KR102260018B1 - Method of monitoring damage to concrete structures - Google Patents

Abstract

The present invention relates to a method of monitoring a damage to a concrete structure, which is able to detect damage generated to a concrete structure, comprising: a step (S10) of reading identification information of an RF tag (12) attached to the structure (10) to be diagnosed through a drone (20) having a camera (23), laser sensor (24), and RF reader (25), and transmitting the information to a main server (30); a step (S20) of loading DB data storing wall surface image information of the structure (10) in accordance with the relevant identification information in the main server (30), and checking if the wall surface image information of the relevant structure (10) is stored or not; a step (S30) in which, when there is no wall surface image information of the relevant structure (10) in the main server (30), the drone (20) approaches the structure (10) to be diagnosed, photographs a wall surface image of the structure (10) with a preset frame and resolution, and transmits the photographed wall surface image information to the main server (30); and a step (S40) in which the main server (30) receives the wall surface image information photographed by the drone (20), processes and analyzes the image of the received wall surface image information, and determines whether the structure (10) to be diagnosed is damaged or not.

Description

METHOD OF MONITORING DAMAGE TO CONCRETE STRUCTURES

The present invention relates to a method for monitoring damage to a concrete structure, and to a method for monitoring damage to a concrete structure capable of detecting damage to the concrete structure.

Today's various concrete structures are prone to deterioration due to the increase in dynamic load due to the enlargement of transportation means and the intrusion of carbon dioxide gas, sulfur dioxide gas, various chlorides, or acidic substances due to exposure to harsh industrial environments.

Accordingly, the strong alkaline concrete structure is neutralized and causes corrosion of the reinforcing bars, so safety strength cannot be guaranteed in design, and the aesthetics is damaged.

However, in most cases, regular inspection for damage monitoring of concrete structures is not performed, and even when damage monitoring is performed, reliable and effective defect inspection is not performed because there is a large deviation depending on the inspector or inspection conditions depending on the naked eye. it is not possible.

US 10-1604037 B1

It is an object of the present invention to provide a method for monitoring damage to a concrete structure capable of detecting damage to the concrete structure by analyzing an image taken using a drone.

Other objects and advantages of the present invention will be set forth below and will be learned by way of example of the present invention. Furthermore, the objects and advantages of the present invention may be realized by means and combinations indicated in the claims.

According to one aspect of the present invention, in a method for monitoring damage to a concrete structure, a structure 10 to be diagnosed through a drone 20 having a camera 23 , a laser sensor 24 , and an RF reader 25 . ) reading the identification information of the RF tag 12 attached to the main server 30 and transmitting (S10); Loading the DB data in which the wall image information of the structure 10 according to the identification information is stored in the main server 30 to check whether the wall image information of the structure 10 is stored (S20); When there is no image information on the wall of the structure 10 in the main server 30, the drone 20 approaches the structure 10 to be diagnosed and the wall of the structure 10 with a preset frame and resolution. Taking an image and transmitting the photographed wall image information to the main server 30 (S30); And the main server 30 receives the wall image information photographed from the drone 20, and performs image processing and analysis of the received wall image information to determine whether the structure 10 to be diagnosed is damaged. In the process of taking an image of the wall of the structure 10 including the step S40, in the step S30, the drone 20 approaches the structure 10 and recognizes the wall surface of the structure 10 step (S310); calculating a trajectory for changing the flight posture of the drone 20 in a form that can be closely attached to the wall surface of the structure 10 (S320); and a step (S330) of the drone 20 changing its posture according to the trajectory to approach the wall surface of the structure 10 and photographing a wall image (S330), wherein the drone 20 has a planar structure, It includes a plurality of rotors that generate thrust in an upward direction, and in the step S330 , the drone control unit 36 rotates at least some of the plurality of rotors in a form parallel to the wall surface of the structure 10 . Control to perform photographing after the posture is changed, and a plurality of measurement sensor units 14 for detecting damage to the structure 10 are installed in the structure 10, and the drone 20 includes the measurement sensor unit ( 14) receives the sensing value and transmits it to the main server 30, and the measurement sensor unit 14 is built into the wall of the structure 10 and is installed at a connection point where vertical and horizontal reinforcing bars are in contact with each other. , a first measuring sensor unit 141 for measuring vibration and pressure and a second measuring sensor unit 143 for measuring a deformation acting on the structure 10, the second measuring sensor unit 143 is a fabric sensor manufactured using an electroactive polymer, the electroactive polymer comprises polyvinylidene fluoride trifluoroethylene, and the weave pattern of the fabric sensor is a plain weave, It is characterized by having components.
(1) sensor fiber bundle: a fiber bundle made by twisting 10 PVDF-TrFE fibers produced by twist-stretching method, (2) electrode fiber: 1k tow of carbon fiber having a fiber diameter of about 7 μm in one strand, and ( 3) Insulation fiber: Polyester fiber to secure electrical short circuit and safety between electrode fibers.

delete

delete

delete

delete

delete

delete

delete

According to the method for monitoring damage to a concrete structure of the present invention, since the abnormal state of the surface of the structure is monitored through a drone, there is no need for an administrator to directly diagnose and check, thereby preventing a safety accident from occurring.

1 is a block diagram of a device for monitoring damage to a concrete structure according to the present invention;
2 is a view showing an example of the drone according to FIG. 1,
Figure 3 is a block diagram of the main server according to Figure 1,
4 is a block diagram of the image processing and analysis unit according to FIG. 3;
5 is a block diagram of the shake estimation unit according to FIG. 4;
6 is a flowchart for explaining a method for monitoring damage to a concrete structure according to the present invention;
7 is a flowchart for explaining the process of photographing a wall image of a structure according to the present invention;
8 is a flowchart for explaining the step of recognizing the wall surface of the structure according to the present invention,
9 is a flowchart for explaining the step of determining whether a structure is damaged by performing image processing and analysis according to the present invention;
10 is a view showing an example of a fabric sensor according to the present invention.

Hereinafter, specific embodiments of the present invention will be described in detail with reference to the drawings. However, the spirit of the present invention is not limited to the presented embodiments, and those skilled in the art who understand the spirit of the present invention may add, change, delete, etc. other components within the scope of the same spirit, through addition, change, deletion, etc. Other embodiments included within the scope of the invention may be easily proposed, but this will also be included within the scope of the invention.

In addition, components having the same function within the scope of the same idea shown in the drawings of each embodiment will be described using the same reference numerals.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings showing preferred embodiments.

In describing the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the gist of the present invention, a detailed description thereof will be omitted.

1 is a block diagram of an apparatus for monitoring damage to a concrete structure according to the present invention, FIG. 2 is a view showing an example of the drone according to FIG. 1, FIG. 3 is a block diagram of the main server according to FIG. 1, FIG. 4 is a configuration diagram of the image processing and analysis unit according to FIG. 3 , FIG. 5 is a configuration diagram of the shake estimation unit according to FIG. 4 , FIG. 6 is a flowchart for explaining a method for monitoring damage to a concrete structure according to the present invention, and FIG. 7 is a flowchart for explaining the process of taking an image of a wall of a structure according to the present invention, FIG. 8 is a flowchart for explaining the step of recognizing a wall of a structure according to the present invention, and FIG. 9 is an image processing according to the present invention And it is a flow chart for explaining the step of determining whether the structure is damaged by performing the analysis, Figure 10 is a view showing an example of the fabric sensor according to the present invention.

1 to 5 , the apparatus for monitoring damage to a concrete structure according to the present invention includes a drone 20 for monitoring the health of the structure 10 and a main server 30 for determining whether the structure 10 is damaged. ) is included.

The structure 10 to be diagnosed may refer to various structures made of concrete including bridges (steel bridges), buildings, and the like. For example, the structure 10 may be a structure such as a building.

In addition, the wall surface of the structure 10 may be a painted surface (painting exterior). According to this, the health monitoring apparatus of the structure of the present invention can diagnose deterioration of the painted surface. Here, deterioration may be understood as a broad concept encompassing corrosion, peeling, checking, chalking, and the like.

The drone 20 is an unmanned aerial vehicle, and may also be expressed as an unmanned aerial vehicle.

As shown in FIG. 2 , the drone 20 has a planar structure and includes a plurality of rotors (eg, four) generating thrust in an upward direction, and at least some of these rotors rotate. Thus, it can be changed into a form that can be closely attached to the wall of the structure, and then the posture can be changed parallel to the wall.

An RF tag 12 for identification of the structure 10 and a plurality of measurement sensor units 14 for detecting damage to the structure 10 are attached or installed to the structure 10 .

The drone 20 includes a sensor value receiver 21 , a controller 22 , a camera 23 , a laser sensor 24 , and an RF reader 25 .

The drone 20 may read the identification information of the RF tag 12 attached to the structure 10 to be diagnosed and transmit it to the main server 30 .

Identification information for identifying the structure 10 is stored in the RF tag 12 , and the RF reader 25 of the drone 20 reads the identification information stored in the RF tag 12 to the RF tag ( 12) to be able to identify the structure 10 corresponding to it.

The main server 30 and the drone 20 are connected through a wired or wireless network to enable communication.

The main server 30 includes a sensing value and image receiving unit 32 , an image processing and analyzing unit 34 , and a drone control unit 36 .

The sensing value and image receiving unit 32 receives the sensing value and the image, and the image processing and analyzing unit 34 processes and analyzes the received image.

The image processing and analysis unit 34 includes an image brightness adjustment unit 341 , a shake estimation unit 343 , and a deblurring performing unit 345 for image processing.

The image brightness adjustment unit 341 adjusts the brightness of the photographed wall image, the shake estimator 343 estimates the degree of shake of the wall image whose brightness is adjusted, and the deblurring unit 345 estimates Deblurring is performed on the photographed wall image according to the degree of shaking.

The shake estimator 343 may estimate a PSF parameter indicating the degree of shake of the input image. In this case, the PSF parameter may be composed of a shake angle and a shake magnitude.

The shake estimation unit 343 includes an image conversion unit 3431 , a near-zero dead point extraction unit 3432 , and a point spread function (PSF) parameter calculation unit 3433 .

Here, the image conversion unit 3431 transforms the image into a frequency domain through discrete Fourier transform in order to check the periodicity of the image, and the zero-approximate point extraction unit 3432 is the zero of the frequency domain on the two-dimensional coordinates. An approximate point is extracted, and the PSF parameter calculating unit 3433 calculates a PSF parameter indicating the degree of shaking by using the near dead point.

The deblurring performing unit 345 removes motion blur caused by shaking of the camera 23, estimates edge information by at least one object included in an image, and combines the estimated edge information and the By comparing the captured images with each other, motion blur generated in the process of capturing the image is removed.

According to an embodiment of the present invention, the image processing and analysis unit 34 may detect damage (including cracks or defects) of the structure 10 , particularly the wall surface, based on the damaged area detection model.

The damage area detection model includes a plurality of wall images, a bounding box including a crack or a defective portion in the wall image, and a label (crack, non-crack, defect, non-defect, leak, corrosion, etc.) indicating the characteristics of the bounding box. ) using end-to-end training is possible. That is, in training the damaged region detection model, which is a deep neural network, it is possible to collectively train the entire network by replacing the learning of a plurality of neural networks according to each function, thereby minimizing the memory load.

Moreover, the damage area detection model of the present invention can significantly improve the defect recognition processing speed compared to the conventional method of detecting cracks on the wall using only a synthetic neural network (CNN), and in addition to crack defects in concrete structures, peeling and dropping It has the advantage that it can be applied to various types of defects and damage such as leaks, leaks, and rebar exposure.

The damage region detection model according to an embodiment of the present invention may be trained using Mini Batch Gradient Descent (MBGD) in order to optimally recognize crack defects in a structure. The mini-batch gradient descent method is a type of gradient descent that randomly selects and learns a data set of a certain size, and has the characteristic of being able to properly allocate computational cost and learning effect. Gradient descent is an optimization technique that adjusts variables in a direction in which the setting error function becomes smaller, and is used to update the weights of neural networks during training.

If damage to the structure is detected, it is determined whether to precisely execute inspection and diagnosis, and the administrator determines whether to precisely execute inspection and diagnosis of abnormal parts by checking the damage history transmitted from the main server 20, and , this will be executed.

The drone control unit 36 controls the operation of the drone 20, and the drone 20 rotates at least some of the plurality of rotors to change its posture to a form parallel to the wall surface of the structure 10, and then performs photographing. can be controlled to do so.

For example, the drone control unit 36 allows the drone 20 to approach the structure 10 to recognize the wall surface of the structure 10 , and may be configured to be in close contact with the wall surface of the structure 10 . A trajectory for changing the flight posture of the drone 20 is calculated, and the drone 20 changes its posture according to the trajectory to approach the wall of the structure 10 and take an image.

In this case, the wall recognition of the structure 10 calculates the distance between the wall surface of the structure 10 and the drone 20 using a value detected by the laser sensor 24 provided in the drone 20 . and compare the calculated distance with the required distance, so that the drone 20 maintains the required distance from the wall surface of the structure 10 so that the drone 20 is constantly moved with respect to the wall surface of the structure 10 . You can keep it flying.

The flight dynamics of the drone 20 are expressed as differential equations for angular acceleration, angular velocity, and attitude transformation, and the propulsion force of the optimal condition can be calculated by reflecting the given conditions and additional conditions. Accordingly, the drone 20 can perform a fast attitude change according to the calculated thrust and trajectory.

In this way, since the drone 20 changes its posture according to the trajectory and approaches the wall of the structure 10 to take an image, it is possible to secure a better quality wall image.

If necessary, a controller different from the flight of the drone 20 may be used in the process of changing the posture. When the controller used for flight is used as it is, in the process of changing the posture, the vertical drag and friction force between the drone 20 and the wall cannot be sufficiently secured, so a fall may easily occur. Therefore, in consideration of the vertical drag and friction between the drone 20 and the wall, a controller capable of stably changing the posture without falling as much as possible can be used.

The drone station 40 is provided with an electromagnetic induction type wireless charging pad 42 for charging the battery of the drone 20, and the drone 20 is positioned in a landing state in a flight standby state.

Accordingly, the drone 20 can charge the battery while landing on the drone station 40 without using a separate charging cable.

Meanwhile, a plurality of measurement sensor units 14 for detecting damage to the structure 10 may be installed in the structure 10 . In this case, the drone 20 may receive the sensing value from the measurement sensor unit 14 and transmit it to the main server 30 .

The measurement sensor unit 14 may include a first measurement sensor unit 141 and a second measurement sensor unit 143 .

The first measurement sensor unit 141 is built in the wall of the structure 10 and is installed at a connection point where vertical and horizontal reinforcing bars are in contact with each other to measure temperature, vibration and pressure.

In general, the structure 10 includes a vertical column formed in a vertical direction and a plurality of horizontal columns extending in a horizontal direction from the vertical column, and reinforcing bars are embedded in each column to increase strength.

The reinforcing bar has resistance as a metal conductor, transmits vibration applied to the structure, has physical properties to transmit heat, and has a resonance frequency different from that of a column (concrete forming a column) in transmitting vibration

The present invention is configured to measure the temperature, vibration and pressure applied to the structure itself through the first measurement sensor unit 141 by using the characteristics of the reinforcing bar.

Preferably, the first measurement sensor unit 141 is built in the wall of the structure 10, and by installing it at the connection point where the vertical and horizontal reinforcing bars are in contact with each other, it is possible to grasp the deterioration, damage, etc. of the structure at a minimum cost.

The second measurement sensor unit 143 may measure the deformation acting on the structure 10 .

The second measurement sensor unit 143 is a fabric sensor manufactured using an electroactive polymer, and the electroactive polymer includes polyvinylidene fluoride trifluoroethylene (PVDF-TrFE).

The fabric sensor has the advantage of being very flexible and durable.

The textile sensor may detect a mechanical minute change that occurs when an external force is applied as an electrical signal and provide it as a numerical value.

The manufacturing of the textile sensor may be divided into an electroactive polymer fiber manufacturing process and a fabric sensor manufacturing process using the electroactive polymer fiber.

The manufacturing process of the electroactive polymer fiber is as follows. Here, as the electroactive polymer, polyvinylidene fluoride trifluoroethylene is used.

(1) Prepare a solution by dissolving PVDF-TrFE powder in a solvent (Methyl ethyl ketone), (2) Pour an appropriate amount of 10 wt% solution on a glass plate, and then use an automatic applicator (ZAA 2300, Zehntner, Switzerland) and an ultra-precise application rod (ZWA 2121) , Zehntner, Switzerland) was used to prepare a film, (3) cut the prepared PVDF-TrFE film into strips of width (2 mm) × length (50 mm) size, twist it by 40π radians, and 450% in the longitudinal direction. elongate.

Thereafter, heat treatment was performed at 140° C. for 24 hours to remove residual solvent in the PVDF-TrFE fiber and improve crystallization, and electrical poling treatment and mechanical stretching treatment were performed to perform piezoelectric performance. was maximized.

Next, the manufacturing process of the fabric sensor using the electroactive polymer fiber is as follows.

The weave pattern of the fabric sensor is plain weave and has three components. (1) sensor fiber bundle: a fiber bundle made by twisting 10 PVDF-TrFE fibers produced by twist-stretching method, (2) electrode fiber: 1k tow of carbon fiber having a fiber diameter of about 7 μm in one strand, and ( 3) Insulation fiber: Polyester fiber to secure electrical short circuit and safety between electrode fibers.

As shown in FIG. 10 , the PVDF-TrFE fiber bundles were aligned on the warp yarns, and electrode fibers and insulating fibers were positioned on the weft yarns, respectively. That is, the sensor fiber bundle was produced by twisting 10 PVDF-TrFE fibers with each other, and as a result of SEM analysis, the average width was observed to be 625 μm. Here, the warp yarn is a yarn laid vertically in the longitudinal direction of the fabric when weaving the fabric, and the weft yarn is a yarn woven while crossing the warp yarn in the horizontal direction.

6 is a flowchart illustrating a method for monitoring the health of a structure according to the present invention.

Referring to FIG. 6 , in a method for monitoring the health of a structure using a drone, a structure 10 to be diagnosed through a drone 20 including a camera 23 , a laser sensor 24 , and an RF reader 25 . Reading the identification information of the RF tag 12 attached to the step (S10); Loading the DB data in which the wall image information of the structure 10 according to the identification information is stored in the main server 30 to check whether the wall image information of the structure 10 is stored (S20); When there is no image information on the wall of the structure 10 in the main server 30, the drone 20 approaches the structure 10 to be diagnosed and the wall of the structure 10 with a preset frame and resolution. Taking an image and transmitting the photographed wall image information to the main server 30 (S30); And the main server 30 receives the wall image information photographed from the drone 20, and performs image processing and analysis of the received wall image information to determine whether the structure 10 to be diagnosed is damaged. In the process of taking an image of the wall of the structure 10 including the step S40, in the step S30, the drone 20 approaches the structure 10 and recognizes the wall surface of the structure 10 step (S310); calculating a trajectory for changing the flight posture of the drone 20 in a form that can be closely attached to the wall surface of the structure 10 (S320); and a step (S330) of taking an image of the wall by approaching the wall of the structure 10 by changing the posture of the drone 20 according to the trajectory.

Referring to FIG. 7 , in the step of recognizing the wall surface of the structure 10 ( S310 ), the wall surface of the structure 10 is performed using a value detected by the laser sensor 24 provided in the drone 20 . calculating a distance between the drone 20 and the drone 20 (S312); And by comparing the calculated distance with the required distance, the drone 20 is constantly maintained with respect to the wall surface of the structure 10 so that the drone 20 maintains the required distance from the wall surface of the structure 10 . It is characterized in that it comprises the step of flying while (S314).

Referring to FIG. 8 , as a step of recognizing the wall of the structure according to the present invention, the step (S40) includes: adjusting the brightness of the photographed wall image (S410); estimating the degree of shaking of the wall image whose brightness has been adjusted (S420); and performing deblurring on the photographed wall image according to the estimated degree of shaking (S430).

Referring to FIG. 9 , the image processing and analysis are performed according to the present invention to determine whether the structure is damaged, and in the step of adjusting the brightness ( S410 ), the first region used for image processing and analysis and the image processing are performed. and dividing the image into a second region not used for analysis, adjusting the image brightness only for the first region, and estimating the degree of shaking (S420), converting the wall image into a frequency domain, In the step (S430) of extracting the near-death point in the frequency domain, calculating a PSF parameter indicating the degree of shaking by using the near-death point, and performing deblurring (S430), at least one object included in the wall image is It is characterized in that the motion blur generated in the process of photographing the wall image is removed by estimating the edge information by means of a method and comparing the estimated edge information and the photographed wall image with each other.

In addition, a plurality of measurement sensor units for detecting damage to the structure 10 are installed in the structure 10 , and the drone 20 receives the sensing value from the measurement sensor unit and the main server 30 ) to be transmitted.

In the above, the configuration and features of the present invention have been described based on the embodiments according to the present invention, but the present invention is not limited thereto, and it is understood that various changes or modifications can be made within the spirit and scope of the present invention. It is intended that such changes or modifications will be apparent to those skilled in the art, and therefore fall within the scope of the appended claims.

10: structure
12: RF tag
14: measurement sensor unit
20 : drone
21: sensor value receiving unit
22: control unit
23 : camera
24: laser sensor
25 : RF reader
30: main server
32: sensing value and image receiving unit
34: image processing and analysis unit
36: drone control unit

Claims (8)

The main server by reading the identification information of the RF tag 12 attached to the structure 10 to be diagnosed through the drone 20 including the camera 23, the laser sensor 24, and the RF reader 25 transmitting to (30) (S10);
Loading the DB data in which the wall image information of the structure 10 according to the identification information is stored in the main server 30 to check whether the wall image information of the structure 10 is stored (S20);
When there is no image information on the wall of the structure 10 in the main server 30, the drone 20 approaches the structure 10 to be diagnosed and the wall of the structure 10 with a preset frame and resolution. Taking an image and transmitting the photographed wall image information to the main server 30 (S30); and
Receiving the wall image information photographed from the drone 20 in the main server 30, and performing image processing and analysis of the received wall image information to determine whether the structure 10 to be diagnosed is damaged (S40),
The process of photographing the wall image of the structure 10 in the step (S30) is,
Recognizing the wall of the structure 10 by the drone 20 approaching the structure 10 (S310);
calculating a trajectory for changing the flight posture of the drone 20 in a form that can be closely attached to the wall surface of the structure 10 (S320); and
The drone 20 changes its posture according to the trajectory and approaches the wall of the structure 10 to take a wall image (S330),
The drone 20 has a planar structure and includes a plurality of rotors that generate thrust in an upward direction, and in step S330 , the drone control unit 36 determines that at least some of the plurality of rotors It rotates to change the posture in a form parallel to the wall of the structure 10, and then controls to perform shooting,
A plurality of measurement sensor units 14 for detecting damage to the structure 10 are installed in the structure 10,
The drone 20 receives the sensing value from the measurement sensor unit 14 and transmits it to the main server 30,
The measurement sensor unit 14 is built into the wall surface of the structure 10 and is installed at a connection point where vertical and horizontal reinforcing bars are in contact with each other to measure temperature, vibration and pressure. A first measurement sensor unit 141 and the structure ( 10) including a second measuring sensor unit 143 for measuring the deformation acting on the
The second measurement sensor unit 143 is a fabric sensor manufactured using an electroactive polymer, and the electroactive polymer includes polyvinylidene fluoride trifluoroethylene,
The weave pattern of the textile sensor is plain weave, (1) sensor fiber bundle: a fiber bundle made by twisting 10 PVDF-TrFE fibers produced by twist-stretching method, (2) electrode fiber: 1 strand has a fiber diameter of about A method of monitoring damage to a concrete structure, characterized in that it has three components: a carbon fiber 1k tow of 7 μm, and (3) insulating fiber: polyester fiber for electrical short circuit between electrode fibers and securing safety.

delete delete delete delete delete delete delete

Images