Corrosion detection method and device, electronic equipment and storage medium

Abstract

The invention provides a corrosion detection method and device, electronic equipment and a computer readable storage medium, and relates to the technical field of nondestructive testing. The corrosion detection method comprises the steps that the corrosion state and the corrosion weight of a target component at the current moment are obtained, and the target component refers to all components of a target object; based on the corrosion state and the corrosion weight, the corrosion risk of the target part at the current moment is evaluated; and based on the corrosion risk of the target component at the current moment, a corrosion component and a corresponding corrosion diffusion component are determined, the corrosion component refers to the target component having a corrosion phenomenon at the current moment, and the corrosion diffusion component refers to the target component to which the corrosion phenomenon diffuses in a future preset time period. The problem that in the prior art, the accuracy and efficiency of a mode for achieving corrosion risk detection are not high is at least solved. The method is suitable for rust detection and processing scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of non-destructive testing, and particularly to a rust detection method, device, electronic device and computer-readable storage medium. Background Art

[0002] The existing methods for realizing rust risk detection include: visual inspection, ultrasonic inspection, electrochemistry / resistance method, and manual marking of rust areas.

[0003] However, the existing methods for realizing rust risk detection have the following technical problems: ① The electrochemistry / resistance method relies on physical contact, and the electrodes are easily worn; ② The efficiency of manual inspection is low, and it is difficult to detect internal rust; ③ The reliability of visual inspection and ultrasonic inspection is poor, and the false alarm rate is high.

[0004] In summary, the existing methods for realizing rust risk detection in the prior art are not high in accuracy and efficiency. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a rust detection method, device, electronic device and computer-readable storage medium for the above-mentioned deficiencies of the prior art. This method can improve the accuracy of rust risk assessment and determination of rusty components and their corresponding rust diffusion components, thereby improving the accuracy and efficiency of rust detection and realizing reliable and rapid rust detection.

[0006] In a first aspect, the present invention provides a rust detection method, including: obtaining the rust state and rust weight of a target component at the current moment, where the target component refers to each component of the target object; evaluating the rust risk of the target component at the current moment based on the rust state and the rust weight; determining the rusty component and its corresponding rust diffusion component based on the rust risk of the target component at the current moment, where the rusty component refers to the target component that has rusted at the current moment, and the rust diffusion component refers to the target component to which the rust phenomenon spreads within a preset time period in the future.

[0007] Preferably, the rust state includes color change, thickness change and environmental humidity. The obtaining of the rust state and rust weight of the target component at the current moment specifically includes: collecting the image, environmental humidity and signal-to-noise ratio of the target component at the current moment; performing color detection on the image at the current moment based on the color space to obtain the color change of the target component at the current moment, where the color space includes at least one of the following: RGB and Lab color spaces; performing thickness detection on the target component based on the ultrasonic detection technology to obtain the thickness change of the target component at the current moment; evaluating the rust weight of the target component at the current moment based on the image and the signal-to-noise ratio.

[0008] Preferably, the rust weight includes a color weight and a thickness weight. Based on the image and the signal-to-noise ratio, the rust weight of the target component at the current moment is evaluated, which specifically includes: evaluating the color confidence and ultrasonic confidence of the target component at the current moment, where the color confidence includes the blurriness and illumination uniformity of the image, and the ultrasonic confidence refers to the signal-to-noise ratio corresponding to the thickness detection of the target component at the current moment; calculating the color weight and thickness weight of the target component at the current moment based on the color confidence and ultrasonic confidence of the target component at the current moment.

[0009] Preferably, based on the rust state and the rust weight, the rust risk of the target component at the current moment is evaluated, which specifically includes:

[0010] According to formula (1), the rust risk of the target component at the current moment is evaluated:

[0011] CRR = w final-color ·ColorScore + w final-ultra ·UltraScore + w humidity ·H (1),

[0012] where CRR represents the rust risk of the target component at the current moment, w final-color represents the color weight of the target component at the current moment, ColorScore represents the color change of the target component at the current moment, w final-ultra represents the thickness weight of the target component at the current moment, UltraScore represents the thickness change of the target component at the current moment, w humidity represents the preset humidity weight, and H represents the environmental humidity of the target component at the current moment.

[0013] Preferably, based on the rust risk of the target component at the current moment, the rusty components and their corresponding rust diffusion components are determined, which specifically includes: generating a spatio-temporal graph convolutional network of the target object based on the rust risk of the target component at the current moment; determining the rusty components and their corresponding rust diffusion components based on the spatio-temporal graph convolutional network of the target object.

[0014] Preferably, the spatio-temporal graph convolutional network includes nodes, node attributes, edges, and weights. Generating a spatio-temporal graph convolutional network of the target object based on the rust risk of the target component at the current moment specifically includes: respectively determining the nodes and edges of the target object with the target component and the physical connection relationship between the target components; determining the weights of the target object with the distance between the target components and the material type corresponding to the physical connection relationship between the target components; evaluating the rust degree and stress value of the target component at the current moment based on the rust risk of the target component at the current moment; determining the node attributes of the target object with the environmental humidity, rust degree, and stress value of the target component at the current moment.

[0015] Preferably, for the spatio-temporal graph convolutional network based on the target object to determine the rusty components and their corresponding rust diffusion components, it specifically includes: determining the nodes whose node attributes meet the rust conditions as the rusty components; predicting the spread speed between the rusty components and other components in the target object based on the material types corresponding to the physical connection relationships between the rusty components and other components in the target object; and determining the rust diffusion components corresponding to the rusty components based on the spread speed and distance between the rusty components and other components in the target object.

[0016] In a second aspect, the present invention further provides a rust detection device, including an acquisition module, an evaluation module, and a determination module. The acquisition module is configured to acquire the rust state and rust weight of a target component at the current moment, where the target component refers to each component of the target object. The evaluation module is connected to the acquisition module and is configured to evaluate the rust risk of the target component at the current moment based on the rust state and the rust weight. The determination module is connected to the evaluation module and is configured to determine the rusty components and their corresponding rust diffusion components based on the rust risk of the target component at the current moment, where the rusty component refers to the target component that has rusted at the current moment, and the rust diffusion component refers to the target component to which the rust phenomenon spreads within a preset time period in the future.

[0017] In a third aspect, the present invention further provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to implement the rust detection method provided in the first aspect above.

[0018] In a fourth aspect, the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the rust detection method provided in the first aspect above.

[0019] A rust detection method, device, electronic device, and computer-readable storage medium provided by the present invention dynamically evaluate the rust risk of a target component through a dynamically changing rust weight, improve the accuracy of rust risk assessment, and then quickly and effectively determine the rusty components and their corresponding rust diffusion components through the dynamically changing rust risk of the target component, improving the accuracy and efficiency of rust detection. Therefore, the present invention can improve the accuracy of rust risk assessment, determination of rusty components and their corresponding rust diffusion components, and further improve the accuracy and efficiency of rust detection, realizing reliable and rapid rust detection. Description of the Drawings

[0020] Figure 1 It is a flowchart of a rust detection method according to Embodiment 1 of the present invention;

[0021] Figure 2Schematic diagram of a rust detection device according to Embodiment 2 of the present invention. Detailed implementation manners

[0022] To enable those skilled in the art to better understand the technical solutions of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0023] It can be understood that the specific embodiments and the accompanying drawings described herein are only for explaining the present invention, rather than limiting the present invention.

[0024] It can be understood that, without conflict, the various embodiments in the present invention and the features in the embodiments can be combined with each other.

[0025] It can be understood that for the convenience of description, only the parts related to the present invention are shown in the drawings of the present invention, and the parts unrelated to the present invention are not shown in the drawings.

[0026] It can be understood that each unit and module involved in the embodiments of the present invention may correspond to only one physical structure, or may be composed of multiple physical structures, or multiple units and modules may also be integrated into one physical structure.

[0027] It can be understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of the present invention may occur in an order different from that marked in the accompanying drawings.

[0028] It can be understood that in the flowcharts and block diagrams of the present invention, the possible architectures, functions, and operations of the systems, devices, equipment, and methods according to the embodiments of the present invention are shown. Among them, each block in the flowchart or block diagram may represent a unit, module, program segment, or code, which contains executable instructions for implementing the specified function. Moreover, each block or combination of blocks in the block diagram and flowchart can be implemented by a hardware-based system for implementing the specified function, or by a combination of hardware and computer instructions.

[0029] It can be understood that the units and modules involved in the embodiments of the present invention can be implemented in software or in hardware. For example, the units and modules can be located in the processor.

[0030] Embodiment 1:

[0031] As Figure 1 shown, this embodiment provides a rust detection method. The rust detection method includes:

[0032] S101. Obtain the rust state and rust weight of the target component at the current moment, where the target component refers to each component of the target object.

[0033] In this embodiment, taking the billboard bracket as the target object, the target components include, but are not limited to: the cross beam, the column, and the connecting bolts of the billboard bracket.

[0034] Specifically, the rust state includes color change, thickness change, and environmental humidity.

[0035] In this embodiment, the rust state refers to a comprehensive description of the rust characteristics or phenomena exhibited by the target component at the current moment, reflecting the changes in the surface and structure of the target component due to the corrosion process. For example: The surfaces of the cross beam, column, and connecting bolts of the billboard bracket are all coated with special color-changing materials. Therefore, when rust occurs on the cross beam, column, and connecting bolts of the billboard bracket (i.e., the chemical substances in the special color-changing material react with the rust products), color changes will occur. For example, the color when no rust occurs is yellow, and the color when rust occurs is red, and the color change is yellow → red; in addition, when rust occurs on the cross beam, column, and connecting bolts of the billboard bracket, thickness changes usually also occur, and environmental humidity is one of the factors causing rust. In this embodiment, by obtaining the color change, thickness change, and environmental humidity of the cross beam, column, and connecting bolts of the billboard bracket at the current moment, the rust degree on the surfaces of the cross beam, column, and connecting bolts and the material corrosion condition can be accurately reflected. Combining the environmental humidity data provides data support for subsequent effective evaluation of the rust process and prediction of the structural performance degradation trend, realizes dynamic monitoring and early warning of the structural safety of the billboard bracket, improves the maintenance efficiency, reduces potential safety hazards, and extends the service life.

[0036] Specifically, S101: Obtain the rust state and rust weight of the target component at the current moment, including steps S1011 - S1014:

[0037] S1011, collect the image, environmental humidity, and signal-to-noise ratio of the target component at the current moment.

[0038] In this embodiment, through the camera, the original images of the cross beam, column, and connecting bolts are collected, and the images of the cross beam, column, and connecting bolts are collected once per hour, or every half hour, or at other frequencies. Among them, the image when no rust occurs is the original image; through the temperature and humidity sensor, the environmental humidity of the cross beam, column, and connecting bolts is monitored in real time; in addition, the signal-to-noise ratio = effective signal amplitude / noise amplitude. The higher the SNR (Signal-to-Noise Ratio), the more reliable the ultrasonic detection result. The range of SNR is 20 - 50 dB. In this embodiment, the signal-to-noise ratio when the ultrasonic probe emits / receives sound waves is also monitored to be used for subsequent evaluation of the ultrasonic confidence of the cross beam, column, and connecting bolts at the current moment, and thus helps to utilize the signal-to-noise ratio subsequently.

[0039] S1012, perform color detection on the image at the current moment based on a color space to obtain the color change of the target component at the current moment, where the color space includes at least one of the following: RGB and Lab color spaces.

[0040] In this embodiment, based on the RGB (Red green blue) color space, detect the red, green, and blue color values of the crossbeam, column, and connecting bolt in the image at the current moment and the original image, and according to the formula ColorScore = (R t - R0) + (G t - G0) + (B t - B0), calculate the color change ColorScore of the crossbeam, column, and connecting bolt at the current moment t, where R t , G t , B t respectively represent the red, green, and blue color values of the image at the current moment t, and R0, G0, and B0 respectively represent the red, green, and blue color values of the original image. The red, green, and blue color values are in the range of 0 to 255. The larger the ColorScore, the more serious the corrosion. For example, the original image of the crossbeam, column, and connecting bolt is yellow (R0 = 200, G0 = 180, B0 = 50), and after corrosion, the image of the crossbeam, column, and connecting bolt at the current moment t becomes red (R t = 220, G t = 50, B t = 30), then ColorScore = (220 - 200) + (180 - 50) + (50 - 30) = 170.

[0041] Since the color change of the target component calculated based on the RGB color space is affected by light, this embodiment can also use the Lab (composed of three elements, one element is the luminance L, and a and b are two color channels) color space. The Lab color space separates the luminance (L, Luminance) from the color (a * , b * ), reducing light interference. The chromaticity differences Δa * , Δb * directly reflect the color shift direction. For example, a positive value of Δa is red-shifted, and a negative value is green-shifted. Therefore, based on the Lab color space, detect the a * and b * of the crossbeam, column, and connecting bolt in the image at the current moment and the original image, and according to the formula calculate the color change ColorScore of the crossbeam, column, and connecting bolt at the current moment t. For example, the original image of the crossbeam, column, and connecting bolt is: a * = 10, b *= 20, the images of the crossbeam, column, and connecting bolts after corrosion at the current moment t are: a * t = 25, b * t == 5, Δa * = 15, Δb * = -15,

[0042] It should be noted that when different metals corrode, the change directions of the color channels (red, green, blue) may be opposite. For example: after the iron target component corrodes, the red (R t ) may increase (rust is reddish), the green (G t ) and blue (B t ) may decrease (rust color is dull). The red, green, and blue color values of the original image are: R0 = 200, G0 = 180, B0 = 50 (bright yellow). The red, green, and blue color values of the corroded image are: R t = 220, G t = 50, B t = 30 (dark red), ΔR = (R t - R0) = +20, ΔG = (G t - G0) = -130, ΔB = (B t - B0) = -20. Therefore, based on the RGB color space, the color change of the target component calculated at the current moment may also be negative. And the degree of corrosion should reflect the "total change amount" of the color. Whether the color turns red or darkens, in this embodiment, it is considered to use the method of absolute value summation to handle negative values, that is, according to the formula ColorScore = |ΔR| + |ΔG| + |ΔB|, calculate the color change of the target component at the current moment. For example: ΔR = +20 → |ΔR| = 20, ΔG = -130 → |ΔG| = 130, ΔB = -20 → |ΔB| = 20, ColorScore = 20 + 130 + 20 = 170.

[0043] S1013, based on ultrasonic detection technology, perform thickness detection on the target component to obtain the thickness change of the target component at the current moment.

[0044] In this embodiment, when the metal rusts, a loose oxide layer (such as rust) will form on the surface. The volume of the oxide is usually larger than the original metal, so the metal seems to "become thicker" to the naked eye. However, the "thickening" is just a layer of rust covering the surface, and the actual thickness of the metal decreases. Therefore, the essence of rusting is that the metal is gradually "consumed" after being oxidized, just like wood being eaten by insects, with its internal structure damaged and the effective load-bearing part becoming thinner. The principle of ultrasonic detection technology is similar to "judging whether a watermelon is ripe by knocking on it and listening to the sound". It mainly emits high-frequency sound waves (such as 1 MHz) through an ultrasonic probe. The sound waves propagate inside the metal. Since the structure of healthy metal is dense, the sound wave propagation speed is fast (for example, the sound speed of steel is about 5900 m / s), while the rust layer is porous and loose, and the sound wave propagation speed is slow (for example, the sound speed of rust is about 2000 m / s), and most of the sound waves will be reflected or absorbed and cannot penetrate. When the sound wave encounters the interface between the metal and the rust layer, it will be reflected back (similar to an echo). The ultrasonic device defaults to measuring the thickness of healthy metal, and the rust layer will be regarded as "interference signal" and filtered out. Therefore, through the time difference between the sound wave emission and reception, combined with the sound speed formula, the measured thickness reduction value = the thickness consumed by the metal due to rusting, and the remaining thickness of the healthy metal layer can be calculated.

[0045] Based on the ultrasonic detection technology, the thickness of the cross beam, column, and connecting bolt is detected to obtain the thickness changes of the cross beam, column, and connecting bolt at the current moment, which specifically includes: obtaining the initial thickness Ultra0 of the cross beam, column, and connecting bolt. The initial thickness refers to the thickness when no rusting phenomenon occurs, such as 10 mm; based on the ultrasonic detection technology, detecting the initial frequency f0 of the cross beam, column, and connecting bolt, and detecting the frequency f of the cross beam, column, and connecting bolt at a preset frequency t , where the preset frequency includes but is not limited to: every hour or every half hour or other frequencies. The initial frequency refers to the frequency when no rusting phenomenon occurs. For example: f0 = 1 MHz, f t = 1.2 MHz, and according to the formula calculate the thickness change UltraScore of the cross beam, column, and connecting bolt at the current moment. Among them, Δf represents the frequency shift of the cross beam, column, and connecting bolt, Δf = f t - f0. Therefore, the thickness change UltraScore of the cross beam, column, and connecting bolt at the current moment can be calculated as UltraScore = 10 mm × 0.2 / 1.0 = 2 mm.

[0046] It should be noted that the camera, the temperature and humidity sensor, and the ultrasonic probe are all powered by solar cells, and transmit the image and the original image at the current moment, the environmental humidity at the current moment, the frequency at the current moment, and the initial frequency through 4G (The 4th generation mobile communication technology) / LoRa (Long Range, a low-power wide-area network communication technology) modules respectively.

[0047] S1014. Evaluate the corrosion weight of the target component at the current moment based on the image and the signal-to-noise ratio.

[0048] Specifically, the corrosion weight includes a color weight and a thickness weight.

[0049] In this embodiment, the corrosion weight refers to a weighted index used to quantitatively and comprehensively evaluate the influence degree of the corrosion state of the target component. The color weight reflects the influence degree of the color change on the corrosion severity. The more obvious the color, the more serious the corrosion, and the higher the weight may be. The thickness weight reflects the influence of the thickness reduction on the structural strength and load-bearing capacity. The greater the thickness loss, the greater the weight.

[0050] Specifically, S1014: Evaluate the corrosion weight of the target component at the current moment based on the image and the signal-to-noise ratio, including: evaluating the color confidence and ultrasonic confidence of the target component at the current moment, where the color confidence includes the blur degree and the illumination uniformity of the image, and the ultrasonic confidence refers to the signal-to-noise ratio corresponding to the thickness detection of the target component at the current moment; calculate the color weight and the thickness weight of the target component at the current moment respectively based on the color confidence and the ultrasonic confidence of the target component at the current moment.

[0051] In this embodiment, the blurriness is used to measure whether the image is clear and affects the accuracy of color feature extraction; the lighting uniformity is used to measure whether the lighting of the image is uniform to avoid inaccurate color judgment caused by shadows or overexposure. Calculate at least one of the variance of the Laplacian operator, gradient intensity, and number of edge detections of the image as the blurriness of the image; calculate at least one of the standard deviation of pixel brightness, coefficient of variation of pixel brightness, color temperature of the image, or brightness histogram as the lighting uniformity of the image. According to the mapping relationship between blurriness and score, and the mapping relationship between lighting uniformity and score, obtain the scores corresponding to blurriness and lighting uniformity respectively. For example, when the variance of the Laplacian operator of the image ≤ 5%, the color confidence is high. When the variance of the Laplacian operator of the image is 3%, the score corresponding to the variance of the Laplacian operator of the image is 0.9. When the standard deviation of pixel brightness ≤ 10%, the color confidence is high. When the standard deviation of pixel brightness is 8%, the score corresponding to the standard deviation of pixel brightness is 0.8. Furthermore, according to the formula ColorConfidence = 0.5 × score1 + 0.5 × score2, calculate the color confidence ColorConfidence of the crossbeam, column, and connecting bolt at the current moment, where score1 represents the score corresponding to at least one of the variance of the Laplacian operator, gradient intensity, and number of edge detections of the image, and score2 represents the score corresponding to at least one of the standard deviation of pixel brightness, coefficient of variation of pixel brightness, color temperature of the image, or brightness histogram.

[0052] After monitoring the signal-to-noise ratio SNR when the ultrasonic probe emits / receives sound waves in this embodiment, according to the formula Calculate the ultrasonic confidence UltraConfidence of the crossbeam, column, and connecting bolt at the current moment.

[0053] Under normal circumstances, humidity will cause reflective interference. The higher the humidity, the lower the color weight and the higher the thickness weight. Therefore, this embodiment can calculate the initial color weight w of the crossbeam, column, and connecting bolt according to the formula where H represents the ambient humidity of the crossbeam, column, and connecting bolt at the current moment, and the range of H is 0–100%. For example, when the ambient humidity is 85%, H = 85, k represents the adjustment parameter for controlling the influence intensity of ambient humidity on the color weight, which is obtained through historical data training. For example, k = 0.1, e represents the natural constant, e is approximately 2.718, which is used for exponential calculation, and the range of w color is 0–1. The initial thickness weight w of the crossbeam, column, and connecting bolt can be calculated according to the formula w color = 1 - w ultra color ultra ultra .

[0054] According to the formula w​final-color = w color × ColorConfidence and formula w final-ultra = w ultra × UltraConfidence, respectively calculate the color weight w of the target component at the current moment final-color and the thickness weight w final-ultra . In this embodiment, through the weight calculation technology based on confidence, the multi-source sensor data quality information is effectively fused, and the adaptive changes of the color weight and the thickness weight are realized, the effectiveness of the color weight and the thickness weight is improved, the dynamic evaluation of the corrosion risk of the target component is realized, so as to improve the accuracy and reliability of the evaluation of the corrosion state of the target component, and provide more powerful technical support for the structural safety monitoring and maintenance.

[0055] It should be noted that if the signal-to-noise ratio when the ultrasonic probe emits / receives sound waves ≥ 20 dB, the ultrasonic confidence is high. After this embodiment monitors the signal-to-noise ratio when the ultrasonic probe emits / receives sound waves, it can also, in the same way as according to the mapping relationship between the blurriness and the score, and the mapping relationship between the illumination uniformity and the score, obtain the scores corresponding to the blurriness and the illumination uniformity respectively, and obtain the score corresponding to the signal-to-noise ratio according to the mapping relationship between the signal-to-noise ratio and the score, so as to be used as the ultrasonic confidence. For example: when SNR = 25 dB, the corresponding score is 0.8.

[0056] S102. Based on the corrosion state and the corrosion weight, evaluate the corrosion risk of the target component at the current moment.

[0057] Specifically, S102: Based on the corrosion state and the corrosion weight, evaluate the corrosion risk of the target component at the current moment, including: According to formula (1), evaluate the corrosion risk of the target component at the current moment:

[0058] CRR = w final-color · ColorScore + w final-ultra · UltraScore + w humidity · H (1),

[0059] where CRR represents the corrosion risk of the target component at the current moment, w final-color represents the color weight of the target component at the current moment, ColorScore represents the color change of the target component at the current moment, w final-ultra represents the thickness weight of the target component at the current moment, UltraScore represents the thickness change of the target component at the current moment, w humidity represents the preset humidity weight, and H represents the environmental humidity of the target component at the current moment.

[0060] In this embodiment, the rust weight further includes a humidity weight. Similarly to the changing color weight and thickness weight, the preset humidity weight w humidity is not a constant value. Since the higher the environmental humidity, the faster the rusting speed, in this embodiment, a mapping relationship between the environmental humidity and the humidity weight is constructed, and according to the mapping relationship between the environmental humidity and the humidity weight, the preset humidity weights of the cross beam, column, and connecting bolt at the current moment are matched to achieve the adaptive change of the humidity weight.

[0061] If the color changes: (Δa * = +20, Δb * = -15) → ColorScore = 35, the thickness changes: UltraScore = 2.0, the environmental humidity: H = 90%, the initial color weight: The initial thickness weight: w ultra = 1 - 0.12 = 0.88, the color confidence: (the score corresponding to the fuzziness of 2% is 0.95, the score corresponding to the light variance of 5% is 0.9) → ColorConfidence = 0.925, the ultrasonic confidence: SNR = 28dB → UltraConfidenc = 0.8, then w final-color = 0.12 × 0.925 ≈ 0.11, w final-ultra = 0.88 × 0.8 ≈ 0.70, CRR = 0.11 × 35 + 0.70 × 2.0 + 0.1 × 90 = 3.85 + 1.4 + 9 = 14.25.

[0062] S103. Based on the rust risk of the target component at the current moment, determine the rusted component and its corresponding rust diffusion component, where the rusted component refers to the target component that has rusted at the current moment, and the rust diffusion component refers to the target component to which the rust phenomenon spreads within a preset time period in the future.

[0063] Specifically, S103: Based on the rust risk of the target component at the current moment, determine the rusted component and its corresponding rust diffusion component, including steps S1031 - S1032:

[0064] S1031. Based on the rust risk of the target component at the current moment, generate a spatio - temporal graph convolutional network of the target object.

[0065] Specifically, the spatio - temporal graph convolutional network includes nodes, node attributes, edges, and weights.

[0066] Specifically, S1031: Generate a spatio-temporal graph convolutional network for the target object based on the corrosion risk of the target component at the current moment, including: determining the nodes and edges of the target object by taking the target component and the physical connection relationship between the target components respectively; determining the weights of the target object by taking the distance between the target components and the material type corresponding to the physical connection relationship between the target components; evaluating the corrosion degree and stress value of the target component at the current moment based on the corrosion risk of the target component at the current moment; and determining the node attributes of the target object by taking the environmental humidity, corrosion degree and stress value of the target component at the current moment.

[0067] In this embodiment, the crossbeam, the column, and the connecting bolt are determined as nodes. For example: nodes = {101: "Main crossbeam (Q235 carbon steel)", 202: "Support column (galvanized square pipe)", 303: "High-strength connecting bolt (304 stainless steel)"}. The physical connection relationships between the crossbeam, the column, and the connecting bolt include, but are not limited to: the crossbeam is connected to the connecting bolt, and the column is connected to the connecting bolt. The connection methods include, but are not limited to: bolt connection, welding connection, hinge connection, lap / socket structure, and connection with a support or bracket. Therefore, the edges can be determined to include: the edges between the node corresponding to the crossbeam and the node corresponding to the connecting bolt, and the edges between the node corresponding to the column and the node corresponding to the connecting bolt. For example: edges = [(101, 303, {"type": "bolt fastening", "interface": "steel - stainless steel"}), (303, 202, {"type": "flange connection", "interface": "stainless steel - galvanized steel"})]. The weights of the target object include the distance weight and the material weight of the target object. Based on the BIM (Building Information Modeling) model coordinates, calculate the distances and distance weights between the crossbeam, the column, and the connecting bolt. For example:

[0068]

[0069] Obtain the material types corresponding to the connection methods of the connection between the crossbeam and the connecting bolt and the connection between the column and the connecting bolt, and calculate the electrochemical corrosion potential difference and material weight of the material types corresponding to the connection methods, as shown in Table 1.

[0070] Table 1 Electrochemical corrosion potential difference and material weight of material types

[0071] Connection type Material combination Potential difference (mV) Weight coefficient Bolt fastening Carbon steel - Stainless steel 320 0.55 Flange connection Stainless steel - Galvanized steel 180 0.75

[0072] In this embodiment, according to the corrosion degree quantification model and the corrosion risks of the crossbeam, the column, and the connecting bolt at the current moment, evaluate the corrosion degrees of the crossbeam, the column, and the connecting bolt at the current moment. For example:

[0073]

[0074]

[0075] In this embodiment, the stress values of the cross beam, the column, and the connecting bolts at the current moment are calculated through a stress-corrosion coupling analysis algorithm. For example:

[0076] # Stress concentration calculation considering the influence of corrosion pits

[0077] Kt = 1 + 2 * sqrt(A_corrosion / (π * d)) # Stress concentration factor

[0078] σ_effective = σ_nominal * Kt + 0.15 * E * (t_rust / t_original)

[0079] # Calculation result of cross beam 101: σ_nominal = 85 MPa → σ_effective = 121 MPa (allowable stress 135 MPa).

[0080] The environmental humidity, corrosion degree, and stress values of the cross beam, the column, and the connecting bolts at the current moment are determined as node attributes. For example: {"101": {"environmental humidity": 92% RH, "surface temperature": 41 °C, "corrosion degree": 0.68, "real-time stress": 121 MPa, "vibration frequency": 12.3 Hz}, "303": {"environmental humidity": 88% RH, "preload loss": 18%, "thread corrosion": 0.31, "shear stress": 67 MPa}}. In this embodiment, the corrosion degree and stress values comprehensively reflect the damage condition and stress state of the components, helping to accurately judge the structural safety. The environmental humidity, as an external influencing factor, helps to understand the driving conditions for corrosion propagation, enhances the environmental relevance of the assessment, improves the accuracy of the spatio-temporal graph convolutional network of the target object, and provides more accurate data support for subsequent determination of the corroded components and their corresponding corrosion-propagating components.

[0081] It should be noted that in this embodiment, it is also possible to directly determine whether the corrosion risk of the cross beam, column, and connecting bolts at the current moment is greater than a preset threshold. Taking the preset threshold as 80 as an example, if the corrosion risk of the cross beam, column, and connecting bolts at the current moment ≥ 80, then the corrosion degree of the cross beam, column, and connecting bolts at the current moment is "red warning", and immediate repair is required. For example: color change = 170, thickness change = 2mm, humidity = 0.8, corrosion risk = 0.6×170 + 0.3×2 + 0.1×0.8 = 102 + 0.6 + 0.08 = 102.68, then the corrosion degree is "red warning", and immediate repair is required. In addition, after determining whether the corrosion risk of the cross beam, column, and connecting bolts at the current moment is greater than the preset threshold in this embodiment, it further includes: if the corrosion risk of the cross beam, column, and connecting bolts at the current moment is within [60, 80), then the corrosion degree of the cross beam, column, and connecting bolts at the current moment is yellow warning (observation); if the corrosion risk of the cross beam, column, and connecting bolts at the current moment is within [0, 60), then the corrosion degree of the cross beam, column, and connecting bolts at the current moment is green safety.

[0082] S1032. Based on the spatio-temporal graph convolutional network of the target object, determine the corroded components and their corresponding corrosion diffusion components.

[0083] Specifically, S1032: Based on the spatio-temporal graph convolutional network of the target object, determine the corroded components and their corresponding corrosion diffusion components, including: determining the nodes whose node attributes meet the corrosion conditions as corroded components; predicting the spread speed between the corroded components and other components in the target object based on the material types corresponding to the physical connection relationships between the corroded components and other components in the target object; determining the corrosion diffusion components corresponding to the corroded components based on the spread speed and distance between the corroded components and other components of the target object.

[0084] In this embodiment, the corrosion conditions include a corrosion degree threshold and a stress threshold, and it is judged whether the corrosion degree and stress value of the cross beam, column, and connecting bolts at the current moment are greater than the corrosion degree threshold and the stress threshold; if the corrosion degree of the cross beam, column, and connecting bolts at the current moment is greater than the corrosion degree threshold or the stress value of the cross beam, column, and connecting bolts at the current moment is greater than the stress threshold, then determine the nodes corresponding to the corrosion degree greater than the corrosion degree threshold or the stress value greater than the stress threshold as corroded components. For example:

[0085] # Corrosion determination criteria (need to be calibrated according to material properties)

[0086] def is_corrosion_node(node): return (node['corrosion degree'] > 0.7) or (node['stress ratio'] > 0.85) # Stress ratio = real-time stress / yield strength

[0087] #Current detection data: Node 101 (main crossbeam): Rusting degree 0.82, stress ratio 0.79 → Determined as a rusted component; Node 303 (bolt): Rusting degree 0.65, stress ratio 0.91 → Determined as a rusted component.

[0088] This embodiment can predict the spread rate between the rusted component and other components in the target object based on the material combination corrosion coefficient table shown in Table 2 and the material types corresponding to the physical connection relationships between the rusted component and other components in the target object.

[0089] Table 2 Material combination corrosion coefficient table

[0090]

[0091]

[0092] This embodiment can also predict the spread rate between the rusted component and other components in the target object based on the spread rate formula and the material types corresponding to the physical connection relationships between the rusted component and other components in the target object. For example:

[0093]

[0094] Based on the spread rate and distance between the rusted component and other components of the target object, the time period required for the rust phenomenon to spread to other components of the target object can be calculated. Determine whether the required time period is less than or equal to the preset time period. If the required time period is less than or equal to the preset time period, then determine that the other components of the target object corresponding to the required time period are the rust spread components corresponding to the rusted component. This embodiment comprehensively considers the connection relationship and materials to establish a more complete rust spread network model, improves the adaptability to the rust development trend under complex working conditions, enhances the authenticity and accuracy of prediction, quickly locates the adjacent components that may be affected by rust, warns the risk spread area in advance, helps to focus on monitoring and maintaining the key nodes with high risk of spread, and prevents the overall structure from failing due to rust spread.

[0095] It should be noted that after determining the rusted component and its corresponding rust spread components in this embodiment, the loss function, that is Optimize and adjust the rust situation, the material combination corrosion coefficient table, and the spread rate formula. Among them, T represents the time step, such as the data of the past 6 months, and λ represents the regularization parameter, such as: 0.01, to prevent the model from overfitting. The structure sparse constraint forces the model to only focus on the physically connected nodes and ignores the irrelevant paths.

[0096] A rust detection method provided in this embodiment dynamically evaluates the rust risk of a target component through dynamically changing rust weights, improving the accuracy of rust risk assessment. Furthermore, based on the dynamically changing rust risk of the target component, it quickly and effectively determines the rusty components and their corresponding rust diffusion components, enhancing the accuracy and efficiency of rust detection and achieving reliable and rapid rust detection.

[0097] Embodiment 2:

[0098] As Figure 2 shown, this embodiment provides a rust detection device, including an acquisition module 21, an evaluation module 22, and a determination module 23. The acquisition module 21 is used to acquire the rust state and rust weight of the target component at the current moment. Here, the target component refers to each component of the target object. The evaluation module 22 is connected to the acquisition module 21 and is used to evaluate the rust risk of the target component at the current moment based on the rust state and the rust weight. The determination module 23 is connected to the evaluation module 22 and is used to determine the rusty components and their corresponding rust diffusion components based on the rust risk of the target component at the current moment. Here, the rusty component refers to the target component that has a rust phenomenon at the current moment, and the rust diffusion component refers to the target component to which the rust phenomenon spreads within a preset time period in the future.

[0099] Specifically, the acquisition module 21 includes: a collection unit 211, a color detection unit 212, a thickness detection unit 213, and a first evaluation unit 214. The collection unit 211 is used to collect the image, environmental humidity, and signal-to-noise ratio of the target component at the current moment. The color detection unit 212 is used to perform color detection on the image at the current moment based on the color space to obtain the color change of the target component at the current moment. Here, the color space includes at least one of the following: RGB and Lab color spaces. The thickness detection unit 213 is used to perform thickness detection on the target component based on ultrasonic detection technology to obtain the thickness change of the target component at the current moment. The first evaluation unit 214 is used to evaluate the rust weight of the target component at the current moment based on the image and the signal-to-noise ratio.

[0100] Specifically, the first evaluation unit 214 includes: a first evaluation subunit and a calculation subunit. The first evaluation subunit is used to evaluate the color confidence and ultrasonic confidence of the target component at the current moment. Here, the color confidence includes the blurriness and illumination uniformity of the image, and the ultrasonic confidence refers to the signal-to-noise ratio corresponding to the thickness detection of the target component at the current moment. The calculation subunit is used to calculate the color weight and thickness weight of the target component at the current moment based on the color confidence and ultrasonic confidence of the target component at the current moment.

[0101] Specifically, the evaluation module 22 includes: a second evaluation unit 221, configured to evaluate the corrosion risk of the target component at the current moment according to formula (1):

[0102] CRR = w final-color ·ColorScore + w final-ultra ·UltraScore + w humidity ·H (1),

[0103] where CRR represents the corrosion risk of the target component at the current moment, w final-color represents the color weight of the target component at the current moment, ColorScore represents the color change of the target component at the current moment, w final-ultra represents the thickness weight of the target component at the current moment, UltraScore represents the thickness change of the target component at the current moment, w humidity represents the preset humidity weight, and H represents the environmental humidity of the target component at the current moment.

[0104] Specifically, the determination module 23 includes: a generation unit 231 and a determination unit 232. The generation unit 231 is configured to generate a spatio-temporal graph convolutional network of the target object at the current moment based on the corrosion risk of the target component at the current moment. The determination unit 232 is configured to determine the corroded component and its corresponding corrosion diffusion component based on the spatio-temporal graph convolutional network of the target object at the current moment.

[0105] Specifically, the generation unit 231 includes: a first determination subunit, a second determination subunit, a second evaluation subunit, and a third determination subunit. The first determination subunit is configured to determine the nodes and edges of the target object at the current moment by using the target component and the physical connection relationship between the target components respectively. The second determination subunit is configured to determine the weights of the target object at the current moment by using the distance between the target components and the material type corresponding to the physical connection relationship between the target components. The second evaluation subunit is configured to evaluate the corrosion degree and stress value of the target component at the current moment based on the corrosion risk of the target component at the current moment. The third determination subunit is configured to determine the node attributes of the target object at the current moment by using the environmental humidity, corrosion degree, and stress value of the target component at the current moment.

[0106] Specifically, the determination unit 232 includes: a fourth determination subunit, a prediction subunit, and a fifth determination subunit. The fourth determination subunit is configured to determine that a node whose node attribute meets the corrosion condition is a corroded component. The prediction subunit is configured to predict the spread rate between the corroded component and other components in the target object based on the material type corresponding to the physical connection relationship between the corroded component and other components in the target object. The fifth determination subunit is configured to determine the corrosion diffusion component corresponding to the corroded component based on the spread rate and distance between the corroded component and other components in the target object.

[0107] It can be understood that the rust detection device provided above executes the rust detection method corresponding to Embodiment 1 provided above. Therefore, the beneficial effects it can achieve can refer to the beneficial effects of the solution corresponding to the rust detection method in Embodiment 1 above, which will not be elaborated here.

[0108] Embodiment 3:

[0109] This embodiment provides an electronic device, including a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to implement the rust detection method in Embodiment 1 above.

[0110] Embodiment 4:

[0111] This embodiment provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the rust detection method in Embodiment 1 above is implemented.

[0112] It can be understood that the above embodiments are merely exemplary embodiments adopted to illustrate the principles of the present invention. However, the present invention is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also regarded as the protection scope of the present invention.

Claims

1. A rust detection method, characterized in that, Including: Obtain the rust state and rust weight of the target component at the current moment, where the target component refers to each component of the target object; Evaluate the rust risk of the target component at the current moment based on the rust state and the rust weight; Determine the rusted component and its corresponding rust diffusion component based on the rust risk of the target component at the current moment, where the rusted component refers to the target component that has rusted at the current moment, and the rust diffusion component refers to the target component that the rust phenomenon spreads to within a preset time period in the future.

2. The rust detection method according to claim 1, wherein The rust state includes color change, thickness change, and environmental humidity. The obtaining of the rust state and rust weight of the target component at the current moment specifically includes: Collect the image, environmental humidity, and signal-to-noise ratio of the target component at the current moment; Based on the color space, perform color detection on the image at the current moment to obtain the color change of the target component at the current moment, where the color space includes at least one of the following: RGB and Lab color spaces; Based on ultrasonic detection technology, perform thickness detection on the target component to obtain the thickness change of the target component at the current moment; Evaluate the rust weight of the target component at the current moment based on the image and the signal-to-noise ratio.

3. The rust detection method according to claim 2, characterized in that, The rust weight includes color weight and thickness weight. Evaluating the rust weight of the target component at the current moment based on the image and the signal-to-noise ratio specifically includes: Evaluate the color confidence and ultrasonic confidence of the target component at the current moment, where the color confidence includes the blur degree and illumination uniformity of the image, and the ultrasonic confidence refers to the signal-to-noise ratio corresponding to the thickness detection of the target component at the current moment; Based on the color confidence and ultrasonic confidence of the target component at the current moment, calculate the color weight and thickness weight of the target component at the current moment respectively.

4. The rust detection method according to claim 3, wherein Evaluating the rust risk of the target component at the current moment based on the rust state and the rust weight specifically includes: According to formula (1), evaluate the rust risk of the target component at the current moment: CRR = w final-color · ColorScore + w final-ultra · UltraScore + w humidity · H(1), Among them, CRR represents the corrosion risk of the target component at the current moment, w final-color represents the color weight of the target component at the current moment, ColorScore represents the color change of the target component at the current moment, w final-ultra represents the thickness weight of the target component at the current moment, UltraScore represents the thickness change of the target component at the current moment, w humidity represents the preset humidity weight, and H represents the environmental humidity of the target component at the current moment.

5. The rust detection method according to claim 1, wherein The determining of the rusted component and its corresponding rust diffusion component based on the rust risk of the target component at the current moment specifically includes: Generate a spatio-temporal graph convolutional network of the target object based on the rust risk of the target component at the current moment; Determine the rusted component and its corresponding rust diffusion component based on the spatio-temporal graph convolutional network of the target object.

6. The rust detection method according to claim 5, wherein, The spatio-temporal graph convolutional network includes nodes, node attributes, edges, and weights. The generating of the spatio-temporal graph convolutional network of the target object based on the rust risk of the target component at the current moment specifically includes: Determine the nodes and edges of the target object with the target components and the physical connection relationships between the target components respectively; Determine the weights of the target object with the distances between the target components and the material types corresponding to the physical connection relationships between the target components; Evaluate the rust degree and stress value of the target component at the current moment based on the rust risk of the target component at the current moment; Determine the node attributes of the target object with the environmental humidity, rust degree, and stress value of the target component at the current moment.

7. The rust detection method according to claim 6, wherein The determining of the rusted component and its corresponding rust diffusion component based on the spatio-temporal graph convolutional network of the target object specifically includes: Determine that the node with the determined node attribute satisfying the rust condition is a rusted component; Predict the spread rate between the rusted component and other components in the target object based on the material type corresponding to the physical connection relationship between the rusted component and other components in the target object; Determine the rust diffusion component corresponding to the rusted component based on the spread rate and distance between the rusted component and other components of the target object.

8. A rust detection device, characterized in that, It includes an acquisition module, an evaluation module, and a determination module. The acquisition module is used to acquire the rust state and rust weight of the target component at the current moment, where the target component refers to each component of the target object. The evaluation module is connected to the acquisition module and is used to evaluate the rust risk of the target component at the current moment based on the rust state and the rust weight. The determination module is connected to the evaluation module and is used to determine the rusted component and its corresponding rust diffusion component based on the rust risk of the target component at the current moment, where the rusted component refers to the target component with rust phenomenon occurring at the current moment, and the rust diffusion component refers to the target component to which the rust phenomenon spreads within a preset time period in the future.

9. An electronic device, characterized in that, It includes a memory and a processor. A computer program is stored in the memory, and the processor is configured to run the computer program to implement a rust detection method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, a rust detection method according to any one of claims 1 to 7 is implemented.

Images