relies on post-analysis, that is, detection and repair measures are only carried out when the disease has occurred or expanded to a certain extent. This passive maintenance strategy not only has a lag, but also leads to further deterioration of the damage of the bridge pile pier structure, increasing the complexity and cost of subsequent maintenance. At the same time, hidden diseases of bridge piles and piers, such as deep cracks and internal material peeling, are often difficult to detect in the early stage, which brings potential risks to the long-term safety of bridge piles and piers. In addition, environmental factors have a significant impact on the adhesion effect of fluorescent agents. The adhesion efficiency of fluorescent agents under adverse underwater conditions is low, resulting in poor quality of fluorescent imaging, which increases the difficulty of subsequent contour extraction and diagnosis of diseased areas. Finally, the existing bridge pile and pier disease detection systems are mostly single detection methods, lacking comprehensive disease assessment and risk prediction functions. When facing multiple disease clustering areas, it is impossible to effectively evaluate the expansion trend of the disease, the degree of material damage, and the interaction between diseases. This single detection method is difficult to provide comprehensive data support for the long-term maintenance and management of bridges. In particular, the early warning mechanism for serious disease areas is still imperfect, and it is difficult to achieve timely preventive maintenance. It often requires manual judgment based on a large amount of historical detection data, which is cumbersome and inefficient. Therefore, in response to the above problems, there is an urgent need for a bridge pile and pier disease detection method that can adapt to complex underwater environments. Summary of the invention

[0003] The present invention solves the problems existing in the above-mentioned prior art and provides an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography, which mainly includes:

[0004] According to the sound wave propagation speed of different materials and the thickness of bridge piles and piers, the sound wave emission frequency and sound wave emission power of the non-destructive testing equipment are determined. The non-destructive testing equipment carried by the underwater robot emits sound waves and obtains the sound wave reflection signal to determine whether the bridge piles and piers are damaged;

[0005] The historical injection parameters of the fluorescent injection device, the underwater environmental parameters at the corresponding detection time, and the adhesion effect of the fluorescent agent are obtained through the historical detection records of the bridge pile piers, and a prediction model for the injection parameters of the fluorescent injection device is constructed to determine the injection parameters of the fluorescent injection device with the best adhesion effect of the fluorescent agent under the current underwater environmental parameters.

[0006] Use an underwater robot equipped with a fluorescent spray device to spray fluorescent agent on the bridge piles and piers with defects, obtain the fluorescence distribution image of the bridge pile surface, and extract the regional contour of the defective area in the fluorescence distribution image;

[0007] Obtain regional characteristic data of the diseased area, build a disease diagnosis model for the diseased area, determine the disease type and severity of the diseased area of ​​the bridge pile pier, and formulate maintenance strategies for diseased areas with high disease severity;

[0008] According to the number and area of ​​the diseased areas of bridge piles and piers, determine whether it is a disease clustering area, and build a structural risk assessment model for disease clustering areas based on the disease characteristic data of historical disease clustering areas to assess the structural risk level of disease clustering areas;

[0009] Through the fluorescence distribution images of historical diseased areas, the area change rate and depth development rate of the diseased areas are obtained, the development rate of the diseased areas is calculated, and preventive maintenance activities are carried out for the diseased areas whose development rate exceeds the preset rate threshold;

[0010] According to the maintenance results of bridge piles and piers, the maintenance effect of bridge piles and piers is judged, and based on the maintenance effect, the early warning threshold and monitoring strategy are continuously adjusted.

[0011] Furthermore, the method determines the sound wave transmission frequency and the sound wave transmission power of the nondestructive testing equipment according to the sound wave propagation speed of different materials and the thickness of the bridge pile pier, transmits the sound wave through the nondestructive testing equipment carried by the underwater robot and obtains the sound wave reflection signal to judge whether the bridge pile pier has a disease, including:

[0012] Through the bridge design drawings, the material composition of the bridge piers and the sound wave propagation speed of each material are obtained, including steel, concrete, and paint; according to the sound wave propagation speed of different materials and the thickness of the bridge piers, the formula is used , determine the sound wave emission frequency of the non-destructive testing equipment; according to the distance of sound wave propagation and the absorption characteristics of the material, based on the sound wave attenuation formula , determine the acoustic wave transmission power of the non-destructive testing equipment, where, is the received sound wave power, is the transmission power, is the attenuation coefficient of the material, which is obtained through the relevant physical material database, and d is the sound wave propagation distance; the sound wave is emitted by the non-destructive testing equipment carried by the underwater robot and the sound wave reflection signal is obtained in real time to obtain the sound wave propagation data in the bridge piles and piers, including the sound wave propagation speed, reflection time, reflection intensity and attenuation rate; the historical sound wave propagation data is obtained through the historical inspection records of the bridge piles and piers, and the historical sound wave propagation data are marked to determine whether the bridge piles and piers have defects, the support vector machine algorithm is used for model training, and a bridge pile pier defect judgment model is constructed; according to the real-time acquired sound wave propagation data in the bridge piles and piers, the bridge pile pier defect judgment model is used to determine whether the bridge piles and piers have defects.

[0013] Furthermore, the method of obtaining historical injection parameters of the fluorescent injection device, underwater environmental parameters corresponding to the detection time, and fluorescent agent adhesion effect through the historical detection records of bridge piles and piers, constructing an injection parameter prediction model for the fluorescent injection device, and determining the injection parameters of the fluorescent injection device with the best fluorescent agent adhesion effect under the current underwater environmental parameters includes:

[0014] If it is determined that the bridge pile pier has defects, the historical injection parameters of the fluorescent injection device, the underwater environmental parameters at the corresponding detection time, and the fluorescent agent adhesion effect are obtained through the historical inspection records of the bridge pile pier. The injection parameters include the fluorescent agent injection angle, pressure and speed, the underwater environmental parameters include water temperature, turbidity and flow rate, and the fluorescent agent adhesion effects include excellent, good and poor. According to the historical injection parameters, the underwater environmental parameters at the corresponding detection time and the fluorescent agent adhesion effect, a recurrent neural network is used for model training to construct a fluorescent injection device injection parameter prediction model. The underwater environmental parameters are monitored in real time through the multifunctional underwater sensor carried by the underwater robot, and the injection parameter prediction model of the fluorescent injection device is used to determine the injection parameters of the fluorescent injection device with an excellent fluorescent agent adhesion effect under the current underwater environmental parameters, and the obtained injection parameters are set to the fluorescent injection device.

[0015] Furthermore, the method of using an underwater robot equipped with a fluorescent spraying device to spray a fluorescent agent on a bridge pile with a defect, obtaining a fluorescent distribution image on the surface of the bridge pile, and extracting a regional contour of the defective area in the fluorescent distribution image includes:

[0016] An underwater robot is equipped with a fluorescent spray device, and fluorescent agent is sprayed on the damaged bridge piles and piers according to the set spray parameters; an optical camera is used to capture the fluorescence distribution image on the surface of the bridge piles and piers, and the fluorescence distribution is preprocessed. The preprocessing includes eliminating the noise caused by suspended particles through a median filter, adjusting the underwater illumination unevenness by an adaptive histogram equalization method, and enhancing the edge features of the fluorescence distribution by a Laplacian operator; a fluorescence intensity threshold standard is set according to the fluorescence intensity value in the processed fluorescence image, and the image is divided into a diseased area and a non-disease area by a threshold segmentation method, and the fluorescence intensity is the grayscale value of the image; the Canny edge detection algorithm is used to extract the regional contour of the diseased area in the fluorescence distribution image.

[0017] Furthermore, the acquisition of regional characteristic data of the damaged area, and the construction of a damage diagnosis model for the damaged area, determine the damage type and severity of the damaged area of ​​the bridge pile pier, and formulate a maintenance strategy for the damaged area with high severity, including:

[0018] The area of ​​the diseased area is determined by calculating the number of pixels within the area contour of the diseased area; the morphological coefficient calculation formula is used , calculate the morphological coefficient of the diseased area, where is the morphological coefficient of the area, L is the perimeter of the diseased area, is the minimum circumscribed rectangular area surrounding the diseased area, and A is the actual area of ​​the diseased area. The location of the diseased area is determined by calculating the center position of the diseased area. The fluorescence intensity change rate in the diseased area is calculated using the Sobel operator to obtain the fluorescence intensity gradient distribution of the diseased area, and the gradient characteristics of the diseased area are determined, including the average fluorescence intensity, the maximum fluorescence intensity, and the standard deviation of the gradient direction. The area, regional morphological coefficient, position, and fluorescence intensity gradient characteristics of the diseased area are standardized and combined into a feature vector to form the regional feature data of the diseased area. The bridge pile pier historical detection records are used to obtain the bridge The regional characteristic data of the historical disease areas of beams, piles and piers are obtained, and the corresponding disease type labels and disease severity labels are annotated for the regional characteristic data of the disease areas. The random forest algorithm is used for model training to build a disease diagnosis model for the diseased areas. The disease types include but are not limited to cracks, spalling and exposed reinforcement, and the disease severity includes high, medium and low. The regional characteristic data of the diseased areas are obtained in real time through optical cameras, and the disease diagnosis model of the trained diseased areas is used to determine the disease type and disease severity of the diseased areas of bridge piles and piers, and formulate and implement bridge pile and pier maintenance strategies for diseased areas with high disease severity.

[0019] Furthermore, judging whether it is a disease clustering area according to the number and area of ​​the diseased areas of the bridge pile piers, and building a structural risk assessment model for the disease clustering area according to the disease characteristic data of the historical disease clustering area, and assessing the structural risk degree of the disease clustering area, includes:

[0020] If the number of diseased areas on the bridge pile pier is greater than the preset number threshold, then the ratio of the area of ​​the diseased area to the total area of ​​the pile pier is used to determine whether the bridge pile pier is a disease clustering area; if the ratio is greater than the preset ratio threshold, then the bridge pile pier is determined to be a disease clustering area, and the position and area of ​​each diseased area in the disease clustering area are obtained; through the historical detection records of the bridge pile pier, the disease characteristic data of the historical disease clustering area is obtained, including the number of diseases, the degree of material damage and the interaction index data, and the decision tree algorithm is used for model training to construct a structural risk assessment model for the disease clustering area; through the multifunctional underwater sensor carried by the underwater robot, the disease characteristic data of the disease clustering area is obtained in real time, and the structural risk assessment model of the disease clustering area is used to assess the structural risk level of the disease clustering area; according to the risk assessment results, the corresponding risk warning threshold is set to monitor the structural health status of the disease clustering area in real time; if the structural risk level exceeds the preset first risk level threshold, the warning signal is activated, and the maintenance strategy for the disease clustering area is dispatched, and the maintenance strategy includes local reinforcement, component replacement, increased monitoring frequency or expanded monitoring area.

[0021] It also includes obtaining material damage degree and interaction index data in historical disease cluster areas through historical inspection records of bridge piles and piers.

[0022] The material damage degree and interaction index data of the historical disease accumulation area are obtained through the historical inspection records of bridge piles and piers, specifically including:

[0023] Through the historical inspection records of bridge piles and piers, the location, type and severity of the diseased areas in the historical disease clustering areas are obtained; the location of each diseased area is mapped to the corresponding disease clustering area; the entire disease clustering area is divided into several grid units, each grid is used as an independent monitoring unit, the number and total area of ​​the diseased areas in each unit are counted, and the number and total area data of the diseased areas in the grid are obtained; according to the number, area and type of the diseased areas, the material damage degree assessment formula of the disease clustering area is used , calculate the degree of material damage in the disease-concentrated area, where: is the degree of material damage in the disease-concentrated area, is the area of ​​the diseased area i, is the severity factor of disease area i, which is defined based on historical experience and historical data, and N is the total number of disease areas. Based on the distance between disease areas in the disease cluster area, the severity of the disease and the type of disease, the disease interaction index formula is used. , calculate the disease interaction index, where is the interaction index of each pair of diseased areas i and j, is the distance between diseased area i and diseased area j, is the difference in severity factor between diseased area i and diseased area j, is the distance attenuation coefficient, which is used to adjust the effect of distance on interaction and is adjusted according to the distribution characteristics of the disease area. is the disease type correlation coefficient. The interaction intensity is adjusted according to the difference in disease types between disease area i and disease area j. Different disease types have different correlation coefficients, which are obtained by fitting the historical data of the disease. Based on the disease type interaction index, the disease cluster area interaction index formula is used. , calculate the disease interaction index of the disease clustering area, where TUU is the disease interaction index of the disease clustering area.

[0024] Furthermore, the fluorescent distribution image of the historical diseased area is used to obtain the area change rate of the diseased area and the depth development rate of the diseased area, calculate the development rate of the diseased area, and perform preventive maintenance activities on the diseased area whose development rate exceeds a preset rate threshold, including:

[0025] Through the historical inspection records of bridge piles and piers, the fluorescence distribution image of the historical diseased area is obtained, and according to the area of ​​the same diseased area at different times, the formula is used , calculate the rate of change of the diseased area ,in, and Separately for the moment and The area of ​​the diseased area; according to the average fluorescence intensity of the diseased area at different times, use the formula , calculate the depth development rate of the diseased area ,in, and Separately for the moment and The average fluorescence intensity of the diseased area; according to the rate of change of the diseased area and the rate of development of the depth of the diseased area, the formula is used , calculate the disease area development rate ,in and It is a weight coefficient used to adjust the impact of the expansion rate and depth rate on the comprehensive development of the disease, and is optimized through experiments or historical data. If the development rate of the diseased area exceeds the preset rate threshold, preventive maintenance activities will be carried out on the diseased area, including local structural reinforcement, crack sealing, coating repair, material reinforcement around the diseased area, and the addition of fixed-point monitoring equipment.

[0026] Furthermore, judging the maintenance effect of the bridge piles and piers according to the maintenance results of the bridge piles and piers, and continuously adjusting the warning threshold and monitoring strategy based on the maintenance effect, includes:

[0027] Obtain regional characteristic data of the diseased area before and after the maintenance of bridge piles and piers and disease characteristic data of the diseased clustering area; evaluate the effect of maintenance activities based on the change in the area of ​​the diseased area and the change in the fluorescence intensity gradient, and use data visualization tools to display the development trend of the diseased area before and after maintenance and the changes in the health status of bridge piles and piers; evaluate the effectiveness of the current disease warning strategy and monitoring strategy based on the maintenance effect of bridge piles and piers; if the warning threshold or the effectiveness of the maintenance measures is lower than the preset threshold, adjust the warning threshold and monitoring frequency to obtain an updated warning and monitoring strategy; implement the updated warning and monitoring strategy, and obtain new maintenance effect data to continuously optimize the maintenance, warning and monitoring strategies of bridge piles and piers. The present invention achieves efficient detection and maintenance optimization of bridge pile pier diseases by combining non-destructive testing technology, fluorescent injection technology and intelligent analysis.

[0028] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0029] The present invention provides an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography. By optimizing the emission frequency and power of non-destructive testing equipment according to the sound wave propagation speed of different materials and the thickness of bridge pile piers, it can accurately determine whether the bridge pile piers have diseases in complex underwater environments. The present invention optimizes the angle, pressure and speed of fluorescent agent injection by combining historical detection records with real-time underwater environmental parameters, greatly improves the adhesion effect of fluorescent agent, and can obtain clear disease images even in turbid waters. After obtaining the image of the diseased area, the contour of the diseased area can be accurately extracted, and the type and severity of the disease can be quickly determined according to the area and depth characteristics of the disease, and targeted maintenance strategies can be formulated, especially the rapid response to serious disease areas, which greatly improves the timeliness of maintenance. Based on the number of diseased areas and historical data, the present invention can evaluate the structural risks of disease-aggregated areas and further improve the safety management level of bridge pile piers. By dynamically tracking the development rate of diseases, potential risks can be discovered in advance, preventive maintenance can be implemented, and finally the early warning threshold and monitoring strategy can be continuously optimized to ensure the long-term safety and stability of bridge pile piers. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography of the present invention;

[0031] Figure 2 It is a schematic diagram of an intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography according to the present invention;

[0032] Figure 3 It is another schematic diagram of the intelligent detection method of bridge pile pier diseases in turbid waters based on fluorescence angiography of the present invention. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] like Figure 1-3 In this embodiment, a method for intelligently detecting bridge pile pier defects in turbid waters based on fluorescence angiography may specifically include:

[0035] Step S101, according to the sound wave propagation speed of different materials and the thickness of the bridge pile pier, determine the sound wave emission frequency and sound wave emission power of the non-destructive testing equipment, emit sound waves through the non-destructive testing equipment carried by the underwater robot and obtain the sound wave reflection signal to judge whether the bridge pile pier has defects.

[0036] Through the bridge design drawings, the material composition of the bridge piers is obtained, and the sound wave propagation speed of each material is obtained, including steel, concrete, and paint. According to the sound wave propagation speed of different materials and the thickness of the bridge piers, the formula is used , determine the sound wave emission frequency of the non-destructive testing equipment. According to the distance of sound wave propagation and the absorption characteristics of the material, based on the sound wave attenuation formula , determine the acoustic wave transmission power of the non-destructive testing equipment, where, is the received sound wave power, is the transmission power, is the attenuation coefficient of the material, obtained through the relevant physical material database, and d is the sound wave propagation distance. The non-destructive testing equipment carried by the underwater robot emits sound waves and obtains the sound wave reflection signal in real time to obtain the sound wave propagation data in the bridge pile pier, including the sound wave propagation speed, reflection time, reflection intensity and attenuation rate. The historical sound wave propagation data is obtained through the historical detection records of the bridge pile pier, and the historical sound wave propagation data is marked to indicate whether the bridge pile pier has defects. The support vector machine algorithm is used for model training to build a bridge pile pier defect judgment model. According to the real-time sound wave propagation data in the bridge pile pier, the bridge pile pier defect judgment model is used to judge whether the bridge pile pier has defects.

[0037] For example, the material composition of the bridge pile is obtained from the bridge design drawings, including steel, concrete and paint, and the sound wave propagation speed of steel is 5800m / s, concrete is 4000m / s, and paint is 1500m / s from the physical material database. If the steel layer thickness of the bridge pile is 10cm, the concrete layer thickness is 50cm, and the paint thickness is 2cm, according to the sound wave propagation speed and thickness of these materials, the formula is used The sound wave emission frequency of the non-destructive testing equipment was calculated to ensure that the sound wave could penetrate the pile and reflect back. A transmission frequency of 100kHz was selected. According to the distance d=62cm of sound wave propagation and the absorption characteristics of the material, the sound wave attenuation formula was used. , determine the acoustic wave transmission power of the non-destructive testing equipment, where, is the received sound wave power, is the transmission power, is the attenuation coefficient of the material, obtained from the relevant physical material database, and d is the distance of sound wave propagation. is 0.1W, the attenuation coefficient The weighted average of steel, concrete and paint is 0.0025Np / cm. The transmission power is determined according to the sound wave attenuation formula. The sound wave propagation speed in the pile pier is 5W to ensure a strong enough reflection signal. The non-destructive testing equipment carried by the underwater robot emits sound waves, obtains the sound wave reflection signal in real time, and obtains the sound wave propagation speed in the pile pier, the reflection time is 0.0001 seconds, the reflection intensity is 0.5W, and the attenuation rate is 0.3. Combined with the historical inspection records of bridge piles and piers, it is marked whether there are defects in the past inspection data, and the support vector machine algorithm is used for model training to build a bridge pile pier disease judgment model. By analyzing the real-time acquired sound wave propagation data, it is judged that the current sound wave propagation speed in the pile pier is 5% slower than the speed in the healthy state, and the reflection intensity is reduced by 20%. The model judges that the pile pier has a disease.

[0038] Step S102, obtaining historical spray parameters of the fluorescent spray device, underwater environmental parameters corresponding to the detection time, and fluorescent agent adhesion effect through historical inspection records of bridge piles and piers, constructing a spray parameter prediction model for the fluorescent spray device, and determining the spray parameters of the fluorescent spray device with the best fluorescent agent adhesion effect under the current underwater environmental parameters.

[0039] If it is determined that the bridge pile pier has defects, the historical injection parameters of the fluorescent injection device, the underwater environmental parameters at the corresponding detection time, and the fluorescent agent adhesion effect are obtained through the historical detection records of the bridge pile pier. The injection parameters include the fluorescent agent injection angle, pressure and speed, the underwater environmental parameters include water temperature, turbidity and flow rate, and the fluorescent agent adhesion effect includes excellent, good and poor. According to the historical injection parameters, the underwater environmental parameters at the corresponding detection time, and the fluorescent agent adhesion effect, a recurrent neural network is used for model training to construct a fluorescent injection device injection parameter prediction model. The underwater environmental parameters are monitored in real time through the multifunctional underwater sensor carried by the underwater robot. The fluorescent injection device injection parameter prediction model is used to determine the injection parameters of the fluorescent injection device with an excellent fluorescent agent adhesion effect under the current underwater environmental parameters, and the obtained injection parameters are set to the fluorescent injection device.

[0040] For example, if it is determined that the bridge pile pier has a disease, the historical injection parameters of the fluorescent injection device are obtained through the historical inspection records of the bridge pile pier, including the injection angle of the fluorescent agent in the previous inspection is 30 degrees, the injection pressure is 2.5MPa, and the injection speed is 3m / s. Combined with the underwater environmental parameters at that time, the water temperature is 15°C, the turbidity is 30NTU, the flow rate is 1m / s, and the fluorescent agent adhesion effect is good. Through these historical data, a recurrent neural network is used to train the injection parameters, environmental parameters, and fluorescent agent adhesion effects, and a prediction model for the injection parameters of the fluorescent injection device is constructed. The multifunctional underwater sensor carried by the underwater robot monitors the current underwater environmental parameters in real time, and detects that the current water temperature is 18°C, the turbidity is 50NTU, and the flow rate is 0.8m / s. The injection parameter prediction model of the fluorescent injection device predicts that under these environmental conditions, in order to ensure the best adhesion effect of the fluorescent agent, the injection angle of the fluorescent injection device should be adjusted to 35 degrees, the injection pressure should be set to 3.0MPa, and the injection speed should be 3.5m / s. These injection parameters should be set to the fluorescent injection device to ensure that the fluorescent agent can effectively adhere to the defective area of ​​the bridge pile pier in the current environment.

[0041] Step S103, using an underwater robot equipped with a fluorescent spraying device to spray fluorescent agent on the bridge piles and piers with defects, obtaining a fluorescent distribution image of the bridge pile pier surface, and extracting the regional contour of the defective area in the fluorescent distribution image.

[0042] An underwater robot was equipped with a fluorescent spray device, and the fluorescent agent was sprayed on the damaged bridge piles according to the set spray parameters. An optical camera was used to capture the fluorescence distribution image on the surface of the bridge piles, and the fluorescence distribution was preprocessed. The preprocessing included eliminating the noise caused by suspended particles through a median filter, adjusting the underwater illumination unevenness by an adaptive histogram equalization method, and enhancing the edge features of the fluorescence distribution by the Laplacian operator. According to the fluorescence intensity value in the processed fluorescence image, a threshold standard of the fluorescence intensity was set, and the image was divided into the diseased area and the non-disease area by the threshold segmentation method. The fluorescence intensity was the gray value of the image. The Canny edge detection algorithm was used to extract the regional contour of the diseased area in the fluorescence distribution image.

[0043] Exemplarily, a fluorescent spray device carried by an underwater robot is used to spray fluorescent agent on the bridge piles and piers with defects according to the set spray parameters, with a spray angle of 35 degrees, a spray pressure of 3.0 MPa, and a spray speed of 3.5 m / s. The fluorescence distribution image on the surface of the bridge piles and piers is captured by an optical camera, and the image shows a distribution with higher fluorescence intensity in the defective area. Then, the fluorescence distribution image is preprocessed, and a median filter is used to remove the noise caused by underwater suspended particles to make the image clearer. Adaptive histogram equalization is used to adjust the problem of uneven underwater illumination, so that the brightness of the fluorescence distribution in each area is more uniform, and the Laplacian operator is used to enhance the edge features in the image to highlight the outline of the defective area. According to the processed fluorescence image, the fluorescence intensity value range of the image is measured to be 0 to 255, and the threshold standard is set to 150. The image is segmented into a defective area and a normal area using a threshold segmentation method, and pixels with a fluorescence intensity higher than 150 are marked as defective areas. Then, the Canny edge detection algorithm is applied to extract the contour of the defective area in the fluorescence distribution image, obtain the clear edge of the defective area, and confirm that the area is the defective part of the bridge pile pier.

[0044] Step S104, obtaining regional characteristic data of the damaged area, and constructing a damage diagnosis model for the damaged area, determining the damage type and severity of the damaged area of ​​the bridge pile pier, and formulating a maintenance strategy for the damaged area with high severity.

[0045] The area of ​​the diseased area is determined by counting the number of pixels within the area outline of the diseased area. The morphological coefficient calculation formula is used , calculate the morphological coefficient of the diseased area, where is the morphological coefficient of the area, L is the perimeter of the diseased area, is the minimum circumscribed rectangular area surrounding the diseased area, and A is the actual area of ​​the diseased area. The location of the diseased area is determined by calculating the center position of the diseased area, and the Sobel operator is used to calculate the rate of change of the fluorescence intensity in the diseased area to obtain the fluorescence intensity gradient distribution of the diseased area, and determine the gradient characteristics of the diseased area, including the average fluorescence intensity, the maximum fluorescence intensity, and the standard deviation of the gradient direction. The area, regional morphological coefficient, position, and fluorescence intensity gradient characteristics of the diseased area are standardized and combined into a feature vector to form the regional characteristic data of the diseased area. The regional characteristic data of the historical diseased areas of bridge piles and piers are obtained through the historical inspection records of bridge piles and piers, and the corresponding disease type labels and disease severity labels are annotated for the regional characteristic data of the diseased areas. The random forest algorithm is used for model training to construct a disease diagnosis model for the diseased area. The disease types include but are not limited to cracks, spalling, and exposed reinforcement, and the disease severity includes high, medium, and low. The regional characteristic data of the diseased area is obtained in real time through an optical camera, and the disease diagnosis model of the trained diseased area is used to determine the disease type and severity of the diseased area of ​​the bridge pile pier, and formulate and implement bridge pile pier maintenance strategies for diseased areas with high disease severity.

[0046] For example, the perimeter of a bridge pile pier defect area is determined by calculating the number of pixels within the contour of the defect area. 85 pixels, area The minimum enclosing rectangle area is 500 pixels. = 600 pixels. The morphological coefficient reflects the compactness of the diseased area relative to its circumscribed rectangle. The smaller the value, the more compact the shape of the area is, close to a rectangle. The larger the value, the more complex or irregular the shape is. , the morphological coefficient of the diseased area is calculated to be 0.034, where is the morphological coefficient of the area, L is the perimeter of the diseased area, is the minimum circumscribed rectangular area surrounding the diseased area, and A is the actual area of ​​the diseased area. The location of the diseased area is determined by centroid calculation, and the centroid coordinates are (200,150). The Sobel operator is used to calculate the rate of change of fluorescence intensity in the diseased area, and the gradient characteristics of fluorescence intensity are obtained, where the average fluorescence intensity is 180, the maximum fluorescence intensity is 255, and the standard deviation of the gradient direction is 0.25. Then, the area, morphological coefficient, position and fluorescence intensity gradient characteristics of the diseased area are standardized and combined into a feature vector to form the regional feature data of the diseased area. Through the historical inspection records of bridge piles and piers, the historical feature data of similar diseased areas are obtained. One of the historical diseased areas has an area of ​​450 pixels, a morphological coefficient of 0.017, an average fluorescence intensity of 170, and a maximum value of 240. The disease type is marked as cracks and the disease severity is medium. The random forest algorithm is used to train the historical data to construct a disease diagnosis model for the diseased area. The optical camera is used to obtain the characteristic data of the current bridge pile pier disease area in real time. The trained diagnostic model is used to determine that the disease area is of the spalling type and the disease severity is high. Based on the diagnosis results, a maintenance strategy for the disease area is formulated and implemented. The maintenance activities include strengthening the bridge pile pier and repairing the surface of the diseased area to prevent further deterioration.

[0047] Step S105, judging whether it is a disease clustering area according to the number and area of ​​the diseased areas of the bridge piles and piers, and constructing a structural risk assessment model for the disease clustering area according to the disease characteristic data of the historical disease clustering area to assess the structural risk level of the disease clustering area.

[0048] If the number of diseased areas on the bridge pile pier is greater than the preset number threshold, then the ratio of the area of ​​the diseased area to the total area of ​​the pile pier is used to determine whether the bridge pile pier is a disease clustering area. If the ratio is greater than the preset ratio threshold, the bridge pile pier is determined to be a disease clustering area, and the position and area of ​​each diseased area in the disease clustering area are obtained. Through the historical detection records of bridge pile piers, the disease characteristic data of the historical disease clustering area, including the number of diseases, material damage degree and interaction index data, are obtained, and the decision tree algorithm is used for model training to construct a structural risk assessment model for the disease clustering area. Through the multifunctional underwater sensor carried by the underwater robot, the disease characteristic data of the disease clustering area is obtained in real time, and the structural risk assessment model of the disease clustering area is used to assess the structural risk degree of the disease clustering area. According to the risk assessment results, the corresponding risk warning threshold is set to monitor the structural health status of the disease clustering area in real time. If the structural risk degree exceeds the preset first risk level threshold, the warning signal is activated, and the maintenance strategy for the disease clustering area is dispatched. The maintenance strategy includes local reinforcement, component replacement, increased monitoring frequency or expanded monitoring area.

[0049] For example, if the number of diseased areas on a bridge pile pier exceeds a preset threshold of 5, the bridge pile pier is judged to be a disease clustering area, and the ratio of the diseased area to the total area of ​​the pile pier is calculated to be 0.15, and this ratio is greater than the preset ratio threshold of 0.1, so the bridge pile pier is confirmed to be a disease clustering area. In the disease clustering area, the position and area of ​​each diseased area are obtained. If the centroid position of a diseased area is (250, 300) and the area is 600 square millimeters, the position of another diseased area is (400, 500) and the area is 800 square millimeters. Through the historical inspection records of bridge pile piers, historical disease characteristic data of similar disease clustering areas are obtained, including 5 diseases, medium material damage degree, and disease interaction index of 0.75. Using these historical data, the model is trained by the decision tree algorithm to construct a structural risk assessment model for disease clustering areas. The multifunctional underwater sensor carried by the underwater robot can obtain the characteristic data of the current disease cluster area in real time, and the number of diseases detected is 15, the degree of material damage is serious, and the disease interaction index is 0.8. The constructed structural risk assessment model is used to assess the structural risk level of the disease cluster area. The results show that the risk level is high, exceeding the preset first risk level threshold. Based on this result, the early warning signal is activated and the maintenance strategy is dispatched. The maintenance strategy includes local reinforcement of the disease area, replacement of damaged parts, increasing the monitoring frequency from once a week to once a day, and expanding the monitoring area to the adjacent parts of the pile pier to prevent further expansion of the disease.

[0050] Among them, the material damage degree and interaction index data of historical disease concentration areas are obtained through the historical inspection records of bridge piles and piers.

[0051] Through the historical inspection records of bridge piles and piers, the location, type and severity of the diseased areas in the historical disease clustering areas are obtained. The location of each diseased area is mapped to the corresponding disease clustering area. The entire disease clustering area is divided into several grid units, each grid is used as an independent monitoring unit, and the number and total area of ​​the diseased areas in each unit are counted to obtain the number and total area data of the diseased areas in the grid. According to the number, area and type of the diseased areas, the material damage degree assessment formula of the disease clustering area is used. , calculate the degree of material damage in the disease-concentrated area, where: is the degree of material damage in the disease-concentrated area, is the area of ​​the diseased area i, is the severity factor of disease area i, which is defined based on historical experience and historical data, and N is the total number of disease areas. Based on the distance between disease areas in the disease cluster area, the severity of the disease and the type of disease, the disease interaction index formula is used. , calculate the disease interaction index, where is the interaction index of each pair of diseased areas i and j, is the distance between diseased area i and diseased area j, is the difference in severity factor between diseased area i and diseased area j, is the distance attenuation coefficient, which is used to adjust the effect of distance on interaction and is adjusted according to the distribution characteristics of the disease area. is the disease type correlation coefficient. The interaction strength is adjusted according to the difference in disease types between disease area i and disease area j. Different disease types have different correlation coefficients, which are obtained by fitting the historical data of the disease. Based on the disease type interaction index, the disease cluster area interaction index formula is used. , calculate the disease interaction index of the disease clustering area, where TUU is the disease interaction index of the disease clustering area.

[0052] For example, if the historical defect area location, defect type and defect severity of a defect cluster area are obtained through the historical inspection records of bridge piles and piers, it is assumed that there are 5 defect areas in the defect cluster area, which are located in different grid units. The number and total area of ​​the defect areas in each grid unit have been counted. There are 2 defect areas in grid 1 with a total area of ​​100 square millimeters, and there is 1 defect area in grid 2 with a total area of ​​50 square millimeters. The remaining grids also have corresponding statistical data. According to the number, area and defect type of these defect areas, the material damage degree assessment formula of the defect cluster area is used. Calculate the degree of material damage in the disease-concentrated area, where: is the degree of material damage in the disease-concentrated area, is the area of ​​the diseased area i, is the severity factor of the damaged area i. If the total number of damaged areas N is 5, the area of ​​damaged area 1 is 100 square millimeters, different types of damage have different effects on the structure, so the severity factors are different. The severity factor w is 0.8, the area of ​​damaged area 2 is 50 square millimeters, and the severity factor is 0.9. The degree of material damage is calculated. = 72. Based on the distance between diseased areas, disease severity and type, the disease interaction index formula is used. , calculate the disease interaction index, where is the interaction index of each pair of diseased areas i and j, is the distance between diseased area i and diseased area j, is the difference in severity factor between diseased area i and diseased area j, is the distance attenuation coefficient, which is used to adjust the effect of distance on interaction and is adjusted according to the distribution characteristics of the disease area. is the disease type correlation coefficient. The interaction strength is adjusted according to the difference in disease types between disease area i and disease area j. Different disease types have different correlation coefficients, which are obtained by fitting the historical data of the disease. If the distance between disease area 1 and disease area 2 is 15 mm, the severity factor of disease area 1 is 0.3, and the severity factor of disease area 2 is 0.2, then the severity factor difference w 12 is 0.1, the distance attenuation coefficient =10, disease type correlation coefficient T 12 = 0.7, substituting these values ​​into the formula, the interaction index between diseased areas 1 and 2 is calculated as UU 12 =0.33. Based on the interaction index between all disease areas, the disease cluster area interaction index formula is used. The overall disease interaction index TUU of the disease clustering area was calculated to be 1.3, and the structural health status of the disease clustering area was assessed to be poor, and further maintenance measures were judged to be needed.

[0053] Step S106, obtaining the area change rate and depth development rate of the diseased area through the fluorescence distribution image of the historical diseased area, calculating the development rate of the diseased area, and performing preventive maintenance activities on the diseased area whose development rate exceeds the preset rate threshold.

[0054] Through the historical inspection records of bridge piles and piers, the fluorescence distribution image of the historical diseased area is obtained, and according to the area of ​​the same diseased area at different times, the formula is used , calculate the rate of change of the diseased area ,in, and Separately for the moment and According to the average fluorescence intensity of the diseased area at different times, the formula , calculate the depth development rate of the diseased area ,in, and Separately for the moment and The average fluorescence intensity of the diseased area is calculated using the formula , calculate the disease area development rate ,in and It is a weight coefficient used to adjust the impact of the expansion rate and depth rate on the comprehensive development of the disease, and is optimized through experiments or historical data. If the development rate of the diseased area exceeds the preset rate threshold, preventive maintenance activities will be carried out on the diseased area, including local structural reinforcement, crack sealing, coating repair, material reinforcement around the diseased area, and the addition of fixed-point monitoring equipment.

[0055] For example, through the historical inspection records of bridge piles and piers, the fluorescence distribution images of the same diseased area at different times are obtained. = 0 days, the area of ​​the diseased area = 100 square millimeters, and at the moment = Area of ​​diseased area at 30 days =150 square millimeters. According to the formula Calculate the change rate of the diseased area =1.67 square millimeters / day. The average fluorescence intensity of the diseased area =180, at time The mean fluorescence intensity =210. According to the formula , calculate the depth development rate of the diseased area = 1 intensity unit / day. According to the area change rate and depth development rate of the diseased area, use the formula , where the weight coefficients τ=0.6 and ν=0.4 are set, and the comprehensive development rate of the diseased area is calculated = 1.4 units / day. If the development rate of the diseased area 1.4 exceeds the preset rate threshold 1, preventive maintenance activities will be carried out on the diseased area, including local structural reinforcement, crack sealing, coating repair, and additional material reinforcement around the diseased area, and additional set-point monitoring equipment to further track the development of the disease and ensure structural safety.

[0056] Step S107, judging the maintenance effect of the bridge piles and piers according to the maintenance results of the bridge piles and piers, and continuously adjusting the warning threshold and the monitoring strategy based on the maintenance effect.

[0057] Obtain the regional characteristic data of the diseased areas before and after the maintenance of bridge piles and piers, and the disease characteristic data of the diseased cluster areas. Evaluate the effectiveness of maintenance activities based on the changes in the area of ​​the diseased areas and the changes in the fluorescence intensity gradient, and use data visualization tools to display the development trend of the diseased areas before and after maintenance and the changes in the health status of bridge piles and piers. Evaluate the effectiveness of the current disease early warning strategy and monitoring strategy based on the maintenance effect of bridge piles and piers. If the early warning threshold or the effectiveness of the maintenance measures is lower than the preset threshold, adjust the early warning threshold and monitoring frequency to obtain an updated early warning and monitoring strategy. Implement the updated early warning and monitoring strategy, obtain new maintenance effect data, and continuously optimize the maintenance, early warning, and monitoring strategies of bridge piles and piers.

[0058] For example, by obtaining the regional characteristic data of the diseased area before and after the maintenance of the bridge pile pier, it is assumed that the area of ​​the diseased area before maintenance is 200 square millimeters, the fluorescence intensity gradient is 3.5, and the area of ​​the diseased area after maintenance is reduced to 150 square millimeters, and the fluorescence intensity gradient decreases to 2.0. In addition, the disease characteristic data of the disease clustering area is obtained, showing that the number of diseases has been reduced from 8 to 6, and the degree of material damage has also improved. According to the change in the area of ​​the diseased area and the decrease in the fluorescence intensity gradient, the effect of the maintenance activity is evaluated as significantly improved, and the development trend chart of the diseased area is displayed using data visualization tools such as Tableau, showing a significant slowdown in the spread of the disease and an improvement in the overall health of the bridge pile pier. According to the maintenance effect of the bridge pile pier, the current disease warning strategy and monitoring strategy are evaluated, and it is found that the warning threshold is set at 200 square millimeters of the diseased area, which is too high and causes the warning to be not timely enough, and the monitoring frequency of once a month is also too sparse. After the maintenance effect evaluation, if it is found that the warning threshold and the effectiveness of the maintenance measures are lower than the preset threshold requirements, the warning threshold will be adjusted to 150 square millimeters, and the monitoring frequency will be increased to once every two weeks. By implementing the updated early warning and monitoring strategies and acquiring new maintenance effect data, it was found that the area of ​​the diseased area was further reduced to 120 square millimeters, and the fluorescence intensity gradient was stabilized at 1.8. The maintenance, early warning and monitoring strategies of bridge piles and piers were continuously optimized to ensure the long-term stability and safety of bridge piles and piers.

[0059] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. An intelligent detection method for bridge pile pier diseases in turbid waters based on fluorescence angiography, characterized in that: The method comprises: According to the sound wave propagation speed of different materials and the thickness of bridge piles and piers, the sound wave emission frequency and sound wave emission power of the non-destructive testing equipment are determined. The non-destructive testing equipment carried by the underwater robot emits sound waves and obtains the sound wave reflection signal to determine whether the bridge piles and piers are damaged; The historical injection parameters of the fluorescent injection device, the underwater environmental parameters at the corresponding detection time, and the adhesion effect of the fluorescent agent are obtained through the historical detection records of the bridge pile piers, and a prediction model for the injection parameters of the fluorescent injection device is constructed to determine the injection parameters of the fluorescent injection device with the best adhesion effect of the fluorescent agent under the current underwater environmental parameters. Use an underwater robot equipped with a fluorescent spray device to spray fluorescent agent on the bridge piles and piers with defects, obtain the fluorescence distribution image of the bridge pile surface, and extract the regional contour of the defective area in the fluorescence distribution image; Obtain regional characteristic data of the diseased area, build a disease diagnosis model for the diseased area, determine the disease type and severity of the diseased area of ​​the bridge pile pier, and formulate maintenance strategies for diseased areas with high disease severity; According to the number and area of ​​the diseased areas of bridge piles and piers, determine whether it is a disease clustering area, and build a structural risk assessment model for disease clustering areas based on the disease characteristic data of historical disease clustering areas to assess the structural risk level of disease clustering areas; Through the fluorescence distribution images of historical diseased areas, the area change rate and depth development rate of the diseased areas are obtained, the development rate of the diseased areas is calculated, and preventive maintenance activities are carried out for the diseased areas whose development rate exceeds the preset rate threshold; According to the maintenance results of bridge piles and piers, the maintenance effect of bridge piles and piers is judged, and based on the maintenance effect, the warning threshold and monitoring strategy are continuously adjusted; The method of obtaining historical injection parameters of the fluorescent injection device, underwater environmental parameters corresponding to the detection time, and fluorescent agent adhesion effect through historical detection records of bridge piles and piers, constructing a prediction model for injection parameters of the fluorescent injection device, and determining the injection parameters of the fluorescent injection device with the best fluorescent agent adhesion effect under the current underwater environmental parameters includes: If it is determined that the bridge pile pier has defects, the historical injection parameters of the fluorescent injection device, the underwater environmental parameters at the corresponding detection time, and the fluorescent agent adhesion effect are obtained through the historical detection records of the bridge pile pier. The injection parameters include the injection angle, pressure and speed of the fluorescent agent, the underwater environmental parameters include water temperature, turbidity and flow rate, and the fluorescent agent adhesion effects include excellent, good and poor. According to the historical injection parameters, the underwater environmental parameters at the corresponding detection time and the fluorescent agent adhesion effect, a recurrent neural network is used for model training to construct a fluorescent injection device injection parameter prediction model. The underwater environmental parameters are monitored in real time through the multifunctional underwater sensor carried by the underwater robot, and the injection parameter prediction model of the fluorescent injection device is used to determine the injection parameters of the fluorescent injection device with an excellent fluorescent agent adhesion effect under the current underwater environmental parameters, and the obtained injection parameters are set to the fluorescent injection device; The method of obtaining regional characteristic data of the damaged area, constructing a damage diagnosis model for the damaged area, determining the damage type and severity of the damaged area of ​​the bridge pile pier, and formulating a maintenance strategy for the damaged area with high severity of the damage includes: The area of ​​the diseased area is determined by calculating the number of pixels within the area contour of the diseased area; the morphological coefficient calculation formula is used , calculate the morphological coefficient of the diseased area, where is the morphological coefficient of the area, L is the perimeter of the diseased area, is the minimum circumscribed rectangular area surrounding the diseased area, and A is the actual area of ​​the diseased area. The location of the diseased area is determined by calculating the center position of the diseased area. The fluorescence intensity change rate in the diseased area is calculated using the Sobel operator to obtain the fluorescence intensity gradient distribution of the diseased area, and the gradient characteristics of the diseased area are determined, including the average fluorescence intensity, the maximum fluorescence intensity, and the standard deviation of the gradient direction. The area, regional morphological coefficient, position, and fluorescence intensity gradient characteristics of the diseased area are standardized and combined into a feature vector to form the regional feature data of the diseased area. The bridge pile pier historical detection records are used to obtain the bridge The regional characteristic data of the historical disease areas of beams, piles and piers are obtained, and the corresponding disease type labels and disease severity labels are annotated for the regional characteristic data of the disease areas. The random forest algorithm is used for model training to build a disease diagnosis model for the diseased areas. The disease types include but are not limited to cracks, spalling and exposed reinforcement, and the disease severity includes high, medium and low. The regional characteristic data of the diseased areas are obtained in real time through optical cameras, and the disease diagnosis model of the trained diseased areas is used to determine the disease type and disease severity of the diseased areas of bridge piles and piers, and formulate and implement bridge pile and pier maintenance strategies for diseased areas with high disease severity.

2. The method according to claim 1, wherein: The method determines the sound wave transmission frequency and the sound wave transmission power of the non-destructive testing equipment according to the sound wave propagation speed of different materials and the thickness of the bridge pile pier, transmits the sound wave through the non-destructive testing equipment carried by the underwater robot and obtains the sound wave reflection signal to judge whether the bridge pile pier has a disease, including: Through the bridge design drawings, the material composition of the bridge piers and the sound wave propagation speed of each material are obtained, including steel, concrete, and paint; according to the sound wave propagation speed of different materials and the thickness of the bridge piers, the formula is used , determine the sound wave emission frequency of the non-destructive testing equipment; according to the distance of sound wave propagation and the absorption characteristics of the material, based on the sound wave attenuation formula , determine the acoustic wave transmission power of the non-destructive testing equipment, where, is the received sound wave power, is the transmission power, is the attenuation coefficient of the material, which is obtained through the relevant physical material database, and d is the sound wave propagation distance; the sound wave is emitted by the non-destructive testing equipment carried by the underwater robot and the sound wave reflection signal is obtained in real time to obtain the sound wave propagation data in the bridge piles and piers, including the sound wave propagation speed, reflection time, reflection intensity and attenuation rate; the historical sound wave propagation data is obtained through the historical inspection records of the bridge piles and piers, and the historical sound wave propagation data are marked to determine whether the bridge piles and piers have defects, the support vector machine algorithm is used for model training, and a bridge pile pier defect judgment model is constructed; according to the real-time acquired sound wave propagation data in the bridge piles and piers, the bridge pile pier defect judgment model is used to determine whether the bridge piles and piers have defects.

3. The method according to claim 1, wherein: The method comprises: using an underwater robot equipped with a fluorescent spraying device to spray a fluorescent agent on a bridge pile with a defect, obtaining a fluorescent distribution image on the surface of the bridge pile, and extracting a regional contour of the defective area in the fluorescent distribution image, including: Use an underwater robot equipped with a fluorescent spray device to spray fluorescent agents on damaged bridge piles and piers according to the set spray parameters; An optical camera is used to capture the fluorescence distribution image on the surface of the bridge pile pier, and the fluorescence distribution is preprocessed. The preprocessing includes eliminating the noise caused by suspended particles through a median filter, adjusting the underwater illumination unevenness by an adaptive histogram equalization method, and enhancing the edge features of the fluorescence distribution by the Laplacian operator. According to the fluorescence intensity value in the processed fluorescence image, a threshold standard of the fluorescence intensity is set, and the image is divided into the diseased area and the non-disease area by the threshold segmentation method, and the fluorescence intensity is the gray value of the image. The Canny edge detection algorithm is used to extract the regional contour of the diseased area in the fluorescence distribution image.

4. The method according to claim 1, wherein: The method of judging whether a bridge pile pier is a disease clustering area based on the number and area of ​​the diseased area, and constructing a structural risk assessment model for the disease clustering area based on the disease characteristic data of the historical disease clustering area, and assessing the structural risk degree of the disease clustering area includes: If the number of defective areas on the bridge pile pier is greater than a preset number threshold, then the ratio of the area of ​​the defective area to the total area of ​​the pile pier is used to determine whether the bridge pile pier is a defect clustering area; if the ratio is greater than the preset ratio threshold, then the bridge pile pier is determined to be a defect clustering area, and the position and area of ​​each defective area in the defect clustering area are obtained; through the historical inspection records of the bridge pile pier, the defect characteristic data of the historical defect clustering area is obtained, including the number of defects, the degree of material damage and the interaction index data, and the decision tree algorithm is used for model training to construct a structural risk assessment model for the defect clustering area; through the multifunctional underwater sensor carried by the underwater robot, the defect characteristic data of the defect clustering area is obtained in real time, and the structural risk assessment model of the defect clustering area is used to assess the structural risk degree of the defect clustering area; Set corresponding risk warning thresholds based on risk assessment results to monitor the structural health status of disease-concentrated areas in real time; If the structural risk level exceeds the preset first risk level threshold, an early warning signal will be activated and a maintenance strategy will be dispatched for the disease-concentrated area. The maintenance strategy includes local reinforcement, component replacement, increased monitoring frequency or expanded monitoring area.

5. The method according to claim 4, wherein: The material damage degree and interaction index data of the historical disease accumulation area are obtained through the historical inspection records of bridge piles and piers, including: Through the historical inspection records of bridge piles and piers, the location, type and severity of the disease areas in the historical disease cluster areas can be obtained; Map the location of each diseased area to the corresponding disease cluster area; The entire disease cluster area is divided into several grid units. Each grid is used as an independent monitoring unit. The number and total area of ​​diseased areas in each unit are counted to obtain the number and total area data of diseased areas in the grid. According to the number, area and type of diseased areas, the material damage degree evaluation formula of the disease cluster area is used. , calculate the degree of material damage in the disease-concentrated area, where: is the degree of material damage in the disease-concentrated area, is the area of ​​the diseased area i, is the severity factor of disease area i, which is defined based on historical experience and historical data, and N is the total number of disease areas. Based on the distance between disease areas in the disease cluster area, the severity of the disease and the type of disease, the disease interaction index formula is used. , calculate the disease interaction index, where is the interaction index of each pair of diseased areas i and j, is the distance between diseased area i and diseased area j, is the difference in severity factor between diseased area i and diseased area j, is the distance attenuation coefficient, which is used to adjust the effect of distance on interaction and is adjusted according to the distribution characteristics of the disease area. is the disease type correlation coefficient. The interaction intensity is adjusted according to the difference in disease types between disease area i and disease area j. Different disease types have different correlation coefficients, which are obtained by fitting the historical data of the disease. Based on the disease type interaction index, the disease cluster area interaction index formula is used. , calculate the disease interaction index of the disease clustering area, where TUU is the disease interaction index of the disease clustering area.

6. The method according to claim 1, wherein: The method of obtaining the area change rate and depth development rate of the diseased area through the fluorescence distribution image of the historical diseased area, calculating the development rate of the diseased area, and performing preventive maintenance activities on the diseased area whose development rate exceeds a preset rate threshold includes: Through the historical inspection records of bridge piles and piers, the fluorescence distribution image of the historical diseased area is obtained, and according to the area of ​​the same diseased area at different times, the formula is used , calculate the rate of change of the diseased area ,in, and Separately for the moment and The area of ​​the diseased area; according to the average fluorescence intensity of the diseased area at different times, use the formula , calculate the depth development rate of the diseased area ,in, and Separately for the moment and The average fluorescence intensity of the diseased area; according to the rate of change of the diseased area and the rate of development of the depth of the diseased area, the formula is used , calculate the disease area development rate ,in and It is a weight coefficient used to adjust the impact of the expansion rate and depth rate on the comprehensive development of the disease, and is optimized through experiments or historical data. If the development rate of the diseased area exceeds the preset rate threshold, preventive maintenance activities will be carried out on the diseased area, including local structural reinforcement, crack sealing, coating repair, material reinforcement around the diseased area, and the addition of fixed-point monitoring equipment.

7. The method according to claim 1, wherein: The maintenance effect of the bridge piles and piers is judged according to the maintenance results of the bridge piles and piers, and the early warning threshold and monitoring strategy are continuously adjusted based on the maintenance effect, including: Obtain regional characteristic data of diseased areas before and after maintenance of bridge piles and piers and disease characteristic data of disease-aggregated areas; evaluate the effectiveness of maintenance activities based on changes in the area of ​​diseased areas and changes in fluorescence intensity gradients, and use data visualization tools to display the development trend of diseased areas before and after maintenance and changes in the health status of bridge piles and piers; evaluate the effectiveness of current disease early warning strategies and monitoring strategies based on the maintenance effect of bridge piles and piers; if the early warning threshold or the effectiveness of maintenance measures is lower than the preset threshold, adjust the early warning threshold and monitoring frequency to obtain updated early warning and monitoring strategies; implement updated early warning and monitoring strategies, obtain new maintenance effect data, and continuously optimize the maintenance, early warning, and monitoring strategies of bridge piles and piers.

Images