A rapid identification method for bridge damage

Abstract

The present invention discloses a method for rapid identification of bridge damage based on sub-pixel edge detection and finite element method, comprising: 1. collecting data using a camera; 2. collecting vehicle weight data using a vehicle-mounted weighing device; 3. performing finite element division on the collected road information according to actual conditions; 4. obtaining the displacement value of each point on the bridge deck according to the bridge structure data; 5. analyzing the collected data; 6. determining the position of the vehicle; 7. identifying damage to each part of the bridge; 8. dividing the danger level based on the degree of damage to each part; 9. adopting a suitable plan to repair the bridge; 10. after taking measures, identifying damage to the bridge deck again. The present invention simplifies the complex stress conditions of the bridge based on the finite element method, making it possible to measure the damage to a specific area of ​​the bridge deck, thereby providing auxiliary decision-making for the traffic management department and providing support for promoting the intelligent and scientific management of local urban traffic.

Description

Technical Field

[0001] The invention relates to an elevated bridge, in particular to a method for quickly identifying bridge damage. Background Art

[0002] Urban bridges, as the main traffic arteries of the city, once they begin to show damage and are not repaired, a large number of vehicles passing by will gradually expand the damage to the bridges, causing irreversible loss of life and huge economic losses.

[0003] The traditional method of detecting bridge damage can only be closed for inspection, which wastes a lot of time and cannot monitor the bridge anytime and anywhere. Even if the bridge is detected to be damaged, the specific location of the damage is unknown, and the only way is to carry out overall renovation and reinforcement, which requires the bridge to be locked for a long time, causing great trouble to urban traffic. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a method for rapid identification of bridge damage, thereby reducing the cost of the monitoring system, providing auxiliary decision-making for traffic management departments, and providing support for promoting intelligent and scientific management of local urban traffic.

[0005] Technical solution: A method for rapid identification of bridge damage according to the present invention comprises the following steps:

[0006] S1. Collect data using a camera, the data including road data, traffic data, bridge structure data, and vehicle information data;

[0007] S2, using the vehicle weighing device to collect vehicle weight data;

[0008] S3, performing finite element division on the collected road information according to the actual situation;

[0009] S4. Obtain the displacement value of each point on the bridge deck according to the bridge structure data;

[0010] S5, analyzing the collected data, including mathematical transformation, filtering and noise reduction;

[0011] S6. Determine the position of the vehicle;

[0012] S7, based on the collected bridge structure data and the collected vehicle weight data, the finite element knowledge is integrated to perform damage identification on various parts of the bridge;

[0013] S8. Classify the risk level based on the degree of damage to each part;

[0014] S9. Take appropriate measures to repair the bridge to ensure its safety and health;

[0015] S10. After taking measures, the bridge deck is re-identified for damage to ensure that no new safety hazards arise due to repairs, thus ensuring the health and safety of the bridge.

[0016] The road data described in step S1 includes the number of lanes on the road and the length and width of the bridge road; the traffic data includes the number of vehicles on each lane; the bridge structure data mainly includes the displacement values ​​of various parts of the bridge deck; and the vehicle information data mainly includes the weight of each vehicle and the location of the vehicle.

[0017] The road data and road traffic data described in step S1 are obtained by directly capturing the camera above the bridge deck; the bridge structure data is captured by the cameras on both sides of the bridge deck, and then the video is subjected to sub-pixel edge detection, and 5G technology is used for data transmission, eliminating a large amount of line laying work.

[0018] The step S2 specifically includes: determining the vehicle license plate number and vehicle type according to the vehicle weight data measured by the vehicle weighing device and combining the vehicle information data, so as to prepare for the subsequent vehicle positioning.

[0019] The S3 is specifically as follows: during the bridge deck displacement measurement, the bridge main beam is divided into regions, the region length is 0.05 times of the bridge road, and the width is 0.5 times of the lane width, and the bridge deck road is divided into units.

[0020] The S4 is specifically as follows: during the bridge deck displacement measurement process, the surface features of each area of ​​the bridge deck on both sides of the main beam are extracted using a sub-pixel edge detection method, and then curve fitting is performed on the extracted features to obtain the bridge deck displacement information.

[0021] Specifically, S6 includes: keeping the shooting time of the cameras on both sides of the bridge consistent with the shooting time of the cameras above the bridge, determining the position of the vehicle in the video and the weight of the vehicle by the license plate number, and establishing a calculation model in combination with the grid divided by the bridge deck as mentioned above.

[0022] A computer storage medium stores a computer program, which implements the above-mentioned method for rapid identification of bridge damage when executed by a processor.

[0023] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the method for quickly identifying bridge damage is implemented.

[0024] Beneficial effects: Compared with the prior art, the present invention has the following advantages:

[0025] 1. The present invention obtains structural deformation data based on machine vision methods, which significantly reduces costs compared to traditional sensor methods;

[0026] 2. The present invention does not require the laying of a large number of lines, and can obtain deformation data of multiple components and multiple positions of the structure through a small number of cameras, thus having a high measurement efficiency;

[0027] 3. The present invention constructs an evaluation system based on different structural indicators to identify damage in different areas of the structure, establish damage levels, and take maintenance measures based on the damage levels to keep the bridge structure in a healthy state, weakening the control of a single structural indicator and emphasizing the determination of the status of different areas of the structure, which has better operability in engineering practice;

[0028] 4. The present invention can monitor the safety and health of bridges through a small number of cameras. Through the displacement changes of the bridge deck and the weight information of the vehicle, the specific location of bridge damage can be determined, reducing the pressure on the maintenance department;

[0029] 5. The present invention can quickly repair the bridge structure by quickly identifying the specific area damage. There is no need to block the bridge for repair. It only needs to close the specific damaged area and occupy a small amount of traffic channel, thereby greatly reducing the traffic pressure caused by repair and reducing the pressure on the traffic management department;

[0030] 6. The invention combines maintenance methods with structural data based on a closed-loop control method to fully grasp the health of the structure, serve urban traffic optimization, and provide useful support and assistance for promoting urban intelligent transportation construction;

[0031] 7. The present invention simplifies the complex stress conditions of the bridge based on the finite element method, making it possible to measure the damage of specific areas of the bridge deck, providing assistance to the transportation department and promoting the standardized development of urban transportation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a flow chart of a system for rapid identification of bridge deck damage based on sub-pixel edge detection and finite element method according to the present invention;

[0033] Figure 2 The schematic diagram for camera installation is as follows;

[0034] Figure 3 This is a schematic diagram of the installation of ribbons on both sides of the bridge;

[0035] Figure 4 This is a schematic diagram of the picture of the Shihu Bridge deck with sub-pixel edge detection;

[0036] Figure 5 This is the displacement time diagram at the node i=1, j=7 on the bridge deck before repair;

[0037] Figure 6 This is the displacement time diagram of the bridge deck at nodes i=1, j=7 after repair;

[0038] Figure 7 Image of the number of frames returned for the peak value of the image. DETAILED DESCRIPTION

[0039] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0040] A method for quickly identifying bridge damage includes the following steps:

[0041] Step 1: Select a good shooting position to ensure that the camera can capture the entire bridge deck area on one side of the bridge span, and ensure that the cameras above the bridge head and the bridge tail can capture the entire bridge deck. The shooting time of the four cameras is strictly synchronized. The camera installation position is as follows: Figure 2 Install ribbons on both sides of the bridge, such as Figure 3 Show.

[0042] Step 2: Use a camera or image acquisition device to capture video or images of the bridge structure in operation.

[0043] Step 3: Use digital image methods to process the collected bridge structure video to obtain monitoring data at different locations on the bridge deck

[0044] Step 4: First, convert the color image into grayscale. The grayscale conversion formula (1)

[0045] A=0.299R+0.587G+0.114B (1)

[0046] Step 5. Then use the Canny operator to preliminarily obtain the shape of the cable in the image. The Canny algorithm uses four gradient operators to calculate the gradients in the horizontal, vertical and diagonal directions respectively. However, four gradient operators are usually not used to calculate the four directions separately. The commonly used edge difference operator calculates the horizontal and vertical differences Gx and Gy, as shown in formula (2). The gradient angle θ ranges from radians -π to π, and then it is approximated to four directions, representing the horizontal, vertical and two diagonal directions (0°, 45°, 90°, 135°). It can be divided into ±iπ / 8 (i=1,3,5,7), and the gradient angle falling in each area is given a specific value, representing one of the four directions, as shown in formula (3). Calculate the gradient modulus and direction. Here, the Sobel operator is selected to calculate the gradient, as shown in formula (4). Compared with other edge operators, the edge obtained by the Sobel operator is thick and bright:

[0047]

[0048] θ=atan2(G x +G y )(3)

[0049]

[0050] Step 6: Extract the edge points of the bridge deck using the sub-pixel edge detection method. The Gaussian sub-pixel edge detection formula is as shown in (5)-(6).

[0051]

[0052] g: Gaussian model

[0053] M: represents the intensity of the corresponding pixel position

[0054] (x,y): a pixel position in the image

[0055] A: The intensity of the background

[0056] B: Peak intensity in bright areas

[0057] (u,v): The location of the peak in the bright area

[0058] σ: Gaussian model variance

[0059] Then, the extracted edges are fitted with the least squares quadratic function, as shown in formula (7):

[0060]

[0061] Where f(x) is a quadratic polynomial expression, x is the unknown number, and a, b, and c are the polynomial coefficients.

[0062] Step 7. Through the above steps, a curve (quadratic function) is obtained. The curve represents an area of ​​the bridge deck, and the change of the curve represents the displacement of the area. Then a fixed value of 100 is substituted to obtain the corresponding function value. This function value is the ordinate value of the area on the image. The displacement value of the bridge deck area is obtained by the change of the ordinate value on each frame of the image.

[0063] Step 8: By using the camera above the bridge head and the vehicle weighing device at the bridge head, the vehicle information and vehicle weight are combined to obtain the position of the vehicle at any time.

[0064] Step 9. Use the finite element method to divide the bridge deck into rectangular units. Use four-node units. Design the unit width based on the lane width. Divide the unit length l based on the lane length L. Let m = L / I, then the number of units is m*n, where L is the lane length, n is the number of lanes, and l is the unit length.

[0065] Step 10: Use the Gaussian sub-pixel edge detection method to extract the border of the bridge deck side span ribbon lights, using formulas (1)-(4). Then divide the total span of the bridge deck L into several segments, each with a length of l, and perform a function fitting on each segment boundary, such as formulas (8)-(9)

[0066] y=k*x+b (8)

[0067]

[0068] Step 11: Use a high-speed camera placed above the bridge to take pictures, read the vehicle's location information, and calculate the vehicle weight F according to the vehicle's location. ij Distributed to each node, where i represents the node number on the width, j represents the node number on the length, and the unit selects a four-node unit. The force is evenly distributed on the four nodes. Only the effect of the force on the applied area is considered, and the effect of the force on other areas is not considered. The distributed force at each node is If the vehicle is located in multiple units, the area of ​​the unit where it is located and its weight are divided, and then the weight is distributed to the nodes.

[0069] Step 12: After obtaining the fitting function of the boundary, substitute the node abscissa x of the units on both sides of the bridge deck into formula (8) to obtain the displacement value y at the nodes on both sides: 1j and nj .

[0070] Step 13. According to the knowledge of structural mechanics, the ratio of the width to the length of the bridge deck is very small. It can be considered that the displacement value in the width shows regular changes. At the same time, the bridge deck is divided into units. It can be considered that the bridge is in an elastic structure within the unit area. Through the displacement values ​​of the bridge deck on both sides, according to formula (10), the displacement value at any node can be obtained.

[0071]

[0072] Step 14: Obtain the vehicle weight F according to the weighing device, and obtain the stiffness value of each area of ​​the bridge deck according to formula (11).

[0073] k=F / y (11)

[0074] Step 15: Take pictures with a high-speed camera placed above the bridge to obtain the position of the vehicle at each moment, thereby determining the position of the applied force F. ij , thus obtaining the stiffness value of the corresponding area.

[0075] k ij =F ij / y ij (12)

[0076] Step 16: Analyze the calculated stiffness values ​​of each area of ​​the bridge deck to obtain the stiffness damage index p of each area of ​​the bridge deck. ij , p ij The calculation formula is as follows (13), where k i 'j represents the measured regional stiffness value, k ij Represents the regional standard stiffness value.

[0077] p ij =|1-k i ' j / k ij |*100%(13)

[0078] Step 17: Evaluate the structural condition based on the collected stiffness values ​​of each area. There are two main evaluation methods: one is the damage evaluation level q of each area, and the other is based on the overall damage index Q. t The overall damage level Q is obtained. The damage level distribution of each area is shown in Table 3. The overall damage assessment index is as shown in formula (14)(15)(16)(17), where A ij It represents the weight of the stiffness damage index of each area. The Gaussian series G(x) is adopted along the length of the bridge deck, and the quadratic function series F(x) is adopted in the direction perpendicular to the length of the bridge deck.

[0079]

[0080]

[0081] Among them A ij Indicates the weight of the stiffness damage index of each area. The Gaussian series G(i) is taken along the length of the bridge deck, and the quadratic function series F(j) is taken in the direction perpendicular to the length of the bridge deck. i, j represent the number of nodes in length and width. Thus, the overall damage index Q is obtained. t , and then according to the overall damage index Q t The overall damage level Q is obtained, and the overall damage level is shown in Table IV.

[0082] Step 18: Select a suitable traffic maintenance plan based on the damage level of each area and the overall damage level, giving priority to the overall damage level. The overall damage level is determined as shown in Table 1, the regional damage level is determined as shown in Table 2, and the traffic maintenance plan is shown in Table 3.

[0083] Step 19, when the overall damage level is safe, then consider the damage level of each area, and take the approach of giving priority to the area with the largest damage level, repair it, and then identify the damage of the bridge deck again. If the damage level of each area is at a safe level, it will no longer be repaired. If there are still areas at a level outside the safe level, then take repair measures for the area with the largest damage level again, and then identify the regional damage again to form a closed-loop control to improve the stability of the calculation model and the accuracy of the adjustment of measures. The flow chart is as follows: Figure 1 shown.

[0084] Case 1: Take the Shihu Bridge in Suzhou as an example. The Shihu Bridge is located on the extension of Baodai West Road and crosses the Beijing-Hangzhou Grand Canal from east to west. The bridge type adopts a single-tower, double-cable, non-back-cable, all-steel cable-stayed bridge. The main span of the bridge is 100m, the bridge width is 20m, and there are 4 lanes. The units are divided into 4*20, with a total of 80 units and 105 nodes. The bridge deck is equipped with traffic monitoring, and the passing vehicles can be identified through video calls. The displacement of the bridge deck can be obtained by calculating the bridge deck in the video, and the vehicle load can be calculated by the weighing devices at both ends of the bridge. A camera is installed on the west bank of the south side of Shihu Bridge, about 70m away from the main bridge, to collect bridge structure data. Because there are too many nodes and areas, the areas with nodes i=1, j=7, i=2, and j=8 are selected, such as Figure 4 The picture shown is a picture of the Shihu Bridge deck with sub-pixel edge detection.

[0085] The displacement time diagram of the bridge deck at nodes i=1 and j=7 before repair is shown in the figure below: Figure 5 As shown in the figure, the displacement time diagram of the bridge deck at nodes i=1 and j=7 after repair is as follows: Figure 6 shown.

[0086] The horizontal and vertical coordinates of the peak are obtained by automatic peak finding. The horizontal coordinate represents time, the vertical coordinate represents the displacement of the bridge deck, and the peak value represents the vehicle passing through the position at that time point. After finding the peak value, the horizontal coordinate, that is, the time point, is returned in the form of returning the number of image frames (the number of frames divided by the shooting frequency is time) to obtain the position of the vehicle. Figure 7 Shown is the picture frame number at which the peak value of the image is returned.

[0087] The vertical coordinate of the peak value can be used to calculate the displacement of the bridge deck. The difference between the peak value and the initial value of the bridge deck is the displacement of the bridge deck. The displacement values ​​at each node are obtained by formulas (8)-(10). When the corresponding vehicle passes by, the displacement values ​​of the four nodes in this area are The unit is pixel.

[0088] The weights of the vehicles passing through this point are determined by the on-board weighing devices at both ends of the bridge to be 7003kg, 7096kg, and 8154kg, respectively. The regional stiffness values ​​of different vehicles passing through this point are then obtained by formula 12. The values ​​at different times are The unit is kg / pixel, and the standard value of regional stiffness is Therefore, the corresponding stiffness damage index is for Substituting the regional damage index into Table 2, we find that the damage level of the region is q = 2, which is at a dangerous level and needs to be repaired. First, we calculate the stiffness damage index p of each region. ijThen, according to formulas (14)-(17), the overall damage index Q can be obtained t , the final Q t =0.73%. By referring to Table 1, it can be found that the overall damage is at the level of Q=0, which is a safe level. Then, corresponding to Table 3, measure 3 is taken. After taking measures, the bridge deck is tested again. The weight of the vehicle passing through this area is measured to be 5689kg, 8706kg, and 6020kg. When the corresponding vehicle passes, the displacement values ​​of the four nodes in this area are, Then the corresponding stiffness damage index after repair is for By looking up Table 1, it can be found that the overall damage is at the level of q=0, which is a safe level. The overall damage level is also at a safe level. There are no alarms in other areas, and the stiffness values ​​are all at a safe level, so the maintenance is completed.

[0089] Table 1 Evaluation of bridge deck overall damage level Q

[0090] <![CDATA[Overall stiffness damage index Q t > Overall damage level Q <![CDATA[0<Q t ≤7%]]> Q=0, the level is safety level <![CDATA[7%<Q t ≤15%]]> Q=1, the level is risk level <![CDATA[15%<Q t ]]> Q=2, the level is dangerous

[0091] Table 2 Damage level q of bridge deck area ij Evaluation

[0092] <![CDATA[Regional stiffness damage index p ij > <![CDATA[Overall damage level q ij > <![CDATA[0<p ij ≤3%]]> <![CDATA[q ij =0, level is safe]]> <![CDATA[3<p ij ≤5%]]> <![CDATA[q ij =1, the level is risk level]]> <![CDATA[8%<p ij ]]> <![CDATA[q ij =2, the level is dangerous]]>

[0093] Table 3 Corresponding measures to be taken

[0094]

[0095]

Claims

1. A method for rapid identification of bridge damage, It is characterized in that The following steps are involved: S1. Collect data using a camera, the data including road data, traffic data, bridge structure data, and vehicle information data; S2, using the vehicle weighing device to collect vehicle weight data; S3, performing finite element division on the collected road information according to the actual situation; specifically, using the finite element method to divide the bridge deck into rectangular units; S4. Obtain the displacement value of each point on the bridge deck according to the bridge structure data; specifically, use the digital image method to process the collected bridge structure data to obtain monitoring data at different positions of the bridge deck; extract the edge points of the bridge deck by the sub-pixel edge detection method, and perform the least squares quadratic function fitting on the extracted edges, bring in the preset fixed value, and obtain the corresponding function value as the ordinate value of the corresponding area of ​​the bridge deck, and obtain the displacement value of the corresponding area of ​​the bridge deck by the change of the ordinate value on each frame of the picture; S5, analyzing the collected data, including mathematical transformation, filtering and noise reduction; S6. Determine the position of the vehicle; S7. Based on the collected bridge structure data and the collected vehicle weight data, the finite element knowledge is integrated to identify the damage of each part of the bridge; specifically, the Gaussian sub-pixel edge detection method is used to extract the boundary of the ribbon lights on the side of the bridge deck, and the total span of the bridge deck is divided into several sections, and a function fitting is performed on each section boundary; combined with the vehicle-related information, the node coordinates of the units on both sides of the bridge deck are substituted to obtain the displacement value at the node, combined with the knowledge of structural mechanics, the displacement value of any node is obtained, and combined with the vehicle weight data collected by the weighing device, the stiffness value of each area of ​​the bridge deck is obtained; S8. Classify the risk level based on the degree of damage to each part; S9. Take appropriate measures to repair the bridge to ensure its safety and health; S10. After taking measures, the bridge deck is re-identified for damage to ensure that no new safety hazards arise due to repairs, thus ensuring the health and safety of the bridge.

2. A method for rapid identification of bridge damage according to claim 1, It is characterized in that The road data described in step S1 includes the number of lanes on the road and the length and width of the bridge road; the traffic data includes the number of vehicles on each lane; the bridge structure data mainly includes the displacement values ​​of various parts of the bridge deck; and the vehicle information data mainly includes the weight of each vehicle and the location of the vehicle.

3. A method for rapid identification of bridge damage according to claim 1, It is characterized in that The road data and road traffic data described in step S1 are obtained by directly capturing the camera above the bridge deck; the bridge structure data is captured by the cameras on both sides of the bridge deck, and then the video is subjected to sub-pixel edge detection, and 5G technology is used for data transmission, eliminating a large amount of line laying work.

4. A method for rapid identification of bridge damage according to claim 1, It is characterized in that The step S2 specifically includes: determining the vehicle license plate number and vehicle type according to the vehicle weight data measured by the vehicle weighing device and combining the vehicle information data, so as to prepare for the subsequent vehicle positioning.

5. A method for rapid identification of bridge damage according to claim 1, It is characterized in that The S3 is specifically as follows: during the bridge deck displacement measurement, the bridge main beam is divided into regions, the region length is 0.05 times of the bridge road, and the width is 0.5 times of the lane width, and the bridge deck road is divided into units.

6. A method for rapid identification of bridge damage according to claim 1, It is characterized in that The S4 is specifically as follows: during the bridge deck displacement measurement process, the surface features of each area of ​​the bridge deck on both sides of the main beam are extracted using a sub-pixel edge detection method, and then curve fitting is performed on the extracted features to obtain the bridge deck displacement information.

7. A method for rapid identification of bridge damage according to claim 1, It is characterized in that Specifically, S6 includes: keeping the shooting time of the cameras on both sides of the bridge consistent with the shooting time of the cameras above the bridge, determining the position of the vehicle in the video and the weight of the vehicle by the license plate number, and establishing a calculation model in combination with the rectangular units divided by the bridge deck as mentioned above.

8. A computer storage medium having a computer program stored thereon, It is characterized in that When the computer program is executed by a processor, a method for quickly identifying bridge damage according to any one of claims 1 to 7 is implemented.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, It is characterized in that When the processor executes the computer program, a method for quickly identifying bridge damage according to any one of claims 1 to 7 is implemented.

Images