Inspection management system based on machine vision and artificial intelligence
By using an inspection and management system based on machine vision and artificial intelligence, the damage status of steel structures can be accurately determined, solving the problems of inefficiency and misjudgment in traditional detection methods, and achieving efficient and comprehensive damage assessment and early warning.
Patent Information
- Application Number
- CN202511838078.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-08
- Publication Date
- 2026-03-27
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Traditional inspection methods are time-consuming and labor-intensive, difficult to cover all areas, prone to missed detections, and subject to subjective bias in human assessment, making it difficult to identify minute cracks and damage in hidden areas, thus failing to comprehensively assess the overall safety of steel structures.
An inspection and management system based on machine vision and artificial intelligence is adopted. The risk analysis module extracts the structural distribution characteristics and connection node information of the steel structure area, calculates the degree of damage risk, and combines the inspection and detection module to identify damage characteristics and the assessment module to analyze the severity of damage, so as to achieve accurate judgment and early warning.
It improves the targeting and efficiency of damage detection, comprehensively assesses the impact of damage, ensures structural safety, reduces misjudgments and missed detections, and enhances the accuracy and comprehensiveness of detection.
Smart Images

Figure CN121746455A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the field of inspection management, and in particular to an inspection management system based on machine vision and artificial intelligence. BACKGROUND
[0002] As the core hub of power transmission and distribution in the power system, the stable operation of the substation is directly related to the power supply reliability and power safety of the regional power grid. Inside the substation, there are a large number of steel structures supporting power equipment, busbars, and frameworks, including equipment supports, disconnecting switch frameworks, main transformer foundation frameworks, cable trench supports, etc. These steel structures are the key bearing structures that ensure the stable installation and safe operation of power equipment.
[0003] The steel structures not only need to bear the weight load of the power equipment itself, but also need to adapt to the dynamic influences such as vibration and temperature change generated during equipment operation. The structural integrity and mechanical stability directly determine the installation precision and operation safety of the power equipment. Once the steel structure is damaged, it may cause equipment tilt, displacement, and even serious power accidents such as line short circuit and equipment failure, resulting in large-scale power outages and causing significant losses to industrial production, residents' lives, and the social economy.
[0004] In addition, the steel structures are mostly modularly designed and contain a large number of welded joints, bolted joints, and corner structures, and are exposed to the outdoor or semi-outdoor environment for a long time, facing multiple environmental erosions such as ultraviolet radiation and humidity changes. At the same time, some steel structures need to bear movable parts, such as the movable contact connection structure of the disconnecting switch and the rotatable support for equipment maintenance, and the connecting joints are prone to fatigue damage due to repeated friction and stress alternation. These factors all pose strict requirements on the long-term service performance of the steel structures. Therefore, timely and accurate detection and evaluation of the damage state of the steel structures of the power equipment in the substation are important foundations for ensuring the safe and stable operation of the substation and prolonging the service life of the equipment.
[0005] Chinese Patent Application Publication No. CN118398143A discloses a steel structure member damage evaluation method and system, which includes: acquiring a stress data sequence of a steel structure member to be detected; screening all structural regions according to non-structural abnormal quantities to obtain all suspected stress concentration damage structural regions; acquiring a confidence factor of each suspected stress concentration damage structural region; acquiring a steel structure determination factor of each suspected stress concentration damage structural region; acquiring updated stress data of each suspected stress concentration damage structural region; acquiring a final abnormal value of each structural region according to the updated stress data; and performing stress concentration damage detection on the steel structure member to be detected according to the final abnormal value of the structural region. The application improves the accuracy of the stress concentration damage detection result of the steel structure member.
[0006] However, the prior art also has the following problems, Traditional detection is mostly observed by relevant staff through naked eyes, with the help of simple tools, or by using ultrasonic waves, magnetic powder detection and other local detection methods, which requires a lot of time and manpower, and it is difficult to cover all areas, and it is easy to miss detection. Moreover, artificial experience judgment is relied on, and there is subjective deviation in the evaluation of damage degree, and the results of different detection personnel may differ greatly. At the same time, it is difficult for the naked eye to identify tiny cracks and damage in hidden areas, which leads to early damage being ignored. In addition, more attention is paid to the identification of single damage points, and the overall steel structure is not comprehensively evaluated, which leads to an incomplete judgment of the overall safety of the structure, reducing the accuracy and efficiency of detection. SUMMARY
[0007] Therefore, the present application provides a kind of based on machine vision and artificial intelligence's inspection management system to overcome the prior art, traditional detection consumes a lot of time and manpower, and it is difficult to cover all areas, and it is easy to miss detection, and artificial experience judgment is relied on, and there is subjective deviation in the evaluation of damage degree, and the results of different detection personnel may differ greatly. At the same time, it is difficult for the naked eye to identify tiny cracks and damage in hidden areas, which leads to early damage being ignored. In addition, more attention is paid to the identification of single damage points, and the overall steel structure is not comprehensively evaluated, which leads to an incomplete judgment of the overall safety of the structure, reducing the accuracy and efficiency of detection.
[0008] To achieve the above purpose, the present application provides a kind of based on machine vision and artificial intelligence's inspection management system, it includes: A risk analysis module is used to call the corresponding steel structure data of the power equipment, to extract the structure distribution characteristics of the corresponding steel structure region, and to calculate the damage risk degree representation value of the steel structure region by combining the structure distribution characteristics and the number of movable components connected by the connecting node, to mark the steel structure region. An inspection detection module is connected with the risk analysis module, and is used to detect and analyze the damage of the steel structure region based on the marking results of each steel structure region, including, Respectively identify the exposed area ratio of the steel structure region and the surface deformation of the corner region to determine whether the surface state reaches the damage destruction stage, and analyze the damage severity representation parameter of the steel structure region based on the detail state characteristics corresponding to the damage destruction stage to determine whether to issue a warning prompt signal; A damage evaluation module is connected with the inspection detection module, and responds to the warning prompt signal issued by the inspection detection module to evaluate the damage of the steel structure region, including, Determine the structural part corresponding to the direction of maximum principal stress, call the surface wrinkle width and wrinkle height difference corresponding to the structural part, and combine it with the average proportion of adjacent exposed area of adjacent steel structure area in the region where the structural part is located to evaluate the damage interference degree characterization coefficient, determine the corresponding damage interference level, and set the interference degree label. The structural distribution features include the number of bends and the average bend radius, while the detailed state features include the extension length of cracks and the color variation of exposed areas.
[0009] Furthermore, the risk analysis module is used to calculate the damage risk level characterization value of the steel structure area, including: The sum of the ratio of the number of bends to the bend number threshold and the ratio of the mean bend radius threshold to the mean bend radius is used as the first damage risk feature; The ratio of the number of movable components connected to a connection node to a threshold number of movable components is used as the second damage risk characteristic. The first damage risk feature and the second damage risk feature are weighted and summed to determine the damage risk level characterization value.
[0010] Furthermore, the risk analysis module is used to mark the steel structure area, including: If the damage risk level characterization value of the steel structure area is greater than or equal to the damage risk level characterization threshold, then the steel structure area is marked.
[0011] Furthermore, the inspection and detection module is used to perform inspection and damage detection analysis on the steel structure area, including: If any steel structure area is marked, then inspection damage detection analysis is performed on the steel structure area.
[0012] Furthermore, the inspection module is used to determine whether the surface condition has reached the damage stage, including: If the exposed area of the steel structure region is greater than the exposed area percentage threshold and the surface deformation of the corner region is greater than the surface deformation threshold, then the surface condition is determined to have reached the damage and failure stage.
[0013] Furthermore, the inspection and detection module is used to analyze the damage severity characterization parameters of the steel structure area, including: The ratio of the crack extension length to the extension length threshold is used as the first severity feature; The ratio of the color change amplitude of the exposed area to the color change amplitude threshold is used as the second severity feature; The sum of the first severity feature and the second severity feature is used as the parameter representing the severity of the injury.
[0014] Furthermore, the inspection and detection module is used to determine whether to issue a warning signal, including: If the damage severity characterization parameter of the steel structure area is greater than or equal to the damage severity characterization parameter threshold, a warning signal will be issued.
[0015] Furthermore, the damage assessment module is used to assess the damage interference degree characterization coefficient, including: The sum of the ratio of the surface wrinkle width to the surface wrinkle width threshold and the ratio of the wrinkle height difference to the wrinkle height difference threshold corresponding to the structural part is used as the first damage interference feature. The ratio of the average proportion of adjacent exposed areas in adjacent steel structure regions to the threshold of the average proportion of adjacent exposed areas is used as the second damage interference feature. The first damage interference feature and the second damage interference feature are weighted and summed to determine the damage interference degree characterization coefficient.
[0016] Furthermore, the damage assessment module is used to determine the corresponding damage interference level, including: Used to pre-set the correspondence between the damage interference level and the range of damage interference degree characterization coefficients; Used to determine the range of damage interference degree characterization coefficients to which the damage interference degree characterization coefficients corresponding to the steel structure region belong; The damage interference level corresponds one-to-one with the range of damage interference degree characterization coefficients.
[0017] Furthermore, the damage assessment module is used to set interference level labels, including: The interference level label corresponds one-to-one with the damage interference level.
[0018] Compared with existing technologies, this invention includes a risk analysis module, which calls upon the steel structure data corresponding to the power equipment to extract the structural distribution characteristics of several corresponding steel structure areas. Combining these structural distribution characteristics with the number of movable components connected to the connection nodes, it calculates the damage risk level characterization value of the steel structure areas for marking. An inspection and detection module, connected to the risk analysis module, performs inspection and damage detection analysis on the steel structure areas based on the marking results. A damage assessment module, connected to the inspection and detection module, assesses the damage to the steel structure areas in response to early warning signals issued by the inspection and detection module. This invention accurately determines the damage state of the steel structure surface, comprehensively assesses the impact of damage, ensures structural safety, and improves the targeting and efficiency of damage detection.
[0019] In particular, this invention includes a risk analysis module to identify high-risk structural weak areas, considering key mechanical indicators of the steel structure. Specifically, the more bends a steel structure has, the denser the stress concentration points; the smaller the bend radius (i.e., the sharper the angle), the more pronounced the local stress concentration. These types of areas are more prone to fatigue damage or cracking under load. The risk analysis module extracts these characteristics to identify potential weak areas caused by structural morphology. Furthermore, the number of movable components connected to a node reflects the frequency and intensity of dynamic loads borne by that area. The more movable components, the more frequent the wear and stress alternation at the node, and the higher the risk of damage. This invention combines this dynamic factor with static structural characteristics, enabling damage risk assessment to consider both "inherent structural weaknesses" and "external load effects," significantly improving the comprehensiveness and accuracy of risk prediction. Therefore, this invention characterizes the degree of damage risk of the steel structure corresponding to the steel structure area by calculating the damage risk level characterization value of the steel structure area, providing data support for subsequent marking of the steel structure area. This invention accurately determines the damage state of the steel structure surface, thereby comprehensively assessing the damage impact, ensuring structural safety, and improving the pertinence and efficiency of damage detection.
[0020] In particular, this invention incorporates a patrol inspection module to focus on high-risk steel structure areas, avoiding ineffective scanning of low-risk areas. It employs a "screen first, focus later" approach, concentrating inspection resources on potentially damaged structural parts. By using the percentage of exposed area in the steel structure and the surface deformation of corner areas, a dual-indicator system is used to determine whether the surface condition has reached the damage stage. Considering that exposed areas are prone to corrosion, a significant contributing factor to damage, the percentage of exposed area reflects the degree of damage to the anti-corrosion layer of the steel structure; the deformation of corner areas reflects the physical deformation of the structure caused by stress. The combination of these two indicators provides cross-validation for the damage stage, avoiding misjudgments that might occur with a single indicator and ensuring the accuracy of damage stage determination. Furthermore, this invention extracts the crack extension length and the color change amplitude of the exposed area as detailed state features, transforming them into quantifiable damage severity characterization parameters. The crack extension length directly reflects the damage propagation trend; the color change amplitude of the exposed area indirectly reflects the cumulative effect of damage over time. Based on these damage severity characterization parameters, the physical damage degree of the steel structure area is characterized, and the impact of the damage on the steel structure's function is quantified, reflecting the urgency of damage intervention and providing data support for subsequent determination of whether to issue an early warning signal. This invention accurately determines the damage state of the steel structure surface, comprehensively assesses the damage impact, ensures structural safety, and improves the targeting and efficiency of damage detection.
[0021] In particular, this invention includes a damage assessment module that activates a damage assessment mechanism for the corresponding steel structure area upon receiving an early warning signal from the inspection module. The degree of damage interference in the steel structure area is assessed by focusing on the wrinkle characteristics of the corresponding structural part along the direction of maximum principal stress and considering the exposed condition of adjacent steel structure areas. Focusing on the critical stress-bearing areas of the steel structure, i.e., the structural parts along the direction of maximum principal stress, is the key node for the steel structure's load-bearing capacity. Correspondingly, damage in this area has a far greater impact on the overall structural stability than in other areas. By analyzing the width and height difference of the surface wrinkles in this structural part, the core points of damage's impact on the structural mechanical performance are identified. The wrinkle width reflects the lateral spread of damage, while the height difference reflects the severity of structural deformation; the combination of these two factors can be used to determine the extent of damage to the load-bearing capacity of critical parts. Furthermore, the proportion of adjacent exposed areas in adjacent steel structure areas is taken into consideration to fully account for the spread of damage in the steel structure. Since exposed areas are prone to rust spread, a higher proportion of exposed area in adjacent steel structure areas indicates a greater risk of damage spreading to the surrounding areas. This invention comprehensively presents the interference range of damage on the entire steel structure through a two-dimensional analysis of local damage and diffusion trends. Based on the assessment of the above situation, a damage interference degree characterization coefficient is determined to characterize the direct damage degree of the damage to key stress-bearing parts, and the diffusion interference effect of damage to adjacent sub-regions is quantified, providing data support for subsequent determination of the corresponding damage interference level and interference degree label. This invention accurately determines the damage state of the steel structure surface, thereby comprehensively assessing the damage impact, ensuring structural safety, and improving the pertinence and efficiency of damage detection. Attached Figure Description
[0022] Figure 1 This is a functional block diagram of the inspection management system based on machine vision and artificial intelligence, which is an embodiment of the invention. Figure 2 This is a logic decision diagram for marking the steel structure area in an embodiment of the invention. Figure 3 This is a logic diagram for determining whether the surface condition has reached the damage stage in an embodiment of the invention. Figure 4 This is a logic diagram for determining whether to issue a warning signal in an embodiment of the invention. Detailed Implementation
[0023] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.
[0024] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.
[0025] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.
[0026] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.
[0027] Please see Figure 1 The diagram shown is a functional block diagram of the inspection management system based on machine vision and artificial intelligence according to an embodiment of the present invention. The inspection management system based on machine vision and artificial intelligence according to an embodiment of the present invention includes: The risk analysis module is used to call the steel structure data corresponding to the power equipment to extract the structural distribution characteristics of several corresponding steel structure areas. Combining the structural distribution characteristics and the number of movable components connected to the connection nodes, it calculates the damage risk level characterization value of the steel structure area to mark the steel structure area. The inspection and detection module, connected to the risk analysis module, is used to perform inspection damage detection and analysis on the steel structure areas based on the marking results for each steel structure area, including... The proportion of exposed area in the steel structure region and the surface deformation of the corner region are identified respectively to determine whether the surface condition has reached the damage and failure stage. Based on the detailed state characteristics corresponding to the damage and failure stage, the damage severity characterization parameters of the steel structure region are analyzed to determine whether to issue a warning signal. A damage assessment module, connected to the inspection and detection module, performs a damage assessment of the steel structure area in response to an early warning signal issued by the inspection and detection module, including: Determine the structural part corresponding to the direction of maximum principal stress, call the surface wrinkle width and wrinkle height difference corresponding to the structural part, and combine it with the average proportion of adjacent exposed area of adjacent steel structure area in the region where the structural part is located to evaluate the damage interference degree characterization coefficient, determine the corresponding damage interference level, and set the interference degree label. The structural distribution features include the number of bends and the average bend radius, while the detailed state features include the extension length of cracks and the color variation of exposed areas.
[0028] Specifically, the structural data includes structural distribution characteristics, the number of movable components connected to the connection nodes, the proportion of exposed area in the steel structure area, the surface deformation of the corner area, detailed state characteristics, the surface wrinkle width and wrinkle height difference corresponding to the structural parts, and the average proportion of adjacent exposed area of adjacent steel structure areas in the area where the structural parts are located.
[0029] During implementation, sample models of steel structures corresponding to several power equipment in the substation and data on the steel structures themselves are stored in advance, including structural distribution characteristics and the number of movable components connected to the connection nodes. By using high-definition cameras, drones and other equipment to collect images of the steel structure surface, and by using image processing algorithms, including edge detection and feature extraction, relevant damage features in the images can be quickly identified, such as the difference in the width and height of surface wrinkles corresponding to structural parts. By comparing with the sample model, the proportion of exposed area in the steel structure region, the adjacent exposed area of the adjacent steel structure region in the region where the structural part is located, and the surface deformation of the corner region are determined. Furthermore, pixel-level analysis is used to capture micro-cracks and determine the crack extension length; The brightness of the exposed area is determined by analyzing the full spectrum of reflection from the surface of the steel structure using a color analyzer. This brightness is then compared with the brightness of the corresponding area in the sample model to determine the range of color change in the exposed area. This will not be elaborated further.
[0030] Specifically, there are no specific limitations on the method for determining the structural part corresponding to the direction of the maximum principal stress; it can be locked by combining mechanical simulation with feature localization. In some possible implementations, the first step is to obtain a three-dimensional model of the steel structure, its material parameters, and its actual stress conditions. Using finite element analysis tools, such as ANSYS and ABAQUS, stress simulation calculations are performed on the steel structure, and an overall stress cloud map is output. The area with the darkest color or the largest value in the cloud map is the high stress concentration area. By tracing along the principal stress direction, the specific component segment or node with the largest stress value that is consistent with the stress transmission path of the structure is found. For example, corner joints, the periphery of bolted joints, and the connection section of movable components are the structural parts corresponding to the direction of the maximum principal stress. The finite element analysis tool directly outputs the principal stress direction data of the high stress concentration region.
[0031] Specifically, there are no restrictions on the specific structure of the risk analysis module, the inspection and detection module, and the damage assessment module. Each module or its units can be composed of logic components or combinations of logic components. Logic components include field-programmable processors, computers, or microprocessors in computers.
[0032] Specifically, the risk analysis module is used to calculate the damage risk level characterization value of the steel structure area, including: The sum of the ratio of the number of bends to the bend number threshold and the ratio of the mean bend radius threshold to the mean bend radius is used as the first damage risk feature; The ratio of the number of movable components connected to a connection node to a threshold number of movable components is used as the second damage risk characteristic. The first damage risk feature and the second damage risk feature are weighted and summed to determine the damage risk level characterization value.
[0033] Specifically, in practice, the damage risk of steel structures primarily stems from the mechanical defects inherent in their structural form. The number and radius of these bends are core inherent characteristics that determine the local stress state of the structure. A greater number of bends means more stress concentration points on the structural surface. Under load, the local stress at these stress concentration points is much higher than in other smooth areas, making them highly susceptible to cracking or deformation due to fatigue effects—they are "naturally high-risk areas" for damage. Conversely, a smaller bend radius results in a more significant stress concentration effect. Theoretically, the local stress at sharp bends can far exceed the material's yield strength, directly leading to early failure of the steel structure. Their impact on damage risk is fundamental and decisive. These two characteristics directly reflect the inherent weakness of the structure and are the main internal factors contributing to damage. In contrast, the number of movable components connected by joint nodes reflects more the dynamic influence of external loads, such as vibration and friction, and is an external triggering factor for damage. The inherent weakness of the steel structure is a prerequisite for external loads to cause damage. If the structure itself has no obvious stress concentration areas, the risk of damage will be significantly reduced even with frequent dynamic loads. Therefore, in implementation, the structural distribution characteristics that are the internal causes of damage to the steel structure, namely the number of bends and the average bend radius, are assigned a slightly higher weight to the first damage risk characteristic. Then, when performing weighted summation, the weight of the first damage risk characteristic is set to 0.6, and the weight of the second damage risk characteristic is set to 0.4.
[0034] In this embodiment, the purpose of setting thresholds for the number of bends, the average bend radius, and the number of movable components is to characterize situations where the steel structure corresponding to the steel structure area has a high degree of damage risk. By acquiring steel structure data corresponding to several power devices, calling the data on the number of bends, the average bend radius, and the number of movable components connected to the connection nodes, the average value of the number of bends, the average value of the average bend radius, and the average value of the number of movable components are calculated. Based on the purpose of setting the above three thresholds, the average value of the number of bends is used as the threshold for the number of bends, the average value of the average bend radius is used as the threshold for the average bend radius, and the average value of the number of movable components is used as the threshold for the number of movable components.
[0035] Specifically, this invention employs a risk analysis module to identify high-risk structural weak areas, considering key mechanical indicators of the steel structure. The more bends a steel structure has, the denser the stress concentration points; the smaller the bend radius, i.e., the sharper the angle, the more pronounced the local stress concentration. These types of areas are more prone to fatigue damage or cracking under load. The risk analysis module extracts these characteristics to identify potential weak areas caused by structural morphology. Furthermore, the number of movable components connected to a node reflects the frequency and intensity of dynamic loads borne by that area. More movable components mean more frequent wear and stress alternation at the node, resulting in a higher risk of damage. This invention combines this dynamic factor with static structural characteristics, enabling damage risk assessment to consider both "inherent structural weaknesses" and "external load effects," significantly improving the comprehensiveness and accuracy of risk prediction. Therefore, this invention characterizes the degree of damage risk of the steel structure corresponding to the steel structure area by calculating the damage risk level characterization value of the steel structure area, providing data support for subsequent marking of the steel structure area. This invention accurately determines the damage state of the steel structure surface, thereby comprehensively assessing the damage impact, ensuring structural safety, and improving the pertinence and efficiency of damage detection.
[0036] Specifically, please refer to Figure 2 As shown, this is a logic decision diagram for marking steel structure areas according to an embodiment of the present invention. The risk analysis module is used to mark the steel structure areas, including: If the damage risk level characterization value of the steel structure area is greater than or equal to the damage risk level characterization threshold, then the steel structure area is marked. If the damage risk level characterization value of the steel structure area is less than the damage risk level characterization threshold, then there is no need to mark the steel structure area.
[0037] The damage risk level characterization threshold is predetermined. The damage risk level characterization value calculated is determined when the number of bends is equal to the bend number threshold, the average bend radius threshold is equal to the average bend radius threshold, and the number of movable components connected to the connecting node is equal to the movable component number threshold.
[0038] Specifically, the inspection and detection module is used to perform inspection and damage detection analysis on the steel structure area, including: If any steel structure area is marked, then inspection damage detection analysis is performed on the steel structure area.
[0039] Specifically, please refer to Figure 3 As shown, this is a logic diagram for determining whether the surface condition has reached the damage stage according to an embodiment of the present invention. The inspection module is used to determine whether the surface condition has reached the damage stage, including: If the exposed area of the steel structure region is greater than the exposed area percentage threshold and the surface deformation of the corner region is greater than the surface deformation threshold, then the surface condition is determined to have reached the damage and failure stage.
[0040] In this embodiment, the purpose of setting the exposed area percentage threshold and the surface deformation threshold is to characterize the situation where the damage to the steel structure area is relatively severe. By acquiring steel structure data corresponding to several power devices, calling the exposed area data of the steel structure area and the surface deformation data of the corner area, the average exposed area percentage and the average surface deformation are calculated and used as the benchmark values under normal conditions. Based on the purpose of setting the above two thresholds, the exposed area percentage threshold is determined to be the product of the average exposed area percentage and the exposed deviation coefficient, and the surface deformation threshold is determined to be the product of the average surface deformation and the deformation deviation coefficient. The exposed deviation coefficient is selected in the interval [1.15, 1.2], with 1.15 preferred in implementation, and the deformation deviation coefficient is selected in the interval [1.2, 1.3], with 1.2 preferred in implementation.
[0041] Specifically, this invention incorporates an inspection module that prioritizes high-risk steel structure areas, avoiding ineffective scanning of low-risk areas. It employs a "screen first, focus later" approach, concentrating inspection resources—such as machine vision image acquisition and analysis capabilities—on potentially damaged structural components. The invention uses a dual-indicator approach—the percentage of exposed area in the steel structure and the surface deformation of corner areas—to determine whether the surface condition has reached the damage / failure stage. Exposed areas are prone to corrosion, a significant contributing factor to damage; the percentage of exposed area reflects the degree of damage to the anti-corrosion layer. The deformation of corner areas reflects the physical deformation caused by stress; excessive deformation may indicate structural failure. This combination of indicators provides cross-validation of the damage / failure stage, avoiding misjudgments that might arise from a single indicator and ensuring the accuracy of damage / failure stage determination.
[0042] Specifically, the inspection and detection module is used to analyze the damage severity characterization parameters of the steel structure area, including: The ratio of the crack extension length to the extension length threshold is used as the first severity feature; The ratio of the color change amplitude of the exposed area to the color change amplitude threshold is used as the second severity feature; The sum of the first severity feature and the second severity feature is used as the parameter representing the severity of the injury.
[0043] In this embodiment, the purpose of setting the extension length threshold and the color change amplitude threshold is to characterize situations where the physical damage to the steel structure area is severe, the damage has a significant impact on the function of the steel structure, and the urgency of intervention is high. By acquiring steel structure data corresponding to several power devices, calling the extension length data of cracks and the color change amplitude data of exposed areas, the mean extension length and the mean color change amplitude are calculated and used as the benchmark values under normal conditions. Based on the purpose of setting the above two thresholds, the extension length threshold is determined as the product of the mean extension length and the extension deviation coefficient, and the color change amplitude threshold is determined as the product of the mean color change amplitude and the color deviation coefficient. The extension deviation coefficient is selected within the interval [1.1, 1.15], with 1.1 being preferred in practice, and the color deviation coefficient is selected within the interval [1.15, 1.2], with 1.15 being preferred in practice.
[0044] Specifically, this invention extracts the extension length of cracks and the color change amplitude of exposed areas as detailed state features, transforming them into quantifiable damage severity characterization parameters. The extension length of cracks directly reflects the damage propagation trend; the longer the crack extension, the more significant the decrease in structural load-bearing capacity. The color change amplitude of exposed areas, such as color deepening and discoloration range expansion due to corrosion, indirectly reflects the cumulative effect of damage over time, such as the degree of corrosion. Furthermore, based on these damage severity characterization parameters, the physical damage degree of the steel structure area is characterized, and the impact of damage on the steel structure's function is quantified, reflecting the urgency of damage intervention and providing data support for subsequent determination of whether to issue an early warning signal. This invention, by accurately determining the damage state of the steel structure surface, comprehensively assesses the damage impact, ensures structural safety, and improves the targeting and efficiency of damage detection.
[0045] Specifically, please refer to Figure 4 The diagram shown illustrates the logic for determining whether to issue a warning signal according to an embodiment of the present invention. The inspection and detection module is used to determine whether to issue a warning signal, and includes: If the damage severity characterization parameter of the steel structure area is greater than or equal to the damage severity characterization parameter threshold, a warning signal will be issued. If the damage severity characterization parameter of the steel structure area is less than the damage severity characterization parameter threshold, then it is determined that no warning signal needs to be issued.
[0046] The threshold value for the damage severity characterization parameter is predetermined. The damage severity characterization parameter is determined by calculating the damage severity characterization parameter when the extension length of the crack is equal to the extension length threshold value and the color change amplitude of the exposed area is equal to the color change amplitude threshold value.
[0047] Specifically, the damage assessment module is used to assess the damage interference degree characterization coefficient, including: The sum of the ratio of the surface wrinkle width to the surface wrinkle width threshold and the ratio of the wrinkle height difference to the wrinkle height difference threshold corresponding to the structural part is used as the first damage interference feature. The ratio of the average proportion of adjacent exposed areas in adjacent steel structure regions to the threshold of the average proportion of adjacent exposed areas is used as the second damage interference feature. The first damage interference feature and the second damage interference feature are weighted and summed to determine the damage interference degree characterization coefficient.
[0048] Specifically, structural locations along the direction of maximum principal stress are the core areas of steel structures where stress is most concentrated and have the greatest impact on overall stability. Surface wrinkles in these locations represent direct physical deformations caused by stress. This includes the width of the wrinkles, which directly reflects the lateral spread of damage along the critical stress direction. A wider wrinkle indicates a broader area of localized plastic deformation, leading to more severe damage to material continuity and potentially a sharp decrease in the load-bearing capacity of that location. The difference in wrinkle height reflects the vertical amplitude of structural deformation. A larger height difference indicates more significant stress concentration within the steel structure, possibly approaching or exceeding the material's yield strength, posing a high risk of sudden fracture. These two characteristics are directly related to the core mechanical properties of the structure and are decisive indicators for determining whether the structure can still safely bear loads. In contrast, the average percentage of adjacent exposed areas reflects the potential spread trend of damage more, with a gradual and indirect impact. Therefore, in implementation, a slightly higher weight is given to the first damage interference feature calculated based on the surface wrinkle width and wrinkle height difference corresponding to the structural location. When performing weighted summation, the weight of the first damage interference feature is set to 0.6, and the weight of the second damage interference feature is set to 0.4. In this embodiment, the purpose of setting the threshold for surface wrinkle width, the threshold for wrinkle height difference, and the threshold for the average proportion of adjacent exposed area is to characterize the situation where the damage has a large degree of direct damage to the key stress-bearing parts and the damage has a strong diffusion interference effect to the adjacent sub-regions. By acquiring steel structure data corresponding to several power devices, and calling the surface wrinkle width data, corresponding wrinkle height difference data, and average percentage data of adjacent exposed area of adjacent steel structure regions, the average surface wrinkle width, average wrinkle height difference, and average percentage of adjacent exposed area are calculated, and these are used as the benchmark values under normal conditions. Based on the purpose of setting the above three thresholds, the surface wrinkle width threshold is determined as the product of the average surface wrinkle width and the width deviation coefficient; the wrinkle height difference threshold is determined as the product of the average wrinkle height difference and the height deviation coefficient; and the average percentage of adjacent exposed area threshold is determined as the product of the average percentage of adjacent exposed area and the area deviation coefficient. The width deviation coefficient is selected within the range [1.1, 1.15], preferably 1.1 in practice; the height deviation coefficient is selected within the range [1.05, 1.1], preferably 1.05 in practice; and the area deviation coefficient is selected within the range [1.1, 1.15], preferably 1.1 in practice.
[0049] Specifically, this invention includes a damage assessment module that activates a damage assessment mechanism for the corresponding steel structure area upon receiving an early warning signal from the inspection module. The degree of damage interference in the steel structure area is assessed by focusing on the wrinkle characteristics of the corresponding structural part along the direction of maximum principal stress and considering the exposed condition of adjacent steel structure areas. Focusing on the critical stress-bearing areas of the steel structure, i.e., the structural parts along the direction of maximum principal stress, is the key node for the steel structure's load-bearing capacity. Correspondingly, damage in this area has a far greater impact on the overall structural stability than in other areas. By analyzing the width and height difference of the surface wrinkles in this structural part, the core points of impact of damage on the structural mechanical performance are identified. The wrinkle width reflects the lateral spread of damage, while the height difference reflects the severity of structural deformation; the combination of these two factors can be used to determine the extent of damage to the load-bearing capacity of critical parts. Furthermore, the proportion of adjacent exposed areas in adjacent steel structure areas is taken into consideration to fully account for the diffusion of damage in the steel structure. Since exposed areas are prone to rust spread, a higher proportion of exposed area in adjacent steel structure areas indicates a greater risk of damage spreading to the surrounding areas. For example, rust may accelerate the failure of the anti-corrosion layer in adjacent areas, leading to cascading damage. This invention comprehensively presents the interference range of damage on the entire steel structure through a two-dimensional analysis of local damage and diffusion trends. Based on the assessment of the above situation, a damage interference degree characterization coefficient is determined to characterize the direct damage degree of the damage to key stress-bearing parts, and the diffusion interference effect of damage to adjacent sub-regions is quantified, providing data support for subsequent determination of the corresponding damage interference level and interference degree label. This invention accurately determines the damage state of the steel structure surface, thereby comprehensively assessing the damage impact, ensuring structural safety, and improving the pertinence and efficiency of damage detection.
[0050] Specifically, the damage assessment module is used to determine the corresponding damage interference level, including: Used to pre-set the correspondence between the damage interference level and the range of damage interference degree characterization coefficients; Used to determine the range of damage interference degree characterization coefficients to which the damage interference degree characterization coefficients corresponding to the steel structure region belong; The damage interference level corresponds one-to-one with the range of damage interference degree characterization coefficients.
[0051] In this embodiment, the corresponding damage interference level is determined as follows: Based on the damage interference degree characterization coefficient, it is divided into three preset intervals, and three levels of damage interference are set simultaneously; If the damage interference degree characterization coefficient corresponding to the steel structure area is within the first preset interval (0, S0], then the first damage interference level 1 is set. If the damage interference degree characterization coefficient corresponding to the steel structure area is within the second preset interval (S0, 1.2S0], then the second damage interference level 2 is set. If the damage interference degree characterization coefficient corresponding to the steel structure area is within the third preset interval (1.2S0, +∞), then the third damage interference level 3 is set.
[0052] The damage interference degree characterization coefficient threshold S0 is predetermined. The damage interference degree characterization coefficient is determined by calculating the following conditions: the surface wrinkle width corresponding to the structural part is equal to the surface wrinkle width threshold, the wrinkle height difference is equal to the wrinkle height difference threshold, and the average proportion of adjacent exposed area of adjacent steel structure area is equal to the average proportion of adjacent exposed area threshold.
[0053] Specifically, the damage assessment module is used to set interference level labels, including: The interference level label corresponds one-to-one with the damage interference level.
[0054] In this embodiment, the interference level label is set in the following manner: The first level of damage interference, grade 1, corresponds to a yellow label; The second level of damage interference, 2, is correspondingly marked with an orange label; The third level of damage interference, 3, is marked with a red label.
[0055] Specifically, substations contain numerous and dispersed steel structures, generating a large amount of inspection data and assessment results after each inspection. Presenting these data solely in text or numerical form requires meticulous verification of parameters to prioritize risks, a time-consuming and labor-intensive process. Color-coded labels, through visual contrast, directly highlight the severity of risks in steel structure areas, allowing management personnel to quickly focus on key issues without sifting through massive amounts of data, significantly improving the efficiency of summarizing and reviewing inspection results.
[0056] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.
Claims
1. An inspection management system based on machine vision and artificial intelligence, characterized in that, include: The risk analysis module is used to call the steel structure data corresponding to the power equipment to extract the structural distribution characteristics of several corresponding steel structure areas. Combining the structural distribution characteristics and the number of movable components connected to the connection nodes, it calculates the damage risk level characterization value of the steel structure area to mark the steel structure area. The inspection and detection module is used to perform inspection damage detection and analysis on the steel structure areas based on the marking results for each steel structure area, including: The proportion of exposed area in the steel structure region and the surface deformation of the corner region are identified respectively to determine whether the surface condition has reached the damage and failure stage. Based on the detailed state characteristics corresponding to the damage and failure stage, the damage severity characterization parameters of the steel structure region are analyzed to determine whether to issue a warning signal. The damage assessment module, in response to an early warning signal issued by the inspection and detection module, performs a damage assessment on the steel structure area. include, Determine the structural part corresponding to the direction of maximum principal stress, call the surface wrinkle width and wrinkle height difference corresponding to the structural part, and combine it with the average proportion of adjacent exposed area of adjacent steel structure area in the region where the structural part is located to evaluate the damage interference degree characterization coefficient, determine the corresponding damage interference level, and set the interference degree label. The structural distribution features include the number of bends and the average bend radius, while the detailed state features include the extension length of cracks and the color variation of exposed areas.
2. The inspection management system based on machine vision and artificial intelligence according to claim 1, characterized in that, The risk analysis module is used to calculate the damage risk level characterization value of the steel structure area, including: The sum of the ratio of the number of bends to the bend number threshold and the ratio of the mean bend radius threshold to the mean bend radius is used as the first damage risk feature; The ratio of the number of movable components connected to a connection node to a threshold number of movable components is used as the second damage risk characteristic. The first damage risk feature and the second damage risk feature are weighted and summed to determine the damage risk level characterization value.
3. The inspection management system based on machine vision and artificial intelligence according to claim 2, characterized in that, The risk analysis module is used to mark the steel structure area, including: If the damage risk level characterization value of the steel structure area is greater than or equal to the damage risk level characterization threshold, then the steel structure area is marked.
4. The inspection management system based on machine vision and artificial intelligence according to claim 3, characterized in that, The inspection and detection module is used to perform inspection and damage detection analysis on the steel structure area, including: If any steel structure area is marked, then inspection damage detection analysis is performed on the steel structure area.
5. The inspection management system based on machine vision and artificial intelligence according to claim 1, characterized in that, The inspection module is used to determine whether the surface condition has reached the damage stage, including: If the exposed area of the steel structure region is greater than the exposed area percentage threshold and the surface deformation of the corner region is greater than the surface deformation threshold, then the surface condition is determined to have reached the damage and failure stage.
6. The inspection management system based on machine vision and artificial intelligence according to claim 1, characterized in that, The inspection and detection module is used to analyze the damage severity characterization parameters of the steel structure area, including: The ratio of the crack extension length to the extension length threshold is used as the first severity feature; The ratio of the color change amplitude of the exposed area to the color change amplitude threshold is used as the second severity feature; The sum of the first severity feature and the second severity feature is used as the parameter representing the severity of the injury.
7. The inspection management system based on machine vision and artificial intelligence according to claim 6, characterized in that, The inspection and detection module is used to determine whether to issue an early warning signal, including: If the damage severity characterization parameter of the steel structure area is greater than or equal to the damage severity characterization parameter threshold, a warning signal will be issued.
8. The inspection management system based on machine vision and artificial intelligence according to claim 1, characterized in that, The damage assessment module is used to evaluate the damage interference degree characterization coefficient, including: The sum of the ratio of the surface wrinkle width to the surface wrinkle width threshold and the ratio of the wrinkle height difference to the wrinkle height difference threshold corresponding to the structural part is used as the first damage interference feature. The ratio of the average proportion of adjacent exposed areas in adjacent steel structure regions to the threshold of the average proportion of adjacent exposed areas is used as the second damage interference feature. The first damage interference feature and the second damage interference feature are weighted and summed to determine the damage interference degree characterization coefficient.
9. The inspection management system based on machine vision and artificial intelligence according to claim 8, characterized in that, The damage assessment module is used to determine the corresponding damage interference level, including: Used to pre-set the correspondence between the damage interference level and the range of damage interference degree characterization coefficients; Used to determine the range of damage interference degree characterization coefficients to which the damage interference degree characterization coefficients corresponding to the steel structure region belong; The damage interference level corresponds one-to-one with the range of damage interference degree characterization coefficients.
10. The inspection management system based on machine vision and artificial intelligence according to claim 9, characterized in that, The damage assessment module is used to set interference level labels, including: The interference level label corresponds one-to-one with the damage interference level.
Citation Information
Patent Citations
Steel structural member damage evaluation method and system
CN118398143A
