Steel structure damage identification system based on deep learning and identification method thereof

Abstract

The invention relates to the technical field of steel structure damage identification, and discloses a steel structure damage identification system and method based on deep learning, and the system comprises a crack obtaining module, a number threshold value presetting module and a comparison response module. According to the method, when the number of the cracks is larger than or equal to the approximate threshold value, the degree of approaching the crack number threshold value is high, at the moment, the crack number is approached, and the probability of potential safety hazards becomes high, so that the number of the cracks is graded firstly; and then shortening the time interval of damage detection and identification according to the incremental equal proportion of levels, so as to improve the frequency of damage detection and identification when the number of cracks is close to the threshold value of the number of cracks, so that problems can be found in time, and safety accidents are avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of steel structure damage identification, and specifically to a steel structure damage identification system and an identification method based on deep learning. Background Art

[0002] Due to advantages such as high strength, light self-weight, strong plasticity, and convenient construction, steel structures have been widely used in many fields of building construction. During the use of steel structures, problems such as fatigue damage, corrosion, cracks, stiffness degradation, stability failure, and accidental damage may cause a decrease in bearing capacity, stiffness, and stability, thus leading to safety problems.

[0003] In order to avoid problems in steel structures and potential safety hazards, it is necessary to identify damage to steel structures. There are various ways of damage identification. For example, by applying a magnetic field to the steel structure and using magnetic powder to show the location of cracks, which is suitable for detecting surface and near-surface defects. Using thermal imaging technology to detect the temperature difference of the structure, so as to identify potential problems caused by stress or physical changes.

[0004] A steel structure damage detection method combining a fully connected neural network and a transfer function is disclosed in the Chinese patent with the publication number CN112131781B. It uses ANSYS software to perform transient analysis on the steel structure frame model to obtain the node acceleration perpendicular to the detection surface. Through MATLAB software, Fourier transform and spectrum division are performed on the acceleration of each node to obtain the transfer function, and then the change in the transfer function is obtained by subtracting the transfer functions. The change in the transfer function is used as the input parameter of the fully connected neural network, and the error is corrected by backpropagation, so as to obtain the damage index value of each detection position of the steel structure. This damage detection method requires relatively complex calculations, and the required various equipment is also relatively complex. Although more accurate damage detection results are obtained, it is very difficult to apply it at the construction site of buildings. Summary of the Invention

[0005] Aiming at the problems existing in the above-mentioned prior art, the purpose of the present invention is to provide a steel structure damage identification system and an identification method based on deep learning, so as to be able to detect and identify steel structure damage and avoid the occurrence of safety hazards.

[0006] To achieve the above object, the present invention provides the following technical solution: A steel structure damage identification system based on deep learning, comprising: a crack acquisition module, which acquires the surface image data of a single steel structure through a visual sensor and records the number of cracks on the surface of the steel structure according to the image data. Among them, for the cracks on the surface of the steel structure whose two endpoints intersect, they are recorded as one crack, and for the cracks on the surface of the steel structure whose two endpoints do not intersect, they are recorded as two cracks; a quantity threshold presetting module, which sets the crack quantity threshold on the surface of the steel structure according to the model specification and surface area of a single steel structure; a comparison and response module, which compares the number of cracks on the surface of the steel structure with the crack quantity threshold and makes different responses according to the comparison result. If the number of cracks is greater than or equal to the crack quantity threshold, repair or replace the steel structure. If the number of cracks is less than the crack quantity threshold, perform a secondary judgment.

[0007] In some embodiments, in the secondary judgment, a crack quantity close to the threshold is set, and the crack quantity close to the threshold is less than the crack quantity threshold. When the number of cracks is less than the crack quantity threshold, the number of cracks on the surface of the steel structure is compared with the crack quantity close to the threshold, and different responses are made according to the comparison result.

[0008] In some embodiments, if the number of cracks is less than the crack quantity close to the threshold, it indicates that the number of cracks on the surface of the steel structure is far from the crack quantity threshold, and the steel structure continues to be used; if the number of cracks is greater than or equal to the crack quantity close to the threshold, it indicates that the number of cracks is close to the crack quantity threshold. At this time, the number of cracks is classified according to the degree of closeness to the crack quantity threshold, and for different levels, the time interval for damage detection and identification is dynamically adjusted.

[0009] In some embodiments, in the classification of the number of cracks, the range between the crack quantity close to the threshold and the crack quantity threshold is evenly divided into three equal parts, and each equal part represents an interval level, which are the first-level crack quantity, the second-level crack quantity, and the third-level crack quantity in ascending order from small to large. The level of the crack quantity on the surface of the steel structure is judged according to the number of cracks.

[0010] In some embodiments, in the dynamic adjustment of the damage detection and identification time interval, an initial damage detection time interval is set. For different levels of the number of cracks, taking the initial damage detection and identification time interval as a reference, the damage detection and identification time interval is shortened in equal proportion in ascending order.

[0011] In some embodiments, under the influence of the recognition that two crack endpoints intersect as one crack, the recorded number of cracks is less than the actual number of cracks on the surface of the steel structure. Considering the comprehensive damage detection and identification time interval, the condition of being far from or exceeding the crack quantity threshold, it is judged that the recorded number of cracks being less than the actual number of cracks causes the deviation of the first-level crack quantity level, and a third judgment is made on this deviated level to correct the level.

[0012] In some embodiments, during the three - time judgment process, the recorded number of cracks is further identified in detail to identify the number of intersections of the endpoints of two cracks. The actual number of cracks is the sum of their intersections. According to the total number of cracks, the crack number level it falls into is re - judged, and then corresponding measures are re - taken.

[0013] In some embodiments, the basis for judging whether there is a boundary at the intersection of the endpoints of two cracks is whether there is an obvious corner formed by two obvious segments in the crack. The corner is set to 90°. If there is an obvious corner less than 90° between two segments in the crack, it is determined that there are two cracks; otherwise, it is one crack.

[0014] The present invention further provides the following technical solutions:

[0015] The present invention further provides a steel structure damage identification method based on deep learning, including the following steps: Step 1, obtaining steel structure surface image data through a visual sensor and obtaining the number of surface cracks based on the image data; Step 2, setting a crack number threshold according to the specification model and surface area of the steel structure; Step 3, comparing the number of cracks with the crack number threshold. If the number of cracks is equal to or greater than the crack number threshold, repair or replacement is carried out; if the number of cracks is less than the crack number threshold, secondary judgment is carried out; Step 4, setting a number of cracks close to the threshold. During the secondary judgment, the number of cracks is compared with the number of cracks close to the threshold. If the number of cracks is less than the number of cracks close to the threshold, the steel structure is used normally; if it is less than the number of cracks close to the threshold, the number of cracks is first classified, and then the dynamic adjustment of the damage detection and identification time interval is carried out according to the level of the number of cracks.

[0016] The present invention further provides a computer - readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the above - mentioned steel structure damage identification system based on deep learning.

[0017] The beneficial effects of the technical solutions provided by the present invention compared with the prior art are as follows:

[0018] First, in the present invention, the recorded number of cracks is compared with the crack number threshold. When the number of cracks is greater than or equal to the crack number threshold, it indicates that the number of cracks has exceeded the range that can be used normally, and repair or replacement is required. When the number of cracks is less than the crack number threshold, different judgments need to be made because of different degrees of less. Specifically, a crack number close to the threshold is set, and the number of cracks is compared with the crack number close to the threshold. When it is less than the crack number close to the threshold, it indicates that the degree of approaching the crack number threshold is low, and the steel structure can be used normally. When it is greater than or equal to the crack number close to the threshold, it indicates that the degree of approaching the crack number threshold is high. At this time, it has already approached the crack number threshold, and the probability of potential safety hazards becomes high. Therefore, for this situation, first, the number of cracks is classified, and then the time interval for damage detection and identification is shortened in equal proportion according to the increase in the level, so as to improve the frequency of damage detection and identification when approaching the crack number threshold, so as to be able to detect problems in time and avoid safety accidents.

[0019] Second, when recording the number of cracks, for the convenience of identification, two cracks with intersecting endpoints are regarded as one. On this basis, the initial recording of the number of cracks is facilitated. However, this will lead to the problem that the actual number of cracks is greater than the recorded number of cracks. This problem of actual being greater than recorded has little impact on the cracks in other intervals, but it mainly affects the interval of the number of primary cracks. Therefore, for this interval, more detailed identification is added to obtain the actual number of cracks, and then it is judged whether there is a problem of deviation from the level in the number of primary cracks in this interval. If there is a deviation, it is corrected to the corresponding level, and then corresponding processing is carried out. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Schematic diagram of the intersection of the endpoints of two cracks;

[0021] Figure 2 Schematic diagram of the non - endpoint intersection of two cracks;

[0022] Figure 3 Schematic diagram of the determination of crack endpoint boundaries;

[0023] Figure 4 Schematic diagram of the logical structure of the invention. DETAILED DESCRIPTION OF THE INVENTION

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] It is understood that the term "a" should be understood as "at least one" or "one or more". That is, in one embodiment, the number of an element can be one, while in other embodiments, the number of the element can be multiple. The term "a" should not be understood as a limitation on the quantity.

[0026] The steel structure damage identification system based on deep learning provided by the present invention, as Figure 4 shown, includes:

[0027] Crack acquisition module: This module acquires the image data of the surface of a single steel structure through a vision sensor. The number of cracks on the surface of the steel structure is recorded in this image data, and these recorded crack numbers are used as the basis for subsequent judgment. For these surface cracks of the steel structure, there are various manifestations and may have the problem of interweaving. To this end, in order to accurately record the number of surface cracks, a method for judging the number of crack lines is given here. As Figure 1 shown, in the figure, two cracks intersect, and the intersection point is an end point. In this case, it is recognized as one crack. As Figure 2 shown, in the figure, two cracks intersect, and the intersection point is in the middle, not at the end point. This non-endpoint intersection situation is recognized as two cracks. The core function of this module is to accurately capture the image data of the surface of the steel structure through advanced vision sensor technology, especially record the crack information therein. Through image analysis technology, the number and morphology of cracks are identified, especially paying attention to the interweaving situation of cracks. Specifically, for the interweaving situation where the intersection point is at the end point of the crack, the system recognizes it as one crack; while for non-endpoint interweaving, it is recognized as two cracks. This accurate judgment mechanism ensures the accurate recording of the number of cracks and provides reliable data support for subsequent crack assessment and management.

[0028] Quantity threshold presetting module: Set the crack quantity threshold on the surface of this steel structure according to the model specifications and surface area of a single steel structure. This crack quantity threshold represents the maximum number of cracks that the surface of this steel structure can accept. As long as the number of cracks exceeds this threshold, this steel structure is considered to be no longer suitable for continued use and needs to be repaired or replaced. This module is responsible for reasonably setting the threshold of the crack quantity according to the model specifications and surface area of the steel structure to serve as the evaluation standard for structural safety. The set crack quantity threshold is based on industry standards, specific application environments, and load requirements to ensure the balance between safety and service life. The formulation of the crack quantity threshold is not only based on theoretical calculations but also adjusted in combination with on-site experience and historical data.

[0029] Comparison and response module: Compare the recorded number of cracks on the steel structure surface with the crack number threshold, and make different responses according to the comparison results. If the number of cracks is greater than or equal to the crack number threshold, it indicates that the number of cracks on the steel structure surface has exceeded the acceptable minimum. This steel structure can no longer be used continuously, otherwise it may pose certain safety hazards. In this case, for this steel structure, certain repair operations need to be carried out. If the defects of the steel structure cannot be compensated by repair, the steel structure needs to be replaced. If the number of cracks is less than the crack number threshold, it indicates that the number of cracks on this steel structure is still within the acceptable range and can continue to be used. For this situation, a secondary judgment is made on the steel structure. This module judges whether further measures need to be taken for the steel structure by comparing the detected number of cracks with the preset crack number threshold. If the number of cracks is equal to or exceeds the crack number threshold, it is regarded that there is a potential safety risk in the structure and repair or replacement is required. When the number of cracks is lower than the crack number threshold, the structure is considered to be within the safe range. Such a comparison and analysis mechanism can help managers make decisions quickly, ensuring the stability and safety of the steel structure. Through this systematic judgment method, the maintenance process can be optimized, work efficiency can be improved, and the accident risk caused by cracks can be effectively reduced.

[0030] Secondary judgment process: When the number of cracks on the steel structure surface is less than the crack number threshold, there are two situations. It may be a low degree of being less than the crack number threshold or a high degree. If it is a low degree of being less than the crack number threshold, it means that although the number of cracks on the steel structure surface has not reached the crack number threshold, it is already close to the crack number threshold. In this case, compared with the situation of a high degree of being less than the crack number threshold, the possibility of problems with the steel structure is greater. Therefore, in the secondary judgment, a crack number close to the threshold is set, and this crack number close to the threshold is less than the crack number threshold. Compare the number of cracks on the steel structure surface with the crack number close to the threshold, and make different responses according to the comparison results. If the number of cracks is less than the crack number close to the threshold, it indicates that the number of cracks on the steel structure surface has not yet approached the crack number threshold. In this case, the steel structure can continue to be used. If the number of cracks is greater than or equal to the crack number close to the threshold, it indicates that the number of cracks on the steel structure surface has approached the crack number threshold. In this case, grade the number of cracks on the steel structure surface according to the degree of approaching the crack number threshold, and dynamically adjust the time interval of damage detection and identification for different levels of crack numbers.

[0031] Process for grading the number of surface cracks in a steel structure: Divide the range between the crack number threshold and the value close to the crack number threshold into three intervals, namely the first interval, the second interval, and the third interval. The first interval is close to the value close to the crack number threshold, and the third interval is close to the crack number threshold. When the number of surface cracks in the steel structure is less than the crack number threshold and greater than or equal to the value close to the crack number threshold, obtain the crack number at this time and grade it according to the interval into which the crack number falls at this time. When the crack number at this time falls into the first interval, it is a first-level crack number; when it falls into the second interval, it is a second-level crack number; and when it falls into the third interval, it is a third-level crack number.

[0032] Process for dynamically adjusting the time interval for damage detection and identification: Set the time interval for damage detection and identification of the steel structure under initial conditions to once every N days. When the number of cracks on the surface of the steel structure is between the value close to the crack number threshold and the crack number threshold, as the crack number level increases, proportionally shorten the time interval for damage detection and identification. Because as the crack number level increases, the possibility of potential safety hazards in the steel structure is higher, so it is necessary to shorten the time interval for damage detection and identification in order to be able to detect problems in a timely manner and avoid potential safety hazards.

[0033] Generally speaking, in the loss identification of surface cracks in a steel structure, when the crack number is below but close to the crack number threshold, more detailed analysis is required to ensure safety. Set the "value close to the crack number threshold" as a reference point. If the crack number is higher than or equal to this value, it is necessary to conduct hierarchical management according to its proximity to the crack number threshold. This process divides the crack number between the value close to the crack number threshold and the crack number threshold into three intervals to conduct first-level, second-level, and third-level risk assessments on the cracks, thereby determining corresponding maintenance measures. The first-level risk indicates the smallest potential safety hazard, while the third-level risk indicates approaching the crack number threshold, requiring more frequent inspections and possible repairs. For this reason, the time interval for damage detection and identification adopts a dynamic adjustment mechanism, initially once every N days, and proportionally shortens the detection interval as the crack level increases to ensure that potential structural problems can be detected and addressed in a timely manner. This method not only optimizes the use of resources but also improves the response speed to changes in the structural state, ensuring the safety and reliability of the steel structure throughout its life cycle.

[0034] For example, assume that the crack number threshold for the surface of a single steel structure is 100. Initially, the number of cracks on the surface of the steel structure obtained by a vision sensor is 70. Since 70 is less than 100, a secondary judgment is performed. It is set that the number of cracks on the surface of the steel structure close to the threshold is 70, and the range between the number of cracks close to the threshold and the crack number threshold is 60 - 100. This range is evenly divided into three parts, which are 70 (including) - 80, 80 (including) - 90, and 90 (including) - 100. That is, the primary crack number is 70 (including) - 80, the secondary crack number is 80 (including) - 90, and the tertiary crack number is 90 (including) - 100. Since the number of cracks on the surface of the steel structure is 70, it is the primary crack number. Since the time interval for damage detection and identification under the initial conditions is once every N days, the time interval should be proportionally shortened at this time. Assume that it can be shortened to once every 0.9N days, and so on. If it is the secondary crack number, the time interval is once every 0.8N days; if it is the tertiary crack number, the time interval is once every 0.7N days.

[0035] In addition, when obtaining the number of cracks on the steel structure surface, two cracks with intersecting endpoints are regarded as one crack, which may lead to the recorded number of cracks being less than the actual number of cracks on the steel structure surface. However, the impact of this problem on the number of cracks in each interval varies. Considering the above, the crack number intervals are divided into five, namely: interval A, the number of cracks < the number of cracks approaching the threshold; interval B, the number of cracks approaching the threshold ≤ the number of cracks < the maximum value of the first-level crack number; interval C, the maximum value of the first-level crack number ≤ the number of cracks < the maximum value of the second-level crack number; interval D, the maximum value of the second-level crack number ≤ the number of cracks < the crack number threshold; interval E, the number of cracks ≥ the crack number threshold. For a clearer description, combined with the above example, the five intervals are: interval A, the number of cracks < 70; interval B, 70 ≤ the number of cracks < 80; interval C, 80 ≤ the number of cracks < 90; interval D, 90 ≤ the number of cracks < 100; interval E, the number of cracks ≥ 100. For interval A, since the number of cracks is far from the crack number threshold, even if two cracks are regarded as one, resulting in the actual number of cracks being more than the recorded number, the impact on this interval is minimal. For interval E, the number of cracks in this interval is already greater than or equal to the crack number threshold and requires repair or replacement. Therefore, even if two cracks are regarded as one, resulting in the actual number of cracks being more than the recorded number, it has no impact on this interval. For intervals B, C, and D, generally speaking, they are actually intervals where the number of cracks is close to the threshold. If two cracks are regarded as one in this combined interval, resulting in the actual number of cracks being more than the recorded number, the final impact may be that interval B may be upgraded to interval C, interval C may be upgraded to interval D, interval D may be upgraded to interval E, etc. For interval C, it is itself the second-level crack number interval, and the time interval for damage identification has been shortened twice. Therefore, if there is a problem with the steel structure, it will not fail to be detected and identified. Similarly, the same applies to interval D. The time intervals for detection and identification in these two intervals have been shortened at least twice in proportion. In the worst case, it was not detected during the first damage detection and identification, but the next damage detection and identification is not long after the first one. Therefore, the probability of a potential safety hazard not being discovered is relatively low. Therefore, if two cracks are regarded as one, resulting in the actual number of cracks being more than the recorded number, the interval with the greatest impact is interval B. The time interval for damage detection in this interval is only shortened for the first time. If there are many cases where two cracks coincide as one and the next damage detection and identification is a long time later, it may lead to the problem that a potential safety hazard cannot be discovered in time. Therefore, for interval B, three additional judgments are made.

[0036] Three judgment processes: When the number of steel structure cracks is within interval B, that is, the number of primary cracks, record the number of steel structure cracks L at this time. At the same time, conduct secondary detail identification to identify how many cases there are where the endpoints of two cracks on the surface of the steel structure intersect. Suppose there are X cases where the endpoints of two cracks intersect. Then the actual number of cracks in interval B is L + X. Then, determine which interval L + X falls into, and based on the interval it falls into at this time, take the corresponding measures for this interval. If 70 ≤ L + X < 80, it is still within interval B. Although the number of cracks is corrected at this time, it still does not exceed the range of interval B. The detection period remains once every 0.9N days, and the change trend of the cracks can continue to be observed. If 80 ≤ L + X < 90, it enters interval C. After correction, the number of cracks reaches within the range of interval C. At this time, the cracks are already in the state of secondary cracks, and the detection period is shortened to once every 0.8N days. If 90 ≤ L + X < 100, it enters interval D. After correction, the number of cracks reaches within the range of interval D. At this time, the cracks are already in the state of tertiary cracks, and the period is shortened to once every 0.7N days. If L + X ≥ 100, it enters interval E. After correction, the number of cracks reaches within the range of interval E. At this time, the steel structure needs to be repaired or replaced.

[0037] Generally speaking, according to the interval division and identification strategy of the number of steel structure cracks, the number of cracks is divided into five intervals: Interval A is where the number of cracks is less than 70, interval B is from 70 (inclusive) to 80, interval C is from 80 (inclusive) to 90, interval D is from 90 (inclusive) to 100, and interval E is 100 and above. Under this framework, focus on interval B because during the crack identification and detection process, two cracks with intersecting endpoints are considered as one, resulting in the actual number of cracks being higher than the recorded value, which has the greatest impact on interval B. Therefore, an additional three judgments are taken: First, record the number of cracks L in interval B, then identify the number of intersecting cracks X in detail, correct the actual number of cracks to L + X, and then determine the new interval it is in based on the corrected number. If it is still within interval B after correction, maintain the detection period of once every 0.9N days; if it enters interval C, it is regarded as secondary cracks, and the detection period is shortened to once every 0.8N days; if it enters interval D, it is regarded as tertiary cracks, and the period is reduced to once every 0.7N days; if it reaches interval E, it needs to be repaired or replaced immediately. This strategy ensures that during the crack identification process, even if there are overlaps, potential safety hazards can be identified and handled in a timely manner.

[0038] It is worth mentioning that the boundary at the endpoints of two cracks may not be obvious in reality. Therefore, the basis for judging whether there is a boundary at the intersection of the endpoints of two cracks here is whether there is an obvious corner formed by two distinct segments in the crack. This corner is set at 90°. That is, if there is an obvious angle less than 90° between two segments in the crack, it is determined that there are two cracks, otherwise it is one crack. As Figure 3 shown in Figure 1, the angle at the corner of the two segments is significantly less than 90°, so Figure 1 is determined to have two cracks. AsFigure 3 As shown in Figure 2, the angle at the corner of the two segments is significantly greater than 90°, so Figure 2 is identified as a crack.

[0039] In the embodiments disclosed by the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. The embodiments disclosed by the present invention include a computer program product, which includes a computer program carried on a computer-readable medium. The computer program contains program codes for executing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by a central processing unit, the above-mentioned functions defined in the methods of the present application are executed. It should be noted that the computer-readable medium described above in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wire segments, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory, an optical fiber, a portable compact disk read-only memory, an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, the computer-readable storage medium can be any tangible medium that contains or stores a program, and the program can be used by or combined with an instruction execution system, apparatus, or device. In the present application, the computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries the computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal medium can also be any computer-readable medium other than the computer-readable storage medium, and the computer-readable medium can send, propagate, or transmit a program for use by or combined with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any suitable medium, including but not limited to: wireless segments, wire segments, optical cables, RF, etc., or any suitable combination of the above.

[0040] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0041] Those skilled in the art should understand that the above description is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application.

Claims

1. A steel structure damage identification system based on deep learning, characterized in that: include: A crack acquisition module, which acquires image data of a single steel structure surface through a visual sensor, and records the number of cracks on the steel structure surface according to the image data, wherein two cracks on the steel structure surface with their endpoints intersecting are recorded as one, and two cracks on the steel structure surface with their non-endpoints intersecting are recorded as two; A quantity threshold preset module, which sets the crack quantity threshold on the surface of the steel structure according to the model specification and surface area of ​​the single steel structure; The comparison response module compares the number of cracks on the surface of the steel structure with the crack number threshold, and makes different responses based on the comparison results. If the number of cracks is greater than or equal to the crack number threshold, the steel structure is repaired or replaced. If the number of cracks is less than the crack number threshold, a secondary judgment is performed.

2. The steel structure damage identification system based on deep learning according to claim 1 is characterized in that: In the secondary judgment, a crack number approach threshold is set, and the crack number approach threshold is less than the crack number threshold. When the crack number is less than the crack number threshold, the crack number on the surface of the steel structure is compared with the crack number approach threshold, and different responses are made according to the comparison results.

3. The steel structure damage identification system based on deep learning according to claim 2 is characterized in that: If the number of cracks is less than the crack number approach threshold, it indicates that the number of cracks on the steel structure surface is far away from the crack number threshold and the steel structure continues to be used; if the number of cracks is greater than or equal to the crack number approach threshold, it indicates that the number of cracks is close to the crack number threshold. At this time, the number of cracks is graded according to the degree of proximity to the crack number threshold, and the time interval for damage detection and identification is dynamically adjusted for different levels.

4. The steel structure damage identification system based on deep learning according to claim 3 is characterized in that: In the crack number classification, the range between the number of cracks close to the threshold and the crack number threshold is divided into three equal parts. Each part represents an interval level, from small to large, they are the first-level crack number, the second-level crack number and the third-level crack number. The level of crack number is determined according to the number of cracks on the surface of the steel structure.

5. The steel structure damage identification system based on deep learning according to claim 4 is characterized in that: In the dynamic adjustment of the damage detection and identification time interval, the initial damage detection and identification time interval is set. For different levels of crack numbers, the damage detection and identification time interval is proportionally shortened in order from small to large, with the initial damage detection and identification time interval as a reference.

6. The steel structure damage identification system based on deep learning according to claim 5 is characterized in that: Based on the influence of the intersection of two crack endpoints being identified as one crack, the recorded number of cracks is smaller than the actual number of cracks on the surface of the steel structure. Combined with the damage detection identification time interval and the conditions of being away from or exceeding the crack number threshold, it is judged that the recorded number of cracks is smaller than the actual number of cracks, causing the first-level crack number to deviate from the level. The deviation level is judged three times to correct the level.

7. The steel structure damage identification system based on deep learning according to claim 6 is characterized in that: During the three judgment processes, the recorded number of cracks is re-identified in detail to identify the number of intersections of the two crack endpoints. The actual number of cracks is the sum of the intersections of the two. The level of crack quantity it falls into is re-judged based on the total number of cracks, and then a new response is made.

8. The steel structure damage identification system based on deep learning according to claim 7 is characterized in that: The basis for judging whether there is a boundary at the intersection of the endpoints of two cracks is whether there are two obvious sections in the crack that form a clear turning angle. The turning angle is set to 90°. If there is an angle of less than 90° between the two sections in the crack, it is judged to be two cracks, otherwise it is a single crack.

9. The steel structure damage identification method based on deep learning according to any one of claims 1 to 8, characterized in that: The following steps are involved: Step 1: Obtain surface image data of the steel structure through a visual sensor, and obtain the number of surface cracks based on the image data; Step 2: Set the crack quantity threshold according to the steel structure specifications and surface area; Step 3: Compare the number of cracks with the crack number threshold. If the number of cracks is equal to or greater than the crack number threshold, repair or replace the cracks; if the number of cracks is less than the crack number threshold, perform a secondary judgment. Step 4: Set a crack number approaching threshold. In the secondary judgment, compare the crack number with the crack number approaching threshold. If the crack number is less than the crack number approaching threshold, the steel structure is in normal use. If the number of cracks is less than the threshold, the number of cracks is first graded, and then the damage detection and identification time interval is dynamically adjusted according to the level of the number of cracks.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement a steel structure damage identification system based on deep learning as described in any one of claims 1 to 8.

Images