Quantitative detection method and system for steel wire rope corrosion based on magnetoresistance sensing-visual combination

Abstract

The present application relates to steel wire rope corrosion detection, specifically to a steel wire rope corrosion quantitative detection method and system based on magnetoresistance sensing-visual combination, a magnetoresistance sensor array collects magnetoresistance signals at each position on the steel wire rope, a magnetoresistance collection module sends the collected magnetoresistance signals to a signal conditioning module for signal conditioning; when the steel wire rope starts to run, a CCD camera synchronously collects images of the steel wire rope, a visual collection card sends the collected visual signals to an image processing module for image processing; a synchronous encoder is used to record the moving position of the steel wire rope in real time, and a computer records the magnetoresistance signals and visual signals at the same position according to the moving position of the steel wire rope; the computer builds a weighted decision fusion model, determines the corrosion depth grade according to the magnetic field abnormal value, determines the corrosion area grade according to the corrosion area ratio and the rust spot fractal dimension, and then obtains the final corrosion grade; the present application can effectively overcome the defect that it is difficult to accurately and quantitatively detect the corrosion of the steel wire rope.

Description

Technical Field

[0001] This invention relates to the detection of steel wire rope corrosion, specifically to a quantitative detection method and system for steel wire rope corrosion based on a combination of magnetoresistive sensing and vision. Background Technology

[0002] Steel wire rope is a commonly used traction and support material in modern machinery, widely applied in construction, mining, and transportation. Corrosion of steel wire ropes leads to a decrease in strength and can even cause strand breakage. Traditional methods for detecting steel wire rope corrosion, such as visual inspection, tapping, and caliper measurement, suffer from drawbacks including high subjectivity, low accuracy, and inability to quantitatively detect corrosion. Single technologies, such as electromagnetic detection based on leakage flux and eddy currents, are insensitive to pitting corrosion, while purely visual methods are easily affected by factors such as light and oil contamination, and struggle to assess the depth of corrosion.

[0003] Unlike using Hall effect sensors or induction coil sensors to detect changes in the magnetic field of steel wire rope corrosion, magnetoresistive sensors are high-precision magnetic field sensors capable of measuring minute changes in the magnetic field. This invention aims to solve the aforementioned problems by achieving integrated non-destructive quantitative detection of steel wire ropes through multi-source information fusion analysis, combining external morphological observation with internal depth sensing. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] In view of the above-mentioned shortcomings of the existing technology, the present invention provides a quantitative detection method and system for steel wire rope corrosion based on magnetoresistive sensing-vision combination, which can effectively overcome the shortcomings of the existing technology in accurately and quantitatively detecting steel wire rope corrosion.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision integration includes the following steps:

[0009] S1. Use a magnetizer to magnetize the steel wire rope until it is saturated;

[0010] S2. Start the wire rope traction device and synchronous encoder. The wire rope starts to move. The magnetoresistive sensor array collects the magnetoresistive signals at various positions on the wire rope. The magnetoresistive acquisition module sends the collected magnetoresistive signals to the signal conditioning module for signal conditioning.

[0011] S3. When the wire rope starts to move, the CCD camera synchronously acquires images of the wire rope, and the vision acquisition card sends the acquired visual signals to the image processing module for image processing.

[0012] S4. The synchronous encoder is used to record the movement position of the wire rope in real time. The computer records the magnetoresistive signal and visual signal at the same position based on the movement position of the wire rope.

[0013] S5. The computer combines magnetoresistive signals and visual signals to construct a weighted decision fusion model. It determines the corrosion depth level based on the magnetic field anomaly value, and determines the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spot, thus obtaining the final corrosion level.

[0014] Preferably, in step S2, the wire rope traction device and synchronous encoder are activated, the wire rope begins to move, and the magnetoresistive sensor array collects magnetoresistive signals at various positions on the wire rope. The magnetoresistive acquisition module sends the collected magnetoresistive signals to the signal conditioning module for signal conditioning, including:

[0015] S21. Start the wire rope traction device and synchronous encoder. The wire rope starts to move. Each magnetoresistive sensor in the magnetoresistive sensor array collects the circumferential leakage magnetic signal at each position on the wire rope.

[0016] S22. The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic signals acquired by each magnetoresistive sensor to obtain a one-dimensional leakage magnetic signal curve, and sends it to the signal conditioning module.

[0017] S23. The signal conditioning module conditions the leakage magnetic signal curve and sends the conditioned leakage magnetic signal curve to the computer.

[0018] Preferably, in step S3, when the wire rope begins to move, the CCD camera simultaneously acquires images of the wire rope, and the vision acquisition card sends the acquired visual signals to the image processing module for image processing, including:

[0019] S31. When the wire rope starts to move, the CCD camera simultaneously acquires images of the wire rope.

[0020] S32. The vision acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module.

[0021] S33. The image processing module performs grayscale processing on the image and then enhances the contrast to obtain a clearer image, and then sends the processed image to the computer.

[0022] Preferably, in S5, the computer combines magnetoresistive signals and visual signals to construct a weighted decision fusion model. Based on the magnetic field anomaly value, the corrosion depth level is determined; based on the corrosion area ratio and the fractal dimension of the rust spots, the corrosion area level is determined; and finally, the final corrosion level is obtained, including:

[0023] S51. Calculate the magnetic field anomaly value using the leakage magnetic field signal curve, and determine the corrosion depth level and the corresponding corrosion depth reliability.

[0024] S52. Extract the two-dimensional morphological features of the rusted area from the image, calculate the rust area ratio and the fractal dimension of the rust spots, and determine the rust area level and the corresponding rust area confidence level.

[0025] S53. Based on the weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence, corrosion area level, and corrosion area confidence.

[0026] Preferably, in step S51, the magnetic field anomaly value is calculated using the leakage magnetic field signal curve, and the corrosion depth level and corresponding corrosion depth reliability are determined, including:

[0027] S511. Considering that under constant excitation, according to Ohm's law for magnetic circuits and the magnetic properties of materials, the magnetic reluctance R of the wire rope... m Changes occur, causing distortion of the magnetic field distribution in the space near its surface. For uniform corrosion, due to the poor magnetic permeability of iron oxides, the relative permeability of the corrosion region is approximately considered to be... The magnetic field anomaly value decreases. According to magnetic field theory, the surface normal magnetic induction intensity is related to the material state. The following formula is used to calculate the magnetic field anomaly value:

[0028] ;

[0029] in, Let B be the magnetic field anomaly value at position x along the axial direction of the wire rope. measured (x) represents the leakage magnetic signal curve at position x along the axial direction of the wire rope, B baseline This represents the baseline value of the magnetic field at a defect-free location on the wire rope. Let x be the corrosion depth at position x along the axial direction of the wire rope. The cross-sectional area loss is at position x along the axial direction of the wire rope. k1 and k2 are constants, determined through calibration experiments, reflecting the influence of the geometric steepness of the defect on the leakage magnetic field.

[0030] S512. By using the calibration curve and signal-to-noise ratio, the magnetic field anomaly value is determined. Preliminary estimate of corrosion depth level L m and the corresponding rust depth confidence level C x .

[0031] Preferably, step S52 involves extracting the two-dimensional morphological features of the rusted area from the image, calculating the rust area ratio and the fractal dimension of the rust spots, and determining the rust area level and the corresponding rust area confidence level, including:

[0032] S521. Segment the rusted areas in the image, separating them from the background, and set the mask function:

[0033] ;

[0034] Among them, Mask(x) p ,y p ) represents pixel coordinates (x) p ,y p The mask value at H(x) is 1 when it represents a rusted area and 0 when it represents the background. p ,y p ), S(x p ,y p ) are pixel coordinates (x p ,y p Hue and saturation at ) H min H max These are the minimum and maximum values ​​of the hue, S. th This is the saturation threshold.

[0035] S522. By removing small noise points, connecting fracture areas, and extracting features from rusted areas, the rust area ratio is calculated:

[0036] ;

[0037] Among them, R area Pixel cor Pixel total These represent the total number of pixels showing rust spots on the surface of the steel wire rope and the total number of pixels, respectively.

[0038] S523. To describe the roughness and severity of rust, the box counting method is used to approximate the fractal dimension of rust spots:

[0039] ;

[0040] in, The required side length to cover the outline of the rust spot is The number of boxes, D f The fractal dimension of the rust spots;

[0041] S524. Based on image clarity and segmentation effect, the ratio of rust area R... area and the fractal dimension D of rust spots f Preliminary estimate of corrosion area level L v and the corresponding rust area reliability C v .

[0042] Preferably, in S53, based on a weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence level, corrosion area level, and corrosion area confidence level, including:

[0043] The final corrosion grade is calculated using the following formula:

[0044] ;

[0045] Among them, L final For the final corrosion grade, the confidence level C for internal or uniform rust with strong magnetic reluctance signals and inconspicuous visual signals is determined. x Automatically increasing the confidence level C of the corrosion area for pitting corrosion with weak magnetoresistance signals and obvious visual signals. v Automatically increases in size.

[0046] A quantitative detection system for steel wire rope corrosion based on magnetoresistive sensing and vision integration includes:

[0047] The magnetoresistive sensor array has a ring structure with multiple magnetoresistive sensors evenly distributed. When the wire rope is magnetized and pulled by the wire rope traction device, it collects the circumferential leakage magnetic signals at various positions on the wire rope.

[0048] The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic field signals acquired by each magnetoresistive sensor, obtains a one-dimensional leakage magnetic field signal curve, and sends it to the signal conditioning module.

[0049] The signal conditioning module conditions the leakage flux signal curve and sends the conditioned leakage flux signal curve to the computer.

[0050] When the steel wire rope begins to move, a pair of CCD cameras simultaneously acquire images of the steel wire rope. Each CCD camera is equipped with a supplementary light to facilitate clear imaging.

[0051] The visual acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module;

[0052] The image processing module performs grayscale processing on the image and then contrast enhancement to obtain a clearer image, and then sends the processed image to the computer.

[0053] The computer records the magnetoresistive and visual signals at the same position of the wire rope in real time based on the synchronous encoder. It then constructs a weighted decision fusion model by combining the magnetoresistive and visual signals. Based on the magnetic field anomaly value, it determines the corrosion depth level, and based on the corrosion area ratio and the fractal dimension of the rust spot, it determines the corrosion area level, thus obtaining the final corrosion level.

[0054] Preferably, the computer constructs a weighted decision fusion model by combining magnetoresistive signals and visual signals, determines the corrosion depth level based on magnetic field anomalies, determines the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spots, and then obtains the final corrosion level, including:

[0055] The magnetic field anomaly value was calculated using the leakage magnetic field signal curve, and the corrosion depth level and the corresponding corrosion depth confidence level were determined.

[0056] Two-dimensional morphological features of the rusted area are extracted from the image, the rust area ratio and rust spot fractal dimension are calculated, and the rust area level and corresponding rust area confidence level are determined.

[0057] Based on a weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence level, corrosion area level, and corrosion area confidence level.

[0058] (III) Beneficial Effects

[0059] To address the shortcomings of existing wire rope corrosion detection methods, this invention proposes a quantitative detection method and system for wire rope corrosion based on magnetoresistive sensing and vision integration. This method simultaneously acquires local magnetic field signals and surface image information of the wire rope, utilizes a magnetoresistive sensor to deeply perceive changes in magnetic properties and cross-sectional area loss caused by corrosion, and employs machine vision to accurately identify the surface morphology features of the corrosion. Finally, a weighted decision fusion model is used to output the final corrosion level, achieving precise quantitative detection of wire rope corrosion. This provides an innovative solution for improving the detection accuracy and comprehensive defect evaluation capabilities of wire rope corrosion.

[0060] Currently, nondestructive testing methods for detecting steel wire rope corrosion mainly rely on electromagnetic or visual methods for analysis, with few solutions integrating multiple analysis methods. The quantitative detection method and system for steel wire rope corrosion based on magnetoresistive sensing and vision proposed in this invention has unique engineering application value. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0062] Figure 1 This is a schematic diagram of the process of the present invention;

[0063] Figure 2 This is a schematic diagram of the system of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0065] The following describes the specific process of the quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined, provided by this invention, with specific examples (e.g.) Figure 1 (as shown) and technical effects.

[0066] S1. Use a magnetizer (through magnetizer) to magnetize the wire rope to saturation.

[0067] S2. Start the wire rope traction device and synchronous encoder. The wire rope begins to move. The magnetoresistive sensor array collects magnetoresistive signals at various positions on the wire rope. The magnetoresistive acquisition module sends the collected magnetoresistive signals to the signal conditioning module for signal conditioning, including:

[0068] S21. Start the wire rope traction device and synchronous encoder. The wire rope starts to move (at a speed of 2m / s). Each magnetoresistive sensor in the magnetoresistive sensor array collects the circumferential leakage magnetic signal at each position on the wire rope (collected at a sampling rate of 1kHz).

[0069] S22. The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic signals acquired by each magnetoresistive sensor to obtain a one-dimensional leakage magnetic signal curve, and sends it to the signal conditioning module.

[0070] S23. The signal conditioning module conditions the leakage magnetic signal curve and sends the conditioned leakage magnetic signal curve to the computer.

[0071] S3. When the wire rope begins to move, the CCD camera (CCD high-resolution anti-flicker camera) simultaneously acquires images of the wire rope. The vision acquisition card sends the acquired visual signals to the image processing module for image processing, including:

[0072] S31. When the wire rope starts to move, the CCD camera simultaneously acquires images of the wire rope.

[0073] S32. The vision acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module.

[0074] S33. The image processing module performs grayscale processing on the image and then enhances the contrast to obtain a clearer image, and then sends the processed image to the computer.

[0075] S4. The synchronous encoder is used to record the movement position of the wire rope in real time. The computer records the magnetoresistive signal and visual signal at the same position based on the movement position of the wire rope.

[0076] S5. The computer constructs a weighted decision fusion model by combining magnetoresistive and visual signals. It determines the corrosion depth level based on magnetic field anomalies and the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spots, thus obtaining the final corrosion level, including:

[0077] S51. Calculate the magnetic field anomaly value using the leakage magnetic field signal curve, and determine the corrosion depth level and the corresponding corrosion depth reliability.

[0078] S52. Extract the two-dimensional morphological features of the rusted area from the image, calculate the rust area ratio and the fractal dimension of the rust spots, and determine the rust area level and the corresponding rust area confidence level.

[0079] S53. Based on the weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence, corrosion area level, and corrosion area confidence.

[0080] Specifically, S51 uses the leakage magnetic field signal curve to calculate the magnetic field anomaly value and determine the corrosion depth level and the corresponding corrosion depth reliability, including:

[0081] S511. Considering that under constant excitation, according to Ohm's law for magnetic circuits and the magnetic properties of materials, the magnetic reluctance R of the wire rope... m Changes occur, causing distortion of the magnetic field distribution in the space near its surface. For uniform corrosion, due to the poor magnetic permeability of iron oxides, the relative permeability of the corrosion region is approximately considered to be... The magnetic field anomaly value decreases. According to magnetic field theory, the surface normal magnetic induction intensity is related to the material state. The following formula is used to calculate the magnetic field anomaly value:

[0082] ;

[0083] in, Let B be the magnetic field anomaly value at position x along the axial direction of the wire rope. measured (x) represents the leakage magnetic signal curve at position x along the axial direction of the wire rope, B baseline This represents the baseline value of the magnetic field at a defect-free location on the wire rope. Let x be the corrosion depth at position x along the axial direction of the wire rope. The cross-sectional area loss is at position x along the axial direction of the wire rope. k1 and k2 are constants, determined through calibration experiments, reflecting the influence of the geometric steepness of the defect on the leakage magnetic field.

[0084] S512. By using the calibration curve and signal-to-noise ratio, the magnetic field anomaly value is determined. Preliminary estimate of corrosion depth level L m and the corresponding rust depth confidence level C x .

[0085] For S512:

[0086] 1) The magnetic field anomaly value is obtained through the calibration curve. Preliminary estimate of corrosion depth level L m The following technical means are used to achieve this:

[0087] Preparation of standard rust samples: Select samples with the same material and specifications as the steel wire rope to be tested, and artificially create uniform rust defects of different depths (such as rust depths of 0.1mm, 0.3mm, 0.5mm, 1.0mm, etc., to cover the actual possible rust range).

[0088] Collect leakage magnetic signals of standard samples: Magnetize the standard samples to saturation through the magnetizer of this system, collect leakage magnetic signals at the same rope speed (such as 2m / s in the embodiment) and sampling rate (1kHz), and calculate the magnetic field anomaly value of each standard sample.

[0089] Constructing a calibration curve: Using "actual corrosion depth" as the abscissa and "corresponding magnetic field anomaly value" as the ordinate, fit the function relationship curve between the two (i.e., the calibration curve), and at the same time determine the threshold range of magnetic field anomaly value corresponding to different corrosion depths.

[0090] To test the magnetic field anomaly at a certain location on the steel wire rope Query the actual corrosion depth range corresponding to the calibration curve, match it with the preset corrosion depth level (such as light, moderate, heavy, etc.), and obtain the corrosion depth level. L m .

[0091] 2) The reliability of the preliminary estimation of the corresponding corrosion depth based on the signal-to-noise ratio. C x The following technical means are used to achieve this:

[0092] Typically, the signal from the defect-free area of ​​the wire rope (near the magnetic field baseline Bbaseline) is selected as the noise reference. The signal variance (representing the noise intensity) of this area is calculated, and then the variance of the leakage magnetic signal in the rusted area (representing the signal + noise intensity) is calculated. The signal-to-noise ratio (SNR) is obtained by comparing the two.

[0093] The reliability of the signal-to-noise ratio (SNR) and corrosion depth was established through experimental calibration. C x Based on the mapping relationship, the reliability of the corresponding corrosion depth can be initially estimated through the signal-to-noise ratio. C x .

[0094] Specifically, in S52, two-dimensional morphological features of the rusted area are extracted from the image, the rust area ratio and fractal dimension of the rust spots are calculated, and the rust area level and corresponding rust area confidence level are determined, including:

[0095] S521. Segment the rusted areas in the image, separating them from the background, and set the mask function:

[0096] ;

[0097] Among them, Mask(x) p ,y p ) represents pixel coordinates (x) p ,y p The mask value at H(x) is 1 when it represents a rusted area and 0 when it represents the background. p ,y p ), S(x p ,y p ) are pixel coordinates (x p ,y p Hue and saturation at ) H min H max These are the minimum and maximum values ​​of the hue, S. th This is the saturation threshold.

[0098] S522. By removing small noise points, connecting fracture areas, and extracting features from rusted areas, the rust area ratio is calculated:

[0099] ;

[0100] Among them, R area Pixel cor Pixel total These represent the total number of pixels showing rust spots on the surface of the steel wire rope and the total number of pixels, respectively.

[0101] S523. To describe the roughness and severity of rust, the box counting method is used to approximate the fractal dimension of rust spots:

[0102] ;

[0103] in, The required side length to cover the outline of the rust spot is The number of boxes, D f The fractal dimension of the rust spots;

[0104] S524. Based on image clarity and segmentation effect, the ratio of rust area... R area and rust fractal dimension D f Preliminary estimate of rust area level L v and the credibility of the corresponding rust area C v .

[0105] For S524, based on the ratio of rusted area... R area and rust fractal dimensionD f Preliminary estimate of rust area level L v and the credibility of the corresponding rust area C v The following technical means are used to achieve this:

[0106] First, compare the rusted area. R area and rust fractal dimension D f Perform preprocessing such as outlier filtering and smoothing;

[0107] By calibrating with standard rust samples, a system was established that... R area + D f "and rust area level" L v Two-dimensional threshold mapping relationship (such as mild, moderate, severe, etc.);

[0108] Substitute the parameters to be detected into the mapping relationship, and combine multi-frame voting to preliminarily estimate the corrosion area level. L v ;

[0109] The image sharpness (e.g., Laplacian variance) and segmentation performance (e.g., edge continuity, background purity) are quantified, and their scores are obtained through weighted fusion. These scores are then combined with dynamic corrections for scenarios such as pitting corrosion to ultimately determine the reliability of the corrosion area. C v (The range is 0 to 1, and the higher the value, the more reliable it is).

[0110] Specifically, S53 uses a weighted decision fusion model to calculate the final corrosion grade by combining the corrosion depth grade, corrosion depth confidence level, corrosion area grade, and corrosion area confidence level, including:

[0111] The final corrosion grade is calculated using the following formula:

[0112] ;

[0113] Among them, L final For the final corrosion grade, for internal or uniform rust with strong magnetic reluctance signals but inconspicuous visual signals, the rust depth confidence level C is... x Automatically increasing the reliability C of the corrosion area for pitting corrosion with weak magnetoresistance signal and obvious visual signal. v Automatically increases in size.

[0114] Based on the aforementioned method for quantitative detection of steel wire rope corrosion using a combination of magnetoresistive sensing and vision, this application further discloses a system for quantitative detection of steel wire rope corrosion based on a combination of magnetoresistive sensing and vision, such as... Figure 2 As shown, it includes:

[0115] The magnetoresistive sensor array is in a ring structure with multiple (ten) magnetoresistive sensors evenly distributed. When the wire rope is magnetized (by passing through a magnetizer) and pulled by the wire rope traction device (at a speed of 2 m / s), the circumferential leakage magnetic signal at each position on the wire rope is collected (at a sampling rate of 1 kHz).

[0116] The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic field signals acquired by each magnetoresistive sensor, obtains a one-dimensional leakage magnetic field signal curve, and sends it to the signal conditioning module.

[0117] The signal conditioning module conditions the leakage flux signal curve and sends the conditioned leakage flux signal curve to the computer.

[0118] CCD cameras (CCD high-resolution anti-flicker cameras) are used to simultaneously capture images of the steel wire rope when it starts to move. Each CCD camera is equipped with a supplementary light to facilitate clear imaging.

[0119] The visual acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module;

[0120] The image processing module performs grayscale processing on the image, then enhances the contrast to obtain a clearer image, and sends the processed image to the computer.

[0121] The computer records the magnetoresistive and visual signals at the same position of the wire rope in real time based on the synchronous encoder. It then constructs a weighted decision fusion model by combining the magnetoresistive and visual signals. Based on the magnetic field anomaly value, it determines the corrosion depth level, and based on the corrosion area ratio and the fractal dimension of the rust spot, it determines the corrosion area level, thus obtaining the final corrosion level.

[0122] Specifically, the computer constructs a weighted decision fusion model by combining magnetoresistive and visual signals. It determines the corrosion depth level based on magnetic field anomalies and the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spots, thus obtaining the final corrosion level, including:

[0123] The magnetic field anomaly value was calculated using the leakage magnetic field signal curve, and the corrosion depth level and the corresponding corrosion depth confidence level were determined.

[0124] Two-dimensional morphological features of the rusted area are extracted from the image, the rust area ratio and rust spot fractal dimension are calculated, and the rust area level and corresponding rust area confidence level are determined.

[0125] Based on a weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence level, corrosion area level, and corrosion area confidence level.

[0126] The technical solution of this application simultaneously acquires local magnetic field signals and surface image information of the wire rope, uses a magnetoresistive sensor to deeply perceive the changes in magnetic properties and cross-sectional area loss caused by corrosion, and uses machine vision to accurately identify the surface morphology features of the corrosion. Finally, a weighted decision fusion model is used to output the final corrosion level, realizing accurate quantitative detection of wire rope corrosion. This provides an innovative solution for improving the detection accuracy of wire rope corrosion and the comprehensive evaluation capability of defects.

[0127] Currently, nondestructive testing methods for detecting steel wire rope corrosion mainly rely on electromagnetic or visual methods for analysis, with few solutions integrating multiple analysis methods. The quantitative detection method and system for steel wire rope corrosion based on magnetoresistive sensing and vision proposed in this invention has unique engineering application value.

[0128] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined method, characterized in that: Includes the following steps: S1. Use a magnetizer to magnetize the steel wire rope until it is saturated; S2. Start the wire rope traction device and synchronous encoder. The wire rope starts to move. The magnetoresistive sensor array collects the magnetoresistive signals at various positions on the wire rope. The magnetoresistive acquisition module sends the collected magnetoresistive signals to the signal conditioning module for signal conditioning. S3. When the wire rope starts to move, the CCD camera synchronously acquires images of the wire rope, and the vision acquisition card sends the acquired visual signals to the image processing module for image processing. S4. The synchronous encoder is used to record the movement position of the wire rope in real time. The computer records the magnetoresistive signal and visual signal at the same position based on the movement position of the wire rope. S5. The computer constructs a weighted decision fusion model by combining magnetoresistive and visual signals. It determines the corrosion depth level based on magnetic field anomalies and the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spots, thus obtaining the final corrosion level, including: S51. Calculate the magnetic field anomaly value using the magnetic flux leakage signal curve, and determine the corrosion depth level and the corresponding corrosion depth reliability, specifically including: S511. Considering that under constant excitation, according to Ohm's law for magnetic circuits and the magnetic properties of materials, the magnetic reluctance R of the wire rope... m Changes occur, causing distortion of the magnetic field distribution in the space near its surface. For uniform corrosion, due to the poor magnetic permeability of iron oxides, the relative permeability of the corrosion region is approximately considered to be... The magnetic field anomaly value decreases. According to magnetic field theory, the surface normal magnetic induction intensity is related to the material state. The following formula is used to calculate the magnetic field anomaly value: ； in, Let B be the magnetic field anomaly value at position x along the axial direction of the wire rope. measured (x) represents the leakage magnetic signal curve at position x along the axial direction of the wire rope, B baseline This represents the baseline value of the magnetic field at a defect-free location on the wire rope. Let x be the corrosion depth at position x along the axial direction of the wire rope. The cross-sectional area loss is at position x along the axial direction of the wire rope. k1 and k2 are constants, determined through calibration experiments, reflecting the influence of the geometric steepness of the defect on the leakage magnetic field. S512. By using the calibration curve and signal-to-noise ratio, the magnetic field anomaly value is determined. Preliminary estimate of corrosion depth level L m and the corresponding rust depth confidence level C x ; S52. Extract the two-dimensional morphological features of the rusted area from the image, calculate the rust area ratio and the fractal dimension of the rust spots, and determine the rust area level and the corresponding rust area confidence level. S53. Based on the weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence, corrosion area level, and corrosion area confidence.

2. The quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined according to claim 1, characterized in that: In S2, the wire rope traction device and synchronous encoder are activated, and the wire rope begins to move. The magnetoresistive sensor array collects magnetoresistive signals at various locations on the wire rope. The magnetoresistive acquisition module sends the collected magnetoresistive signals to the signal conditioning module for signal conditioning, including: S21. Start the wire rope traction device and synchronous encoder. The wire rope starts to move. Each magnetoresistive sensor in the magnetoresistive sensor array collects the circumferential leakage magnetic signal at each position on the wire rope. S22. The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic signals acquired by each magnetoresistive sensor to obtain a one-dimensional leakage magnetic signal curve, and sends it to the signal conditioning module. S23. The signal conditioning module conditions the leakage magnetic signal curve and sends the conditioned leakage magnetic signal curve to the computer.

3. The quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined according to claim 2, characterized in that: When the wire rope in S3 begins to move, the CCD camera simultaneously acquires images of the wire rope. The vision acquisition card sends the acquired visual signals to the image processing module for image processing, including: S31. When the wire rope starts to move, the CCD camera simultaneously acquires images of the wire rope. S32. The vision acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module. S33. The image processing module performs grayscale processing on the image and then enhances the contrast to obtain a clearer image, and then sends the processed image to the computer.

4. The quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined according to claim 1, characterized in that: S52 extracts the two-dimensional morphological features of the rusted area from the image, calculates the rust area ratio and the fractal dimension of the rust spots, and determines the rust area level and the corresponding rust area confidence level, including: S521. Segment the rusted areas in the image, separating them from the background, and set the mask function: ； Among them, Mask(x) p ,y p ) represents pixel coordinates (x) p ,y p The mask value at H(x) is 1 when it represents a rusted area and 0 when it represents the background. p ,y p ), S(x p ,y p ) are pixel coordinates (x p ,y p Hue and saturation at ) H min H max These are the minimum and maximum values ​​of the hue, S. th This is the saturation threshold. S522. By removing small noise points, connecting fracture areas, and extracting features from rusted areas, the rust area ratio is calculated: ； Among them, R area Pixel cor Pixel total These represent the total number of pixels showing rust spots on the surface of the steel wire rope and the total number of pixels, respectively. S523. To describe the roughness and severity of rust, the box counting method is used to approximate the fractal dimension of rust spots: ； in, The required side length to cover the outline of the rust spot is The number of boxes, D f The fractal dimension of the rust spots; S524. Based on image clarity and segmentation effect, the ratio of rust area R... area and the fractal dimension D of rust spots f Preliminary estimate of corrosion area level L v and the corresponding rust area reliability C v .

5. The quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined according to claim 4, characterized in that: S53 uses a weighted decision fusion model that combines corrosion depth level, corrosion depth confidence level, corrosion area level, and corrosion area confidence level to calculate the final corrosion level, including: The final corrosion grade is calculated using the following formula: ； Among them, L final For the final corrosion grade, for internal or uniform rust with strong magnetic reluctance signals but inconspicuous visual signals, the rust depth confidence level C is... x Automatically increasing the reliability C of the corrosion area for pitting corrosion with weak magnetoresistance signal and obvious visual signal. v Automatically increases in size.

6. A quantitative detection system for steel wire rope corrosion based on magnetoresistive sensing-vision combined method, used to execute the quantitative detection method for steel wire rope corrosion based on magnetoresistive sensing-vision combined method as described in claim 1, characterized in that: include: The magnetoresistive sensor array has a ring structure with multiple magnetoresistive sensors evenly distributed. When the wire rope is magnetized and pulled by the wire rope traction device, it collects the circumferential leakage magnetic signals at various positions on the wire rope. The magnetoresistive acquisition module receives and processes the circumferential leakage magnetic field signals acquired by each magnetoresistive sensor, obtains a one-dimensional leakage magnetic field signal curve, and sends it to the signal conditioning module. The signal conditioning module conditions the leakage flux signal curve and sends the conditioned leakage flux signal curve to the computer. When the steel wire rope begins to move, a pair of CCD cameras simultaneously acquire images of the steel wire rope. Each CCD camera is equipped with a supplementary light to facilitate clear imaging. The visual acquisition card receives and processes images acquired by the CCD camera, and sends each frame of the image to the image processing module; The image processing module performs grayscale processing on the image, then enhances the contrast to obtain a clearer image, and sends the processed image to the computer. The computer records the magnetoresistive and visual signals at the same position of the wire rope in real time based on the synchronous encoder. It then constructs a weighted decision fusion model by combining the magnetoresistive and visual signals. Based on the magnetic field anomaly value, it determines the corrosion depth level, and based on the corrosion area ratio and the fractal dimension of the rust spot, it determines the corrosion area level, thus obtaining the final corrosion level.

7. The quantitative detection system for steel wire rope corrosion based on magnetoresistive sensing-vision combined as described in claim 6, characterized in that: The computer constructs a weighted decision fusion model by combining magnetoresistive and visual signals. It determines the corrosion depth level based on magnetic field anomalies and the corrosion area level based on the corrosion area ratio and the fractal dimension of the rust spots, thus obtaining the final corrosion level, including: The magnetic field anomaly value was calculated using the leakage magnetic field signal curve, and the corrosion depth level and the corresponding corrosion depth confidence level were determined. Two-dimensional morphological features of the rusted area are extracted from the image, the rust area ratio and rust spot fractal dimension are calculated, and the rust area level and corresponding rust area confidence level are determined. Based on a weighted decision fusion model, the final corrosion level is calculated by combining the corrosion depth level, corrosion depth confidence level, corrosion area level, and corrosion area confidence level.

Images