Damage information processing device, damage information processing method, and program

Abstract

The present application provides a kind of damage information processing device, damage information processing method and program, which can be easily identified by user in time sequence with the part of unnatural difference.The damage information processing device (10) is the damage information processing device (10) of structure, with processor (20).Processor (20) obtains the damage information of structure, i.e.first damage information and the second damage information of the time later than first damage information in time sequence, extracts the difference of first damage information and second damage information, i.e.difference information, detects only first damage information or first damage information is greater than second damage information in first classification site of difference information, and the damage information of at least one of first damage information and second damage information is associated with the output of the notification display indicating first classification site to display device.

Description

Technical Field

[0001] This invention relates to a damage information processing apparatus, a damage information processing method, and a program, particularly a damage information processing apparatus, a damage information processing method, and a program for processing damage information of structures. Background Technology

[0002] Structures can suffer damage due to aging and other factors. Therefore, inspections related to structural damage are mostly conducted periodically. Moreover, during periodic inspections, it is desirable to compare past and current inspection results to understand the extent of damage progression and to explore techniques that support the periodic inspections.

[0003] Patent document 1 describes a technique for acquiring two images of the same crack taken at different times, and for determining the changes in the length and width of the crack based on the two images.

[0004] Previous technical documents

[0005] Patent documents

[0006] Patent Document 1: Japanese Patent Application Publication No. 2019-211277 Summary of the Invention

[0007] The technical problem to be solved by the invention

[0008] Here, to confirm the extent of damage growth, the difference between two examination results at different times is sometimes extracted. When extracting this difference, it can sometimes become unnatural in the time series for various reasons. An unnatural difference in the time series refers, for example, to a situation where, although no repair was performed, the location of damage in a past examination result cannot be confirmed at the same location in the current examination result. Another example of an unnatural difference in the time series is a situation where, although no repair was performed, the current examination result is smaller than the past examination result in terms of the quantitative value of the damage (length, width, or area). In other words, an unnatural difference in the time series refers to a situation where a result contrary to irreversible change is obtained in the time series.

[0009] Thus, as a reason for the unnatural differences in the output over time, it is believed that if the inspection conditions (inspection personnel / equipment / environment) are different between past and current inspection results, the photographic images of the inspected object may be incomplete, and the damage may be incompletely extracted from the photographic images.

[0010] Therefore, the inspector (user) needs to identify the locations that exhibit unnatural differences in the time series and take countermeasures such as re-evaluating the inspection conditions, re-photographing, and re-extracting the damage for these locations. Patent Document 1, mentioned above, does not mention such unnatural differences in the time series.

[0011] The present invention was made in view of this situation, and its purpose is to provide a damage information processing apparatus, damage information processing method and program that enables users to easily identify parts with unnatural differences in a time series.

[0012] means for solving technical problems

[0013] To achieve the above objectives, one aspect of the damage information processing apparatus of the present invention is a damage information processing apparatus for a structure equipped with a processor. The processor acquires damage information of the structure, namely, first damage information and second damage information at a later time in the time sequence than the first damage information, extracts the difference between the first damage information and the second damage information, namely differential information, detects a first category part in the differential information where only the first damage information or the first damage information is larger than the second damage information, and outputs a notification display representing the first category part in association with the damage information of at least one of the first damage information and the second damage information to a display device.

[0014] According to this method, the difference information between the first damage information and the second damage information is extracted, and a first classification region that exists only in the first damage information is detected within the difference information. Furthermore, a first classification region in the difference information where the first damage information is larger than the second damage information is detected. Here, the first classification region refers to a region that exists only in the first damage information earlier in the time series or a region where the first damage information is larger than the second damage information; it is a region with unnatural difference information in the time series. Moreover, in this method, by outputting a notification display representing the first classification region in association with damage information of at least one of the first and second damage information to a display device, and displaying the notification display on the display device, the user can easily identify the first classification region.

[0015] Preferably, the processor detects a second category region in the differential information where only the second damage information or the second damage information is larger than the first damage information, and outputs a notification display representing the second category region in association with the damage information of at least one of the first damage information and the second damage information to the display device.

[0016] Preferably, the processor detects a third category region where the first and second damage information overlap or are equal, based on differential information, first damage information, and second damage information, and outputs a notification display representing the third category region in association with the damage information of at least one of the first and second damage information to the display device.

[0017] Preferably, the processor switches between displaying and hiding damage information associated with damage information corresponding to a first classification region, damage information corresponding to a second classification region, or damage information corresponding to a third classification region on the display device.

[0018] Preferably, the processor changes the display method of the second classification part according to the size of the difference contained in the difference information.

[0019] Preferably, the processor performs display processing, that is, displays on the display device an information associated with the information obtained by overlaying the first damage information and the second damage information.

[0020] Preferably, the processor acquires a first photographic image and a second photographic image, the first photographic image acquires first damage information, the second photographic image acquires second damage information, and the processor performs display processing to display the first photographic image and the second photographic image side by side on a display device.

[0021] Preferably, if the first or second photographic image does not meet the predetermined conditions, the processor performs display processing, that is, displays a recommendation to retake the image on the display device.

[0022] Preferably, the processor outputs one or more correction methods corresponding to the first damage information or the second damage information of the first classification site to the display device.

[0023] Preferably, the processor receives correction information for the first or second damage information displayed on the display device, and extracts differential information again based on the corrected first or second damage information.

[0024] Preferably, the processor automatically corrects the first damage information or the second damage information corresponding to the first classification site.

[0025] Another aspect of the present invention is a damage information processing method for a structure equipped with a processor. The method is characterized by the following steps performed by the processor: acquiring damage information of the structure, namely first damage information and second damage information at a later time in the time series than the first damage information; extracting the difference between the first damage information and the second damage information, i.e., differential information; detecting a first classification region in the differential information where only the first damage information exists, or where the first damage information is larger than the second damage information; and outputting a notification display representing the first classification region in association with damage information of at least one of the first and second damage information to a display device.

[0026] Another aspect of the present invention is a program that causes a damage information processing device for a structure equipped with a processor to execute a damage information processing method, characterized in that the processor performs the following steps: acquiring damage information of the structure, namely first damage information and second damage information at a later time in the time sequence than the first damage information; extracting the difference between the first damage information and the second damage information, namely differential information; detecting a first classification part in the differential information where only the first damage information exists or the first damage information is larger than the second damage information; and outputting a notification display representing the first classification part in association with the damage information of at least one of the first damage information and the second damage information to a display device.

[0027] Invention Effects

[0028] According to the present invention, by extracting the difference information between the first damage information and the second damage information, a first classification part that exists only in the first damage information or where the first damage information is larger than the second damage information is detected in the difference information, and a notification display representing the first classification part is output to a display device in association with the damage information of at least one of the first damage information and the second damage information, and the notification display is displayed on the display device, so that the user can easily identify the first classification part. Attached Figure Description

[0029] Figure 1 This is a block diagram illustrating an example of the hardware structure of a damage information processing device.

[0030] Figure 2 It is a block diagram of the functions of a CPU.

[0031] Figure 3 This is a flowchart illustrating a display method using a damage information processing device.

[0032] Figure 4 This diagram illustrates the acquisition of damage information A by the damage information acquisition unit.

[0033] Figure 5 This diagram illustrates the acquisition of damage information B by the damage information acquisition unit.

[0034] Figure 6 It is a diagram that uses feature points of a photographic image to illustrate the alignment of the photographic image.

[0035] Figure 7 This is a diagram illustrating the extraction of differential damage information using DP matching.

[0036] Figure 8 This is a diagram representing an example of differential information.

[0037] Figure 9 This is a diagram representing an example of differential information (quantitative values).

[0038] Figure 10 This diagram illustrates an example of displaying a display image (damage diagram and notification display) output from the display output unit on the display unit.

[0039] Figure 11 This is a diagram illustrating one example of how a variation can be displayed.

[0040] Figure 12 This diagram illustrates how a detector detects cracks from photographic images.

[0041] Figure 13 This diagram illustrates an example of accepting corrections to damage information.

[0042] Figure 14 This diagram illustrates another example of the acceptance of corrections to damage information.

[0043] Figure 15 This is a diagram showing the damage of another modified example displayed on the display.

[0044] Figure 16 This diagram illustrates how damage maps can be displayed and hidden. Detailed Implementation

[0045] The preferred embodiments of the damage information processing apparatus, damage information processing method and program involved in the present invention will now be described with reference to the accompanying drawings.

[0046] Figure 1 This is a block diagram illustrating an example of the hardware structure of the damage information processing device 10.

[0047] The damage information processing device 10 can be configured as a computer or a workstation. The hardware structure of the damage information processing device 10 mainly consists of a data acquisition unit 12, a memory 16, an operation unit 18, a CPU (Central Processing Unit) 20, RAM (Random Access Memory) 22, and ROM (Read Only Memory) 24. Furthermore, the damage information processing device 10 outputs display images to a display unit (display device) 26. The display unit 26 is a client-side display unit of a network-connected client-server system, and the damage information processing device 10 functions as a server responding to requests from the client. Therefore, display images output from the damage information processing device 10 are displayed on the display unit 26. Alternatively, the damage information processing device 10 can be mounted on a computer, and the display unit 26 can be configured as a monitor connected to the computer.

[0048] The data acquisition unit 12 is a data input unit, for example, acquiring information (data) stored in the memory 16. Furthermore, the information stored in the memory 16 can be acquired by the data acquisition unit 12, or it can be pre-stored in the memory 16. For example, the data acquisition unit 12 can acquire photographic images or damage information of the object to be inspected, which will be described later.

[0049] The memory 16 functions as a database. It stores photographic images or damage information of the inspected object acquired by the data acquisition unit 12. For example, the memory 16 may store photographic images or damage images of regularly performed structural inspections accumulated in the past. Furthermore, the structures inspected include civil engineering structures such as bridges, tunnels, and dams, as well as buildings, residential houses, and structures such as walls, columns, and beams. Damage detected during inspections includes cracks, peeling, exposed rebar, leaks, loose lime, and corrosion.

[0050] The operation unit 18 consists of a point-and-click device such as a keyboard or mouse. The user inputs commands to the damage information processing device 10 via the operation unit 18.

[0051] CPU 20 performs its functions by expanding RAM 22 and executing programs stored in memory 16 or ROM 24.

[0052] Figure 2 This is a block diagram of the functions of CPU20.

[0053] CPU20 includes a damage information acquisition unit 30, a differential extraction unit 32, a detection unit 34, and a display output unit 36.

[0054] The damage information acquisition unit 30 acquires damage information. Here, damage information refers to information relating to damage to the structure being inspected, and includes information in various forms. For example, damage information may be vector information representing damage (cracks). Additionally, damage information may be area information related to damage (stripping, exposed rebar, water leakage, loose lime, corrosion).

[0055] The damage information acquisition unit 30 acquires two damage information sets for the same damage acquired at different times. For example, the damage information acquisition unit 30 acquires damage information stored in the memory 16 obtained through periodic inspections. The damage information acquisition unit 30 acquires damage information A (first damage information) acquired at time a, damage information related to the same damage, and damage information B (second damage information) acquired at time b. Furthermore, time a is an earlier time than time b; for example, time a is August 2015, and time b is August 2020. Moreover, there are no particular restrictions on time a and time b, as long as they are different times.

[0056] The differential extraction unit 32 extracts the difference between damage information A and damage information B, i.e., differential information. The differential extraction unit 32 extracts the distinct locations of damage information A and damage information B as differential information. For example, when damage information A and damage information B represent the same crack using vector information, the different locations of damage information A and damage information B are extracted as differential information. Alternatively, for example, when damage information A and damage information B represent the same peeled area using area, the different locations of damage information A and damage information B are extracted. Furthermore, the differential extraction unit 32 can also extract differential information of quantitative values ​​(length, width, area) between damage information A and damage information B. For example, when damage information A and damage information B have information related to crack width, differential information of that crack width is extracted.

[0057] The detection unit 34 detects regions in the differential information that exhibit unnatural differences in the time series, which are designated as first-class regions. Here, a first-class region refers to a region in the differential information where only damage information A exists. Furthermore, a first-class region refers to a region where, when the damage information is expressed as a quantitative value, damage information A is larger than damage information B. Even though damage information A was acquired earlier in the time series than damage information B, a region present in damage information A but not in damage information B is unnatural in the time series. Additionally, even though damage information A was acquired earlier in the time series than damage information B, a region where the quantitative value of damage information A is larger than the quantitative value of damage information B is unnatural in the time series. Moreover, damage information A and damage information B do not repair the detected damage.

[0058] Furthermore, the detection unit 34 detects the location where damage grows or appears in the differential information, which is the second classification location. Here, the second classification location refers to the location in the differential information where only damage information B exists. Additionally, the second classification location refers to the location where, when the damage information is expressed as a quantitative value, damage information B is larger than damage information A. If the damage begins to grow or appear from time 'a' when damage information A is acquired, then only damage information B reflects the damage corresponding to that damage.

[0059] Furthermore, the detection unit 34 detects third-classification regions where damage information A and damage information B overlap, based on damage information A, damage information B, and differential information. Additionally, the detection unit 34 detects third-classification regions where damage information A and damage information B are equal when the damage information is expressed as quantitative values, based on damage information A, damage information B, and differential information. Third-classification regions are the regions of damage where no change was observed between the time damage information A was acquired and the time damage information B was acquired.

[0060] Furthermore, the detection unit 34 detects the aforementioned second and third classification regions as needed (user settings, etc.). Therefore, the detection unit 34 can detect only the first classification region, detect both the first and second classification regions, detect both the first and third classification regions, or detect the first, second, and third classification regions.

[0061] The display output unit 36 ​​outputs the display image displayed on the display unit 26. The display output unit 36 ​​performs the following display processing: outputs damage information A and damage information B, and displays damage information A and damage information B on the display unit 26. For example, the display output unit 36 ​​stacks (overlaps) damage information A and damage information B and displays them on the display unit 26. Additionally, the display output unit 36 ​​outputs a notification display indicating a first classification region to the display unit 26 in association with damage information of at least one of the displayed damage information A and damage information B, and displays the notification display on the display unit 26. In this way, by outputting a notification display indicating a first classification region to the display unit 26 and displaying the notification display on the display unit 26, the first classification region can be notified to the user, allowing the user to easily identify the first classification region.

[0062] Furthermore, the display output unit 36 ​​outputs a notification display indicating the second category location in association with damage information of at least one of the displayed damage information A and damage information B to the display unit 26, and displays the notification display on the display unit 26. In this way, by outputting a notification display indicating the second category location to the display unit 26 and displaying the notification display on the display unit 26, the user can be notified of the second category location, and the user can easily identify the growth or occurrence of damage.

[0063] Furthermore, the display output unit 36 ​​outputs a notification display indicating the third classification location in association with damage information of at least one of the displayed damage information A and damage information B to the display unit 26, and displays the notification display on the display unit 26. In this way, by outputting a notification display indicating the third classification location to the display unit 26 and displaying the notification display on the display unit 26, the user can be notified of the third classification location, enabling a more accurate inspection.

[0064] Next, the damage information processing method using the damage information processing device 10 will be described. Figure 3 This is a flowchart illustrating the display method using the damage information processing device 10.

[0065] First, the damage information processing device 10 acquires damage information A and damage information B through the damage information acquisition unit 30 (step S10: damage information acquisition step). Next, the damage information processing device 10 extracts the difference information between damage information A and damage information B through the difference extraction unit 32 (step S11: difference information extraction step). Then, the damage information processing device 10 detects a first classification region in the difference information through the detection unit 34 (step S12: detection step). Furthermore, the detection unit 34 detects a second classification region and / or a third classification region in the difference information according to user settings. Next, the damage information processing device 10 outputs a notification display indicating the first classification region to the display unit 26 through the display output unit 36 ​​(step S13: display step). Additionally, the display output unit 36 ​​outputs a notification display indicating the second classification region and / or the third classification region to the display unit 26.

[0066] Next, we will provide a detailed explanation of each step of the above display method.

[0067] <Steps for Obtaining Damage Information>

[0068] The damage information acquisition step is performed by the damage information acquisition unit 30. The damage information acquisition unit 30 can acquire damage information, or it can extract damage from an input image to generate and acquire damage information. The following description describes the case where a photographic image is input into the damage information acquisition unit 30, and the damage information acquisition unit 30 extracts damage from the photographic image to acquire damage information.

[0069] Figure 4 This diagram illustrates the acquisition of damage information A by the damage information acquisition unit 30.

[0070] exist Figure 4(A) shows a photographic image (first photographic image) 50 acquired by the damage information acquisition unit 30. The damage information acquisition unit 30 acquires a photographic image 50 of the bridge pier C taken at time a. Furthermore, the damage information acquisition unit 30 extracts cracks from the photographic image 50 using various techniques. For example, the damage information acquisition unit 30 extracts cracks from the photographic image 50 using image processing or a detector that has undergone machine learning, and acquires damage information A (vector information).

[0071] exist Figure 4 (B) shows the damage information A acquired by the damage information acquisition unit 30. Furthermore, damage information A is vector information, and a damage map generated based on this vector information is shown. The damage information acquisition unit 30 extracts the crack captured in the photographic image 50 and generates and acquires it as damage information A. Damage information A is vector information and includes information about the location, width, and length of the crack in the photographic image 50.

[0072] Figure 5 This diagram illustrates the acquisition of damage information B by the damage information acquisition unit 30.

[0073] Figure 5 (A) shows a photographic image (second photographic image) 54 acquired by the damage information acquisition unit 30. The damage information acquisition unit 30 acquired the photographic image 54 of the bridge pier C taken at time b. Furthermore, the damage information acquisition unit 30 extracts cracks from the photographic image 54 using various techniques. For example, the damage information acquisition unit 30 extracts cracks from the photographic image 54 using image processing or a detector that has undergone machine learning, and acquires damage information B (vector information).

[0074] Figure 5 (B) shows the damage information B acquired by the damage information acquisition unit 30. Furthermore, the damage information B is vector information, and the diagram illustrates a damage map generated based on this vector information. The damage information acquisition unit 30 extracts the crack captured in the photographic image 54 and generates and acquires it as damage information B. The damage information B is vector information and includes information about the location, width, and length of the crack in the photographic image 54.

[0075] Furthermore, when the input photographic images 50 and 54 are not suitable for acquiring damage information (when predetermined conditions are not met), the display output unit 36 ​​can also output a recommendation to retake the photograph to the display unit 26 and display the recommendation to retake the photograph on the display unit 26. Seeing this display, the user can then input a photographic image that has been retaken.

[0076] As described above, in the damage information acquisition step, the damage information acquisition unit 30 acquires damage information A and damage information B.

[0077] <Differential Information Extraction Steps>

[0078] The differential information extraction step is performed by the differential extraction unit 32. The differential extraction unit 32 extracts the difference between damage information A and damage information B using various methods. Below, examples of the differential information extraction methods of the differential extraction unit 32 will be explained.

[0079] (Example 1)

[0080] In Example 1, the difference extraction unit 32 uses image features to align damage information A and damage information B, and extracts the difference information between damage information A and damage information B. In the example above, the damage information acquisition unit 30 acquires photographic image 50 to acquire damage information A and photographic image 54 to acquire damage information B. In this case, the difference extraction unit 32 aligns photographic images 50 and 54, and based on the information obtained from this alignment, aligns damage information A and damage information B.

[0081] Figure 6 This diagram illustrates the alignment of photographic images 50 and 54 using feature points from photographic images 50 and 54.

[0082] First, the difference extraction unit 32 extracts feature points from photographic images 50 and 54. For example, the difference extraction unit 32 uses ORB (Oriented Fast and Rotated BRIEF) or AKAZE (Accelerated-Kaze) feature point detection techniques to extract feature points from photographic images 50 and 54. Figure 6 The image shows one feature point PA, one of the multiple feature points extracted from photographic image 50. Additionally, the image shows one feature point PB, one of the multiple feature points extracted from photographic image 54.

[0083] Subsequently, the difference extraction unit 32 derives the corresponding feature points between photographic images 50 and 54. The difference extraction unit 32 uses techniques such as Brute-Force and FLANN (Fast Library for Approximate Nearest Neighbors) to derive the corresponding points between photographic images 50 and 54. Figure 6 In the case shown, the straight line COR represents the correspondence between feature point PA and feature point PB.

[0084] As described above, the difference extraction unit 32 derives the correspondence between multiple feature points of photographic image 50 and multiple feature points of photographic image 54 (correspondence point detection). Then, the difference extraction unit 32 uses the detected correspondence points to align photographic images 50 and 54. Specifically, based on the correspondence between photographic images 50 and 54, a coordinate transformation model (coordinate transformation matrix) is generated between the coordinates on photographic image 50 and the coordinates on photographic image 54. Then, the difference extraction unit 32 uses the coordinate transformation model to align damage information A and damage information B, and extracts the difference using the aligned damage information A and damage information B.

[0085] (Example 2)

[0086] In Example 2, the difference extraction unit 32 uses a dynamic programming (DP) matching method to align damage information A and damage information B, and extracts the difference information between damage information A and damage information B. Furthermore, the DP matching-based alignment method is suitable for extracting the difference of damage information related to cracks.

[0087] Figure 7 This is a diagram illustrating the extraction of differential damage information using DP matching.

[0088] Damage information AP and damage information BP are damage information of the same crack at different times. Damage information AP and damage information BP are vector information representing the crack. The differential extraction unit 32 uses DP matching to derive the corresponding parts a to d of damage information AP and damage information BP. Then, the differential extraction unit 32 extracts the difference information between damage information AP and damage information BP for each corresponding part a to d.

[0089] (Example 3)

[0090] In Example 3, the difference extraction unit 32 uses an AI (Artificial Intelligence) method to align damage information A and damage information B, and extracts the difference information between damage information A and damage information B. For example, the difference extraction unit 32 uses a recognition device for image alignment learned (generated) using deep learning to align photographic images 50 and 54, and based on the aligned position information, aligns damage information A and damage information B to extract the difference information.

[0091] As described above, the difference extraction unit 32 performs alignment of damage information A and damage information B using various methods, and after alignment, extracts the difference between damage information A and damage information B.

[0092] <Detection Steps>

[0093] The detection process is performed by the detection unit 34. The detection unit 34 detects locations where only damage information A is present in the differential information, or locations where damage information A is greater than damage information B; these are classified as first-category locations. Additionally, the detection unit 34 detects locations where only damage information B is present in the differential information, or locations where damage information B is greater than damage information A; these are classified as second-category locations. Furthermore, based on the differential information, damage information A, and damage information B, the detection unit 34 detects locations where both damage information A and damage information B are present, or locations where damage information A and damage information B are equal; these are classified as third-category locations.

[0094] Furthermore, the detection unit 34 identifies regions where damage information A is greater than damage information B in the quantitative value of damage information as first-classification regions. Additionally, the detection unit 34 identifies regions where damage information B is greater than damage information A in the quantitative value of damage information as second-classification regions. Finally, the detection unit 34 identifies regions where damage information A and damage information B are equal in quantitative value of damage information as third-classification regions.

[0095] Figure 8 This is a diagram representing an example of the difference between damage information A and damage information B.

[0096] Differential information 52 is obtained by aligning damage information A and damage information B, and subtracting damage information A from damage information B. Furthermore, the differential information 52 shown in the illustration is a damage map generated based on the vector information obtained by subtracting damage information A from damage information B, damage information A, and damage information B. When only damage information A exists, or when the damage information is represented by a quantitative value, location Q (surrounded by a dashed line) in differential information 52 is a location where damage information A is greater than damage information B, and the detection unit 34 detects it as a first-class location. Additionally, when only damage information B exists, or when the damage information is represented by a quantitative value, location R (the location with a solid black line) in differential information 52 is a location where damage information B is greater than damage information A, and the detection unit 34 detects it as a second-class location. Furthermore, when both damage information A and damage information B exist, or when the damage information is represented by a quantitative value, location S (the location with a dashed line) in differential information 52 is a location where damage information A and damage information B are equal, and the detection unit 34 detects it as a third-class location.

[0097] Thus, the detection unit 34 detects the first and second classification regions in the differential information 52. Furthermore, the detection unit 34 detects the third classification region based on damage information A, damage information B, and the differential information 52.

[0098] Figure 9 This is a diagram representing an example of the difference (quantitative value) between damage information A and damage information B.

[0099] exist Figure 9In the middle, for Figure 8 The differential information 52 shown is classified based on quantitative values ​​related to crack width (when damage information is represented by quantitative values). Furthermore, the differential information illustrated represents the value obtained by subtracting the crack width of damage information A from the crack width of damage information B.

[0100] Enlarged views 66 and 68 are magnified views of a portion of the differential information 52. In the crack (number D96_3) portion, the differential information becomes -0.15, and the detection unit 34 detects the crack (number D96_3) portion as a first-class region. Furthermore, in the crack (number D97_3) portion, the differential information becomes 0.05, and the detection unit 34 detects the crack (number D97_3) portion as a second-class region. Additionally, in the crack (number D14_1) portion, the differential information becomes 0.00, and the detection unit 34 detects the crack (number D14_1) portion as a third-class region.

[0101] Thus, the detection unit 34 detects the first and second classification regions in the differential information related to the crack width. Additionally, the detection unit 34 detects the third classification region based on damage information A, damage information B, and differential information 52.

[0102] As described above, the detection unit 34 detects the first classification region, the second classification region, and the third classification region based on the locations where damage information A and damage information B are present in the differential information 52. Furthermore, when the differential information 52 has a quantitative value, the detection unit 34 detects the first classification region, the second classification region, and the third classification region based on the magnitude relationship between damage information A and damage information B (including the case where they are equal).

[0103] <Display Steps>

[0104] The display process is performed by the display output unit 36. The display output unit 36 ​​performs display processing, outputting a notification display indicating the first category to the display unit 26 and displaying the notification on the display unit 26. Furthermore, the display output unit 36 ​​performs display processing, outputting a notification display indicating the second category to the display unit 26 and displaying the notification on the display unit 26. Additionally, the display output unit 36 ​​performs display processing, outputting a notification display indicating the third category to the display unit 26 and displaying the notification on the display unit 26.

[0105] Figure 10 This diagram illustrates an example of a display image (damage diagram and notification display) output from the display output unit 36 ​​being displayed on the display unit 26.

[0106] Damage diagram 53, generated based on damage information A and damage information B, is displayed on display unit 26. In damage diagram 53, a notification display M with a solid white line is shown for the first classification area. Additionally, a notification display N with a solid black line is shown for the second classification area. Furthermore, a notification display P with a dashed line is shown for the third classification area. Thus, in damage diagram 53 displayed on display unit 26, notifications to the first, second, and third classification areas are displayed using color or line type.

[0107] Additionally, display unit 26 displays a white balloon-shaped notification display V1 at the first classification location. Notification display V1 is based on damage information A detected as the first classification location and is displayed at the location notifying the first classification location. Furthermore, display unit 26 displays a black balloon-shaped notification display V2 at the first classification location where damage information A is greater than damage information B. Thus, display unit 26 provides notification to the first classification location in the damage diagram 53 using markers such as balloons.

[0108] As described above, according to this method, the difference information 52 between damage information A and damage information B is extracted, and a first classification region in the difference information 52 is detected. Furthermore, in this method, by displaying a notification indicating the first classification region on the display unit 26 in association with the damage information of at least one of damage information A and damage information B, the user can identify regions with unnatural difference information in the time series.

[0109] Furthermore, in this method, the second classification region in the differential information 52 is detected. Moreover, in this method, by displaying a notification indicating the second classification region on the display unit 26 in association with the damage information of at least one of damage information A and damage information B, the user can identify the location where damage has grown or occurred.

[0110] Furthermore, in this method, the third-classification region is detected based on differential information 52, damage information A, and damage information B. Moreover, in this method, by displaying a notification indicating the third-classification region on the display unit 26 in association with damage information of at least one of damage information A and damage information B, the user can identify regions where the damage has not changed.

[0111] <First Variation>

[0112] Next, a first variation of the above-described embodiment will be described. In this example, a photographic image 62 associated with damage information A corresponding to a region specified by the user and a photographic image 64 associated with damage information B are displayed on the display unit 26.

[0113] Figure 11 This is a diagram illustrating one example of how this example is displayed. Figure 11 (A) is a diagram showing the damage diagram 53 displayed by the display unit 26.

[0114] The user designates a portion of an area F in the damage diagram 53 displayed on the display unit 26. The display output unit 36 ​​accepts the designation of the portion of the area and outputs photographic images 62 and 64 corresponding to the designated area to the display unit 26.

[0115] Figure 11 (B) is a diagram showing the photographic image 62 and photographic image 64 displayed by the display unit 26.

[0116] Photograph 62 is an enlarged view of the region corresponding to region F of photograph 50. Photograph 64 is an enlarged view of the region corresponding to region F of photograph 54. Photograph 62 and photograph 64 are output from display output unit 36 ​​to display unit 26. Then, display output unit 36 ​​performs display processing to display photograph 62 and photograph 64 side by side (comparative display) on display unit 26 in a comparative manner. Alternatively, corresponding damage maps can be overlaid on photograph 62 and photograph 64.

[0117] In this way, by displaying past photographic images 62 and current photographic images 64 of the area specified by the user on the display unit 26, the user can compare and observe the state of the object being inspected in the past and the state of the object being inspected in the present.

[0118] <Second Variation>

[0119] Next, a second variation of the above-described implementation will be described. In this example, the user is prompted with a method for correcting the first classification portion.

[0120] Damage information (damage information A or damage information B) corresponding to the first classification site can be corrected using various methods. Therefore, the display output unit 36 ​​outputs one or more correction methods to the display unit 26. Then, the display unit 26 displays the input correction method and prompts the user with the correction method. Specific examples of correction methods will be described below.

[0121] Damage information corresponding to the first category of locations can be corrected using the inspection time and temperature at the time of inspection. For example, since the volume of concrete changes with temperature, the temperature can be estimated based on the inspection date and time or image features to correct the quantitative value (crack width) in the damage information. Alternatively, the quantitative value in the damage information can also be corrected by the user inputting the date and time or temperature.

[0122] Furthermore, when damage information for the first classification region is generated by a detector that has undergone machine learning, the damage information can be corrected by changing the threshold or other means.

[0123] Figure 12 This diagram illustrates how a detector identifies cracks from a photographic image.

[0124] exist Figure 12 (A) shows a photographic image 70 obtained by taking pictures of the object to be inspected. By inputting the photographic image 70 into a detector (AT) that has been trained through machine learning, cracks captured in the photographic image 70 are detected, and a detection result 72 is output. The detector (AI) output is the probability of cracks, and cracks with a probability above a specified threshold are detected as damage.

[0125] Figure 12 (B) is a diagram representing the detection result 72 output from the detector (AI). For example, the detector (AI) defines a crack as having a probability of 30% or higher, generating detection result 72. Furthermore, in detection result 72, areas with a high probability of cracking are represented by dark lines (indicated by arrow I), and areas with a low probability of cracking are represented by light lines (arrow H). The probability of cracking is displayed using pixel values ​​in detection result 72.

[0126] In cases where only damage information A exists, or where damage information A is larger than damage information B (i.e., the first classification region), the damage detection threshold when generating damage information B is lowered, causing the detector (AI) to re-detect the damage. As a result, damage corresponding to damage information A is also detected in damage information B. Damage information A and damage information B are present in the region detected as the first classification region, or the quantitative value of damage information B is corrected, thus correcting the first classification region. Alternatively, in cases where only damage information A exists, or where damage information A is larger than damage information B (i.e., the first classification region), the damage detection threshold when generating damage information A is raised, causing the detector (AI) to re-detect the damage. As a result, damage information A and damage information B are not present in the region detected as the first classification region, or the quantitative value of damage information A is corrected, thus correcting the first classification region.

[0127] As described above, the display output unit 36 ​​outputs one or more correction methods corresponding to the damage information (damage information A or damage information B) of the first classification part to the display unit 26, and the display unit 26 prompts the user with the correction method by displaying the correction method.

[0128] <Third Variation>

[0129] Next, a third variation of the above-described implementation will be described. In this example, the user manually corrects damage information A or damage information B and then performs differential extraction processing again. As described above, the damage information (damage information A or damage information B) corresponding to the first classification region is corrected. In this example, this correction is performed manually by the user.

[0130] Figure 13 This diagram illustrates an example of accepting corrections to damage information.

[0131] When a damage map 80 of the crack is displayed on the display unit 26 based on damage information, the user adds a damage map 82 of the crack using a pointing device. The addition of damage map 82 is received by the damage information acquisition unit 30 as correction information. Then, the damage information acquisition unit 30 corrects the existing damage information based on the correction information, and then the differential extraction unit 32 extracts the differential information again.

[0132] Figure 14 This diagram illustrates another example of the acceptance of corrections to damage information.

[0133] When a damage map 84 of a damaged area (e.g., peeling) is displayed on the display unit 26 based on damage information, the user adds a damage map 86 of the damaged area using a pointing device. The addition of damage map 86 is received by the damage information acquisition unit 30 as correction information. Then, the damage information acquisition unit 30 corrects the existing damage information based on the correction information, and then the differential extraction unit 32 extracts the differential information again.

[0134] In this way, for linear damage such as cracks, users can correct the damage information by adding or deleting tracing lines. Similarly, for surface damage such as peeling, users can correct the damage information by adding or deleting regions.

[0135] Furthermore, while the above example illustrates the case where the user manually corrects the damage information, it is also possible to correct the damage information automatically. For example, the damage information acquisition unit 30 can delete damage information A from the part detected as a first-class part, or add damage as damage information B to the part detected as a first-class part, thus automatically correcting the damage information.

[0136] <Fourth Variation>

[0137] Next, a fourth variation of the above-described implementation method will be described. In this example, the display of the damage map is changed according to the magnitude of the differential information (quantitative value).

[0138] Figure 15 This is a diagram showing the damage diagram displayed on the display unit 26 in this example. Furthermore, the display unit 26 is controlled by the display output unit 36 ​​to display a display image output from the display output unit 36.

[0139] Damage diagram 55 is generated based on damage information corresponding to the second classification location. Damage diagram 55 changes the line type representing the crack based on the magnitude of the difference in quantitative value difference information (crack width).

[0140] The damage map indicated by arrow W corresponds to a large increase in crack width. This damage map is represented by a thin dashed line. Conversely, the damage map indicated by arrow X corresponds to a moderate increase in crack width. This damage map is represented by a thick dashed line. Finally, the damage map indicated by arrow Y corresponds to a small increase in crack width. This damage map is represented by a solid line.

[0141] In this way, by changing the display method according to the magnitude of the quantitative value in the differential information, the user can be notified of the dangerous areas where damage is growing.

[0142] <Fifth Variation>

[0143] Next, a fifth variation of the above-described implementation will be described. In this example, the display and hiding of damage maps corresponding to the first, second, and third classification regions are controlled.

[0144] Figure 16 This diagram illustrates the display and hiding of damage diagrams corresponding to the first classification region on the display unit 26. Furthermore, the display unit 26 is controlled by the display output unit 36 ​​to display display images output from the display output unit 36.

[0145] Figure 16 (A) is a diagram showing the damage diagram 53 displayed on the display unit 26. Damage diagram 53 displays all damage diagrams generated based on damage information corresponding to the first classification region, the second classification region, and the third classification region. Furthermore, the damage diagram corresponding to the first classification region is displayed with a white line (indicated by arrow Z).

[0146] Figure 16 (B) is a diagram representing damage diagram 53 that hides the damage diagram (arrow Z) corresponding to the first classification region. The display output unit 36 ​​performs display processing, outputs a display image that hides the first classification region to the display unit 26, and displays it on the display unit 26.

[0147] In this way, by controlling the display and hiding of the damage map corresponding to the first classification part, the second classification part, or the third classification part on the display unit 26, the desired damage map can be displayed to the user.

[0148] <Other>

[0149] In the above embodiments, the hardware structure of the processing unit that performs various processes is as shown below, including various processors. These processors include general-purpose processors (CPUs) that execute software (programs) and function as various processing units; programmable logic devices (PLDs) such as FPGAs (Field Programmable Gate Arrays) whose circuit structure can be changed after manufacturing; and dedicated circuits such as ASICs (Application Specific Integrated Circuits) that have circuit structures specifically designed for performing specific processes.

[0150] A processing unit can be composed of one of these various processors, or it can be composed of two or more processors of the same or different types (e.g., multiple FPGAs, or a combination of CPU and FPGA). Alternatively, a single processor can constitute multiple processing units. Examples of a single processor constituting multiple processing units include: first, in computers such as client or server computers, a processor is composed of a combination of one or more CPUs and software, which functions as multiple processing units; second, in systems-on-a-chip (SoC), a processor is used to implement the overall system functionality including multiple processing units using a single integrated circuit (IC) chip. In this way, various processing units can be constructed using one or more of the aforementioned processors in their hardware structure.

[0151] Furthermore, more specifically, the hardware architecture of these various processors is a circuit composed of circuit elements such as semiconductor components.

[0152] The aforementioned structures and functions can be suitably implemented by any hardware, software, or a combination of both. For example, the present invention can also be applied to a program that causes a computer to perform the above-described processing steps (processing sequence), a computer-readable recording medium (non-transitory recording medium) that records such a program, or a computer capable of installing such a program.

[0153] The examples of the present invention have been described above, but the present invention is not limited to the embodiments described above. Needless to say, various modifications can be made without departing from the spirit of the present invention.

[0154] Symbol Explanation

[0155] 10 Damage Information Processing Device

[0156] 12 Data Acquisition Department

[0157] 16-bit memory

[0158] 18 Operations Department

[0159] 20 CPUs

[0160] 22 RAM

[0161] 24ROM

[0162] 26 Display Section

[0163] 30 Damage Information Acquisition Department

[0164] 32 Differential Extraction Unit

[0165] 34 Testing Department

[0166] 36 Display Output Section

Claims

1. A damage information processing device, which is a structural damage information processing device equipped with a processor, characterized in that, The processor Obtain damage information of the structure, namely, first damage information and second damage information at a later time in the time series than the first damage information. The difference between the first damage information and the second damage information is extracted as the difference information. The detection method identifies a first-classification region where only the first damage information or the first damage information is larger than the second damage information exists within the differential information. The notification indicating the first classification location is output to the display device in association with the damage information of at least one of the first damage information and the second damage information.

2. The damage information processing device according to claim 1, wherein, The processor The detection method identifies second-classification regions where only the second damage information or the second damage information is larger than the first damage information. The notification indicating the second category location is output to the display device in association with the damage information of at least one of the first damage information and the second damage information.

3. The damage information processing device according to claim 2, wherein, The processor Based on the differential information, the first damage information, and the second damage information, a third classification region is detected where the first damage information and the second damage information overlap or where the first damage information and the second damage information are equal. The notification display representing the third category of the damage is output to the display device in association with the damage information of at least one of the first damage information and the second damage information.

4. The damage information processing device according to claim 3, wherein, The processor switches between displaying and hiding damage information associated with the first classification region, the second classification region, or the third classification region on the display device.

5. The damage information processing apparatus according to any one of claims 2 to 4, wherein, The processor The display method of the second classification part is changed according to the magnitude of the difference contained in the difference information.

6. The damage information processing apparatus according to any one of claims 1 to 4, wherein, The processor performs display processing, that is, displays on the display device information associated with the information that overlaps the first damage information and the second damage information.

7. The damage information processing apparatus according to any one of claims 1 to 4, wherein, The processor acquires a first photographic image and a second photographic image. The first photographic image contains the first damage information, and the second photographic image contains the second damage information. The processor performs display processing to display the first photographic image and the second photographic image side by side on the display device.

8. The damage information processing device according to claim 7, wherein, If either the first or second photographic image does not meet predetermined conditions, the processor performs display processing, that is, displays a recommendation to retake the image on the display device.

9. The damage information processing apparatus according to any one of claims 1 to 4, The processor One or more correction methods corresponding to the first damage information or the second damage information of the first classification site are output to the display device.

10. The damage information processing apparatus according to any one of claims 1 to 4, wherein, The processor The device accepts correction information for either the first damage information or the second damage information displayed on the display device. Based on the corrected first damage information or second damage information, the differential information is extracted again.

11. The damage information processing apparatus according to any one of claims 1 to 4, wherein, The processor automatically corrects the first damage information or the second damage information corresponding to the first classification site.

12. The damage information processing apparatus according to any one of claims 1 to 4, wherein, The differential information is a quantitative value, and the first classification part represents an unnatural change in the time series that is the opposite of the irreversible change.

13. A damage information processing method, which is a damage information processing device for a structure equipped with a processor, characterized in that, include The processor performs the following steps: The steps of obtaining damage information of the structure, namely, first damage information and second damage information at a time later in the time series than the first damage information; The step of extracting the difference between the first damage information and the second damage information, i.e., the difference information; The step of detecting a first-classification region in the differential information where only the first damage information exists or where the first damage information is larger than the second damage information; as well as The step of outputting a notification display representing the first classification location to a display device in association with damage information of at least one of the first damage information and the second damage information.

14. A recording medium, which is a non-transitory computer-readable recording medium, wherein, The system contains a program that causes a computer to execute the damage information processing method of claim 13.

Images