JPWO2018207528A1 - Structure abnormality diagnosis device - Google Patents

Abstract

The structure abnormality diagnosis device has a vibration information extraction unit, an information storage unit, and an abnormal state diagnosis unit. The vibration information extraction unit extracts the vibration information of the structure reflected in the image from the captured image of the structure. The information storage unit acquires and stores vibration information for a plurality of times. The abnormal state diagnosis unit performs abnormal state diagnosis of the structure using the vibration information accumulated by the information accumulation unit, and outputs the diagnosis result.

Description

  The present invention relates to a structure abnormality diagnosing device, for example, a structure abnormality diagnosing device that diagnoses the presence or absence of a structure abnormality from the vibration of a structure on a captured image.

  In manufacturing equipment, factories, buildings, etc. for industrial products, vibrations of structures are observed in order to prevent accidents by detecting corrosion / deterioration of structures that compose them, aging of machines, etc. It is performed to detect an abnormal state. A vibration sensor is generally used to detect this abnormal state. As the vibration sensor method, there are a capacitance method, a piezoelectric element method, a laser Doppler method, and the like, but it is necessary to install the vibration sensor at a point to be observed. Therefore, especially when a large number of observation points are provided for the purpose of observing the deterioration of the structure, the installation cost of the vibration sensor and the maintenance cost will be incurred, and the line connection between the vibration sensor and the control and monitoring device will be required. There is also the complexity of.

  As a method of determining whether a structure is in an abnormal state from vibration information, for example, a method of observing a change in amplitude of a specific frequency component is also known. For example, in the method described in Patent Document 1, it is determined whether or not a signal peak value is generated at a specific frequency by imaging an object to be monitored with an infrared camera, performing Fourier transform on time-series data indicating a time change of temperature. The configuration is such that an abnormality is detected by determining

JP, 2015-219098, A

  However, in the method described in Patent Document 1, it is necessary to grasp in advance what kind of state the signal abnormality will be in when it occurs. Therefore, it is not possible to deal with a vibration state that has never been experienced or cannot be predicted.

  The present invention has been made in view of such a situation, and an object thereof is to detect a structure abnormality from vibration information of a captured image without previously grasping a vibration state when a structure abnormality occurs. It is to provide a structure abnormality diagnosing device capable of diagnosing whether or not it is in a different state.

In order to achieve the above object, the structure abnormality diagnosis device of the present invention, a vibration information extraction unit that extracts the vibration information of the structure reflected in the image from the captured image of the structure,
An information storage unit that acquires and stores the vibration information for a plurality of times,
An abnormal state diagnosis unit that performs an abnormal state diagnosis of the structure using the vibration information accumulated by the information accumulation unit, and outputs the diagnosis result,
It is characterized by having.

  According to the present invention, the abnormal state of a structure is diagnosed using the accumulated vibration information, and the diagnosis result is output. Therefore, the vibration state when an abnormality occurs in the structure is grasped in advance. Without doing so, it is possible to diagnose whether the structure is in an abnormal state from the vibration information of the captured image. Therefore, it is possible to detect an abnormal state that cannot be predicted based on previous experience.

The block diagram which shows 1st Embodiment of a structure abnormality diagnostic apparatus. The block diagram which shows 2nd Embodiment of a structure abnormality diagnostic device. 3 is a flowchart showing an example of a first processing step according to the first embodiment. The flowchart which shows the example of the 2nd process process by 1st Embodiment. The flowchart which shows the example of the 1st process process by 2nd Embodiment. The flowchart which shows the 2nd example of a process process by 2nd Embodiment. FIG. 4 is an infrared image diagram showing in time series the outdoor test site where gas leakage and background temperature change occur. The graph which shows the temperature change of two points SP1 and SP2 in an outdoor test site. The flow chart which shows an example of gas field extraction processing (# 10 of Drawing 5 and Drawing 6). The graph which shows the data which performed the time averaging process (# 11, # 14 of FIG. 9). The graph which shows the data which performed the difference process (# 12, # 15 of FIG. 9) with the averaged data and original data. The graph which shows the data which performed the standard deviation process (# 13, # 16 of FIG. 9) by 21 frames before and behind two types of difference data. The graph which shows the data which performed the difference process (# 17 of FIG. 9) of two types of standard deviation data. The standard deviation image figure which shows the display image which multiplied the standard deviation (# 13 of FIG. 9) of the difference with the 21-frame average 5000 times. FIG. 10 is a standard deviation image diagram showing a display image obtained by multiplying the standard deviation (# 16 in FIG. 9) of the difference from the average of three frames by 5000 times. The standard deviation difference image figure which shows the display image which multiplied the standard deviation difference (# 17 of FIG. 9) 5000 times. The infrared image figure which shows an example of a structure. The graph which shows the temperature change of the edge part and non-edge part of a structure. The schematic diagram for demonstrating the principle of the temperature change in the edge part of a structure. The figure which shows the data which carried out the difference process (# 17 of FIG. 9) of the standard deviation data about the edge part and non-edge part of a structure with a graph etc. The schematic diagram which shows the specific example of the recording method of a vibration detection area.

  Hereinafter, a structure abnormality diagnosis device and the like embodying the present invention will be described with reference to the drawings. Note that the same parts or corresponding parts in each embodiment and the like are designated by the same reference numerals, and duplicate description will be omitted as appropriate.

  1 and 2 show schematic configurations of the structure abnormality diagnosis devices 10A and 10B according to the first and second embodiments. The structure abnormality diagnosis device 10A has a vibration information extraction unit 1, an information storage unit 2, and an abnormal state diagnosis unit 3 as functional blocks. The structure abnormality diagnosis device 10B includes a vibration information extraction unit 1, an information storage unit. 2. It has an abnormal state diagnosis unit 3 and a gas distribution extraction unit 4 as functional blocks. The vibration information extraction unit 1 receives a captured image of a structure as an input and extracts the vibration information of the structure reflected in the image from the captured image. The information storage unit 2 acquires and stores vibration information for a plurality of times. The abnormal state diagnosis unit 3 performs an abnormal state diagnosis of the structure using the vibration information accumulated by the information accumulation unit 2 and outputs the diagnosis result. The gas distribution extraction unit 4 extracts the distribution information of the gas existing in the space being imaged from the captured image. The arrows in FIGS. 1 and 2 indicate the flow of data between the functional blocks.

  An input signal (captured image information) to the structure abnormality diagnosis device 10A is input to the vibration information extraction unit 1 as shown in FIG. The output result of the vibration information extraction unit 1 is input to the information storage unit 2 and also to the abnormal state diagnosis unit 3. Data is input from the information storage unit 2 to the abnormal state diagnosis unit 3 as necessary. The output signal (abnormal state diagnosis result) from the abnormal state diagnosis unit 3 is output to the outside as an output from the structure abnormality diagnosis device 10A and is also used for input to the information storage unit 2.

  The input signal (captured image information) to the structure abnormality diagnosis device 10B is input to the gas distribution extraction unit 4 and the vibration information extraction unit 1 as shown in FIG. The output results of the gas distribution extraction unit 4 and the vibration information extraction unit 1 are both input to the information storage unit 2. The output result of the vibration information extraction unit 1 is also input to the abnormal state diagnosis unit 3, and data is input from the information storage unit 2 to the abnormal state diagnosis unit 3 as necessary. The output signal (abnormal state diagnosis result) from the abnormal state diagnosis unit 3 is output to the outside as an output from the structure abnormality diagnosis device 10B, and is also used for input to the information storage unit 2. That is, the information accumulating unit 2 accumulates information in which the spatial distribution information of the gas extracted by the gas distribution extracting unit 4 and the structure abnormal state diagnosis result are associated with each other.

  By outputting the result extracted by the gas distribution extraction unit 4 to the outside of the structure abnormality diagnosis device 10B, the gas distribution information can be effectively used. For example, as shown in FIG. 2, it is preferable to use the gas leak warning device 5 as an output device that performs a gas leak notification, a gas leak location display, and the like when a gas leak occurs. Examples of the gas leakage warning device 5 include a control terminal device in a central monitoring room, a personal computer (stationary type, portable type, etc.), and a mobile terminal (smartphone, touch pad, etc.).

  The structure abnormality diagnosis devices 10A and 10B are CPUs (Central Processing Units), RAMs (Random Access Memory), ROMs (Read Only Memory), and HDDs in digital devices such as personal computers and portable devices (smartphones, tablet terminals, etc.). It is composed of (Hard Disk Drive) and the like, and takes captured image information as an input and outputs an abnormal state diagnosis result as an output. The functional blocks are realized by the CPU reading the processing program for structure abnormality diagnosis stored in the HDD, expanding the processing program in the RAM, and executing it.

  Here, the gas distribution information extraction processing by the gas distribution extraction unit 4 will be described. First, FIG. 7 shows, as an example of an input image to the structure abnormality diagnosis devices 10A and 10B, a time-series infrared image of an outdoor test site where a gas leak and a background temperature change occur. These are infrared images G1 to G4 obtained by shooting moving images with an infrared camera, and are actual shooting data when the temperature drastically decreases as a whole because of being shaded by clouds. Further, the input image is an image that captures the black body radiated light generated by an object having an absolute temperature of 0 K or more, and the signal intensity of the image represents the temperature of the subject. The wavelength region of the image is the infrared region, and in this image, the gas distribution is also recorded as the signal intensity based on the light absorption of the gas. A point SP1 where gas can be ejected is set in the outdoor test site, and a point SP2 where gas is not ejected is also set for comparison with the point SP1.

  The infrared image G1 is an infrared image of the outdoor test site taken at time T1 immediately before the sunlight is blocked by the cloud. The infrared image G2 is an infrared image (with gas) of the outdoor test site taken at time T2, which is 5 seconds after the time T1. At time T2, the sunlight is blocked by the clouds, so the background temperature is lower than at time T1. The image G3 is an infrared image (with gas) of the outdoor test site taken at time T3, which is 10 seconds after the time T1. From the time T2 to the time T3, the state where the sunlight is blocked by the cloud continues, so that the background temperature is lower at the time T3 than at the time T2. The image G4 is an infrared image (with gas) of the outdoor test site taken at time T4, which is 15 seconds after the time T1. From time T3 to time T4, the state in which the sunlight is blocked by the cloud continues, so at time T4, the background temperature is lower than at time T3.

  The graphs of FIGS. 8A and 8B show the temperature changes at the two points SP1 and SP2 of the outdoor test site, respectively. In each graph, the vertical axis represents temperature and the horizontal axis represents the number of frames. The frame rate is 30 fps. Therefore, the time from the 45th frame to the 495th frame is 15 seconds.

  As shown in FIG. 8A, the ejection of gas is started at the point SP1 during the time from the time T1 to the time T2. Therefore, for example, as can be seen from the range indicated by the broken line area, the graph (A) showing the temperature change at the point SP1 and the graph (B) showing the temperature change at the point SP2 are different. Since gas is not ejected at the point SP2, the temperature change at the point SP2 shows only the background temperature change. On the other hand, at the point SP1, the gas is ejected, so the gas is floating at the point SP1. Therefore, the temperature change at the point SP1 indicates the temperature change obtained by adding the background temperature change and the temperature change due to the leaked gas.

  As shown in FIG. 8B, the background temperature is lowered by about 4 ° C. in 15 seconds from time T1 to time T4. Therefore, as shown in FIG. 7, it can be seen that the image G4 is darker (that is, darker) as a whole than the image G1 and the background temperature is lowered. However, it can be seen that the temperature change due to the gas ejected at the point SP1 is slight and is not even 0.5 ° C. For this reason, at time T2, time T3, and time T4, the gas is ejected at the point SP1, but since the background temperature change is much larger than the temperature change due to the ejected gas, images G2 and G3 are generated. The image G4 does not show how the gas is coming out from the point SP1.

  From the graphs shown in FIGS. 8A and 8B, it can be seen that the gas is ejected at the point SP1 (that is, it is understood that the gas leakage is generated at the point SP1). However, as described above, it is not known from the infrared image shown in FIG. 7 that gas is ejected at the point SP1 (that is, it is not known that gas is leaked at the point SP1). Therefore, the infrared image is image-processed in consideration of the temperature change of the background, so that the state in which the gas is leaked can be shown by the image. That is, a slight temperature change component (frequency component) due to the presence of gas is extracted by the method described below.

  FIG. 9 shows an example of the processing flow of gas region extraction (# 10 in FIGS. 5 and 6). As the low-frequency extraction time averaging process (# 11 in FIG. 9) for creating data corresponding to the background temperature change, averaging in 21 frames before and after was performed, and as high-frequency extraction time averaging process in 3 frames before and after. Averaging is performed (# 14 in FIG. 9). The graph of FIG. 10 shows the data subjected to the time averaging process (# 11, # 14 in FIG. 9). In FIG. 10, the original data (FIG. 8 (A)) is shown by a dashed-dotted line, the data (Ave3) subjected to the time averaging processing (# 14 in FIG. 9) is shown by a solid line in the three frames before and after, and the time is shown in 21 frames before and after. The data (Ave21) subjected to the averaging process (# 11 in FIG. 9) is indicated by a broken line.

  Further, by taking the difference from the original data (# 12, # 15 in FIG. 9), waveform data as shown in the graphs of FIGS. 11A and 11B can be obtained. In the graph of FIG. 11A, the difference processing (FIG. 8A) between the data (Ave21) subjected to the time averaging process for low frequency extraction (# 11 in FIG. 9) and the original data (FIG. 8A) in the 21 frames before and after is shown. 9 shows the data subjected to # 12) (source-Ave21), and the graph of FIG. 11B shows the data subjected to the high-frequency extraction time averaging process (# 14 of FIG. 9) in the three frames before and after. 9 shows data (original-Ave3) that has been subjected to difference processing (# 15 in FIG. 9) between (Ave3) and original data (FIG. 8A).

  As can be seen from FIG. 11 (A), gas is emitted from around the 90th frame. However, since the waveform data of FIG. 11B is a waveform in which only a very high frequency component is extracted, the information of the frequency corresponding to the fluctuating fluctuation component of the gas is not included, and changes greatly before and after the gas ejection. You can see that there is no. That is, in the difference between the time averaging processed data for high frequency extraction and the original data, only high frequencies such as sensor noise are extracted and gas is not extracted. In other words, the waveform data of FIG. 11 (A) is a waveform in which gas and high frequency noise components such as sensor noise are added, and the waveform data of FIG. 11 (B) is a waveform of only high frequency noise such as sensor noise. However, since the noise components of both waveforms do not have a perfect correlation, the noise components cannot be removed even if the difference is taken with the waveforms of FIGS. 11 (A) and 11 (B) (noise waveform Subtraction at level works much like adding, leaving a noise waveform.)

  Therefore, the standard deviation is calculated in 21 frames before and after the two types of waveforms (source-Ave21, source-Ave3) shown in FIGS. 11A and 11B (# 13 and # 16 in FIG. 9). In FIG. 12, the solid line indicates the standard deviation stdev21 of the original −Ave21, and the broken line indicates the standard deviation stdev3 of the original −Ave3. Further, the difference value Stdev21-3 between the standard deviation stdev21 and the standard deviation stdev3 is calculated (# 17 in FIG. 9). FIG. 13 shows the standard deviation difference value Stdev21-3. It can be seen from FIG. 13 that the value can be corrected to a value close to 0 up to the 90th frame where no gas is emitted. That is, the processing having the effect of taking the absolute value by taking the standard deviation is sandwiched so that the high frequency noise component can be actually subtracted. 14 to 16 show examples in which this processing is performed on the entire image.

  FIG. 14 is a standard deviation image diagram showing a display image obtained by multiplying the standard deviation stdev21 (# 13 in FIG. 9) of the difference between the 21-frame average Ave21 and the original data by 5000, and FIG. 15 shows the 3-frame average Ave3 and the original data. FIG. 10 is a standard deviation image diagram showing a display image obtained by multiplying the standard deviation stdev3 (# 16 in FIG. 9) of the difference by 5000 times. FIG. 16 is a standard deviation difference image diagram showing a display image in which the standard deviation difference Stdev21-3 (# 17 in FIG. 9) is multiplied by 5000. By performing leak source identification processing on the image shown in FIG. 16 by image processing, the leak location of the leak gas PG can be obtained.

  Next, the vibration information extraction processing by the vibration information extraction unit 1 will be described. FIG. 17 shows an infrared image of an example of the structure ST. FIG. 17 (B) is an enlarged view of the partial area PS existing in the image pickup screen Io shown in FIG. 17 (A). In FIG. 17B, E1 indicates an edge portion that is an area including an edge, and E0 indicates a non-edge portion that is an area that does not include an edge.

  FIG. 18 shows changes in temperature at the edge portion and the non-edge portion. Specifically, in FIG. 18, the thick line shows the temperature change of one pixel in the non-edge portion E0, and the thin line shows the temperature change of one pixel corresponding to the edge in the edge portion E1. As can be seen from the graph of FIG. 18, even if the temperature change is small in the non-edge portion, a large temperature change occurs in the edge. The principle of this temperature change will be described with reference to FIG. FIG. 19 shows temperature values of three pixels in the upper and lower neighborhoods of the edge portion E1 in the frame F1 and the frame F2 1/30 seconds after that. In the frame F1 and the frame F2, the position on the structure corresponding to each pixel is shifted downward by 0.1 pixel due to the vibration of the structure. Therefore, in the frame F1, the temperature difference between the center pixel and the bottom pixel was 2 ° C., but in the frame F2 after 1/30 seconds, the temperature between the center pixel and the bottom pixel in the frame F1 was increased. The difference changes by 0.2 ° C. corresponding to 0.1 pixel of 2 ° C., and it seems that there is a temperature change. Therefore, if the same process as the gas detection is performed, the difference value of the standard deviations will be large in the pixel corresponding to the edge, as shown in FIG. On the other hand, since the temperature change caused by the vibration of the structure is small in the non-edge portion as described above, a large temperature change is detected in the edge portion, and if the temperature change is not detected in the non-edge portion in the vicinity thereof, the structure is You can see that it is vibrating. In this way, the vibration information can be acquired as a change in the value of the pixel corresponding to the edge of the structure ST reflected in the captured image.

  20A corresponds to the graph of FIG. 13, and the difference processing (# 17 in FIG. 9) of the standard deviation data is applied to the pixels of the edge portion E1 and the pixels of the non-edge portion E0 of the structure ST. Data Stdev21-3 is shown. 20B is an enlarged view of the region PS including the edge portion E1 and the non-edge portion E0, and FIG. 20C is the differential value Stdev21-3 of the standard deviations of all pixels in the region PS of 8.33 seconds. 20 (D) is an image showing the average value of, and FIG. 20 (D) is an image obtained by binarizing the image of FIG. 20 (C).

  As the evaluation value indicating the vibration amount, for example, the average value of the standard deviation difference value Stdev21-3 for a predetermined time is used. FIG. 20C shows an example in which this average value is imaged. Further, as shown in FIG. 20 (D), the value may be binarized with an appropriate threshold value, and the number of pixels exceeding the threshold value may be counted and used as the vibration amount evaluation value. The larger the vibration, the more the number of pixels affected. Therefore, the vibration information may be acquired as the number of pixels indicating a pixel value (temperature) change of the structure reflected in the captured image. Further, instead of the difference value of the standard deviation, “standard deviation of difference from averaging in 21 frames before and after processing” (large value in FIG. 12) before the process of subtracting the high frequency component may be used.

  Next, an example of abnormality diagnosis processing by the structure abnormality diagnosis device according to the present invention will be described. The flowchart of FIG. 3 shows a first processing step example by the structure abnormality diagnosis device 10A (FIG. 1). The input image to the structure abnormality diagnosis device 10A is input to the vibration information extraction unit 1. The vibration information extraction unit 1 extracts edges existing in the input image by differentiating the input image (# 30). The vibration information is digitized for each extracted edge (# 32), and the location of the edge and the digitized vibration information are stored and recorded in the information storage unit 2 (# 34). In this case, the vibration information to be recorded is preferably the average value of the standard deviation difference value Stdev21-3 for a predetermined time. Then, for each extracted edge, the vibration information is compared with the vibration information at the same location in the accumulated information, and if there is a difference, it is determined that there is abnormal vibration and abnormality detection processing is performed (# 36). The vibration information of the place determined to have the abnormal vibration is recorded (# 38), and the abnormality diagnosis result is output (# 40). The vibration information may be digitized or recorded for all pixels of the input image.

  The flowchart of FIG. 4 shows a second processing step example by the structure abnormality diagnosis device 10A (FIG. 1). In this second processing step example, the operator confirms the photographed image on the monitor screen and designates the area for monitoring the vibration. The input image to the structure abnormality diagnosis device 10A is input to the vibration information extraction unit 1. The vibration information extraction unit 1 digitizes the vibration information for each designated area (# 42), and accumulates and records the location of the area and the digitized vibration information in the information storage unit 2 (# 44). In this case, the vibration information to be recorded may be the average value of the standard deviation difference value Stdev21-3 of each pixel in the region for a predetermined time, but the standard deviation difference value Stdev21-3 for the predetermined time is predetermined. It may be the number of pixels in the region that exceeds the threshold value of. Then, for each designated area, the vibration information is compared with the vibration information of the same area in the accumulated information, and if there is a difference, it is determined that there is abnormal vibration and abnormality detection processing is performed (# 46). The area determined to have abnormal vibration is recorded (# 48), and the abnormality diagnosis result is output (# 50).

  The flowchart of FIG. 5 shows a first processing step example by the structure abnormality diagnosis device 10B (FIG. 2). The input image to the structure abnormality diagnosis device 10B is input to the gas distribution extraction unit 4 and the vibration information extraction unit 1. The gas distribution information is extracted by the gas distribution extraction unit 4 (# 10). If the presence of gas is confirmed, gas leakage source identification processing is performed using image processing, and the calculated gas leakage location is recorded (# 20). Further, the extracted result is effectively utilized by being output to the outside (# 19). For example, if the output destination is the gas leakage warning device 5, the warning is issued based on the gas distribution information.

  On the other hand, the vibration information extraction unit 1 performs the abnormality detection process by the same process as steps # 30 to # 36 of FIG. Then, the gas leakage location and the vibration information of the area determined to have the abnormal vibration are recorded in association with each other (# 52), and the abnormality diagnosis result is output (# 54).

  The flowchart of FIG. 6 shows a second processing step example by the structure abnormality diagnosis device 10B (FIG. 2). The input image to the structure abnormality diagnosis device 10B is input to the gas distribution extraction unit 4 and the vibration information extraction unit 1. The gas distribution extraction unit 4 performs the same processing as steps # 10 to # 20 of FIG. 5 and records the gas leakage location.

  On the other hand, the vibration information extraction unit 1 performs the abnormality detection process by the same process as steps # 42 to # 46 of FIG. Then, the gas leakage location and the area determined to have the abnormal vibration are recorded in association with each other (# 56), and the abnormality diagnosis result is output (# 58).

  As a method of recording the vibration information (# 34 in FIG. 3 and the like), there is a method of recording the vibration amount for each pixel. However, when the vibration is detected in only a part of the screen, the area is recorded. It is more useful to record the location and the vibration information as the area when the abnormal vibration analysis is performed later. FIG. 21 shows a specific example of a recording method of a region where vibration is detected (hereinafter referred to as vibration detection region SA). FIG. 21A shows a vibration detection area SA in the partial area PS of the captured image, and FIGS. 21B to 21D show a method of recording the vibration detection area SA.

  In the recording method shown in FIG. 21 (B), the vibration detection area SA is recorded as a coordinate point SB composed of a set of coordinates of a plurality of points surrounding the vibration detection area SA in which vibration is detected. In the recording method shown in FIG. 21C, the vibration detection area SA is recorded as a circumscribing circle SC centered on the center of gravity of the vibration detection area SA in which vibration is detected and circumscribing the vibration detection area SA. In the recording method shown in FIG. 21D, the vibration detection area SA is recorded as the coordinates and the inclination (angle θ) of the circumscribing rectangle SD circumscribing the vibration detection area SA where the vibration is detected. As the vibration information of the vibration detection area SA, a value obtained by averaging the average value of the difference value Stdev21-3 of the standard deviation of each pixel in the vibration detection area SA for a predetermined time over the entire vibration detection area SA, or the difference between the standard deviations The number of pixels in the vibration detection area SA in which the average value of the value Stdev21-3 for a predetermined time exceeds the threshold value is preferable.

  When comparing the vibration information and the accumulated information, the information in the same area is compared with each other (# 60 in FIGS. 3 to 6), but at that time, all the accumulated data in the same area are extracted and compared. And there is a case where only the data at almost the same time is extracted for comparison. The former is effective when it is expected that the way of vibration does not depend on time. The latter is when there is time dependency in the way of vibration (for example, periodic mechanical vibration is added from other places), and when you want to eliminate the effect, or conversely, periodic mechanical vibration is added. This is effective when you want to detect vibration abnormalities in a closed state.

As an output example of the abnormality detection process,
-Adding data indicating that an abnormality has been detected to the accumulated information and recording it,
・ Displaying the detected abnormality on the monitoring device,
・ Displaying that an abnormality has been detected on the monitoring device and issuing an alarm,
Etc. In the above description, the case where the vibration information is information on the amplitude of vibration has been described, but the frequency of vibration may be used as the vibration information.

  As can be seen from the above description, each of the above-described embodiments includes the following characteristic configurations (C1) to (C7).

(C1): a vibration information extraction unit that extracts the vibration information of the structure reflected in the image from the captured image of the structure,
An information storage unit that acquires and stores the vibration information for a plurality of times,
An abnormal state diagnosis unit that performs an abnormal state diagnosis of the structure using the vibration information accumulated by the information accumulation unit, and outputs the diagnosis result,
A structure abnormality diagnosis device comprising:

  (C2): In the above configuration (C1), the information storage unit further includes a gas distribution extraction unit that extracts distribution information of gas existing in the space captured from the captured image, and the information storage unit includes the gas distribution extraction unit. A configuration for further accumulating information in which the spatial distribution information of the gas extracted by the section and the structural abnormal condition diagnosis result are associated with each other.

  (C3): The vibration information of the structure reflected in the image is extracted from the captured image of the structure, the vibration information is acquired and accumulated for a plurality of times, and the accumulated vibration information is used to store the vibration information. A method for detecting an abnormality in a structure, which comprises diagnosing an abnormal state of a structure and outputting the diagnosis result.

  (C4): a process of extracting vibration information of the structure reflected in the image from a captured image of the structure, a process of acquiring and accumulating the vibration information for a plurality of times, and the accumulated vibration information. A program for diagnosing abnormalities in a structure, which causes a computer to execute a process of diagnosing an abnormal state of the structure and outputting the diagnosis result.

  (C5): A configuration in which the vibration information in the configurations (C1) to (C4) is acquired as a change in the pixel value of the edge portion of the structure reflected in the captured image.

  (C6): A configuration in which the vibration information in the configurations (C1) to (C4) is acquired as the number of pixels indicating a change in the pixel value of the structure reflected in the captured image.

  (C7): In the above configurations (C1) to (C6), the vibration information at the same time is extracted from the vibration information stored by the information storage unit, and the extracted vibration information is used to construct the structure. A configuration that diagnoses abnormal conditions and outputs the diagnostic results.

  As can be seen from the above-described embodiments, according to the embodiment of the structure abnormality diagnosis device, the abnormal state diagnosis of the structure ST is performed using the accumulated information and the extracted vibration information, and Since the diagnostic result is output, whether or not the structure ST is in an abnormal state can be determined from the vibration information of the captured image without grasping the vibration state when the structure ST is abnormal. It is possible to diagnose. Therefore, it is possible to detect an abnormal state that cannot be predicted based on previous experience. This is the same even when the above-described structure abnormality detection method or structure abnormality diagnosis program is used.

  Then, the spatial distribution of the leaked gas PG and the vibration information of the structure ST are extracted using the captured image in which the spatial distribution of the leaked gas PG is reflected as the strength of the signal, so that the gas leak detection and the abnormal vibration detection are performed at the same time. Moreover, since it is not necessary to provide a large number of vibration sensors, installation cost, maintenance cost, etc. can be reduced, and the complexity of line connection can be eliminated. Further, by recording the gas leak detection result and the vibration abnormal state detection result in association with each other, it is possible to utilize the data as predictive maintenance data.

1 Vibration Information Extraction Section 2 Information Storage Section 3 Abnormal State Diagnosis Section 4 Gas Distribution Extraction Section 5 Gas Leakage Warning Device 10A, 10B Structural Abnormality Diagnosis Device G1 to G4 Infrared Image (Captured Image)
Io Imaging screen PG Leakage gas SA Vibration detection area SB Coordinate point SC Outer circle SD Outer rectangle E0 Non-edge portion E1 Edge portion ST Structure

Claims (5)

A vibration information extraction unit that extracts the vibration information of the structure reflected in the image from the captured image of the structure,
An information storage unit that acquires and stores the vibration information for a plurality of times,
An abnormal state diagnosis unit that performs an abnormal state diagnosis of the structure using the vibration information accumulated by the information accumulation unit, and outputs the diagnosis result,
Structure abnormality diagnosis device having.   The information distribution unit further includes a gas distribution extraction unit that extracts distribution information of gas existing in the space being imaged from the captured image, and the information storage unit stores the spatial distribution information of the gas extracted by the gas distribution extraction unit. The structure abnormality diagnosis apparatus according to claim 1, further storing information in which the structure abnormality state diagnosis result is associated with each other.   The structure abnormality diagnosis device according to claim 1, wherein the vibration information is acquired as a change in a pixel value of an edge portion of the structure reflected in the captured image.   The structure abnormality diagnosis device according to claim 1, wherein the vibration information is acquired as the number of pixels indicating a change in pixel value of the structure reflected in the captured image.   The vibration information at the same time is extracted from the vibration information accumulated by the information accumulating unit, the abnormal state diagnosis of the structure is performed using the extracted vibration information, and the diagnosis result is output. The structure abnormality diagnosis device according to any one of items 1 to 4.