Line scanning-based part surface flaw detection method, system and device

Abstract

The application relates to the technical field of nondestructive testing, and provides a part surface flaw detection method, system and equipment based on line scanning, which solves the problems of easy missed detection of defect detection and inability to comprehensively evaluate the expansion risk of composite defects. The method comprises the following steps: scanning the surface of a bogie by using a line scanning device, extracting temperature field distribution data and temperature gradient distribution data from the scanning line data set; generating an abnormal candidate area based on the temperature gradient distribution data; obtaining phase shift data of an eddy current signal in the abnormal candidate area, and assigning a defect classification mark to the abnormal candidate area; analyzing the spatial coordinates of the temperature gradient distribution data and the defect classification mark, and generating a reference parameter; and extracting a target defect area according to the spatial distribution weight of the reference parameter, so as to realize flaw detection of the surface of the bogie. The application improves the accuracy, reliability and efficiency of high-speed rail bogie surface defect detection, and provides strong technical support for guaranteeing the safe operation of high-speed rail.

Description

Technical Field

[0001] This invention belongs to the field of nondestructive testing technology, specifically relating to a method, system, and equipment for surface flaw detection of parts based on line scanning. Background Technology

[0002] As the core equipment of high-speed rail transit, high-speed rail's key components are subjected to high-frequency alternating loads and complex environmental conditions over long periods, making them prone to defects such as surface corrosion and fatigue cracks. Among these, the bogie, as the core component that bears the weight of the car body and transmits traction force, faces severe challenges in defect detection.

[0003] Currently, the mainstream solution for defect detection of bogies, a high-speed rail component, is deep defect detection technology based on ultrasonic phased arrays. This involves emitting ultrasonic signals onto the bogie surface using a phased array probe and receiving feedback information; calculating the defect depth based on ultrasonic propagation time and material sound velocity; and distinguishing defect types based on feedback information and waveform characteristics. Existing solutions have several core problems, such as limited penetration of ultrasonic waves into surface corrosion and oxide layers, resulting in limited information acquisition and a high risk of missed defects; and the inability to comprehensively assess the expansion risk of complex defects through single-data analysis. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for surface flaw detection of parts based on line scanning, so as to solve the problems of easy omission in defect detection and inability to fully assess the expansion risk of compound defects in the prior art.

[0005] To address the aforementioned technical problems, in a first aspect, this application provides a method for surface flaw detection of parts based on line scanning, comprising:

[0006] The surface of the bogie is scanned using a line scanning device, and a scan line dataset of the bogie surface is collected. Temperature field distribution data is then extracted from the scan line dataset.

[0007] Temperature gradient distribution data is extracted from the temperature field distribution data, and anomaly candidate regions are generated based on the temperature gradient distribution data.

[0008] The phase offset data of the eddy current signal within the abnormal candidate region is obtained, and a defect classification identifier is assigned to the abnormal candidate region based on the phase offset data.

[0009] The spatial coordinates of the temperature gradient distribution data and the defect classification identifier are analyzed to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, the target defect region is extracted from the anomaly candidate region to achieve flaw detection on the bogie surface.

[0010] Optionally, phase offset data of eddy current signals within the anomaly candidate region is obtained, and a defect classification identifier is assigned to the anomaly candidate region based on the phase offset data, including:

[0011] The raw phase data of eddy current signals in different frequency bands within the anomaly candidate region are collected, and noise suppression and baseline correction are performed on the raw phase data to generate phase offset data.

[0012] The phase offset data is compensated based on the bogie's material properties, surface oxide layer thickness, and geometric feature parameters of the abnormal candidate region to obtain offset difference data.

[0013] Based on the offset difference data, a defect classification identifier is assigned to the abnormal candidate region.

[0014] Optionally, based on the offset difference data, a defect classification identifier is assigned to the anomaly candidate region, including:

[0015] Based on the bogie's material fatigue property database and historical load data, a depth amplitude decay curve is generated.

[0016] Based on the depth amplitude attenuation curve, the offset difference data is subjected to intensity conversion and feature fusion processing to obtain the vector synthesis result;

[0017] Based on the vector synthesis results, defect classification identifiers are assigned to the abnormal candidate regions.

[0018] Optionally, based on the vector synthesis result, a defect classification label is assigned to the anomaly candidate region, including:

[0019] Based on the anisotropy coefficient and stress concentration factor in the bogie material fatigue property database, a proportional threshold is set;

[0020] Based on the surface curvature parameters and motion speed of the bogie, the cumulative energy value in the direction of the defect depth in the vector synthesis result is calculated to generate a cumulative lower limit value;

[0021] Based on the aforementioned proportional threshold and cumulative lower limit, the vector synthesis result is judged to assign a defect classification identifier to the abnormal candidate region.

[0022] Optionally, the spatial coordinates of the temperature gradient distribution data and the defect classification identifier are analyzed to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, the target defect region is extracted from the anomaly candidate region, including:

[0023] Based on the bogie's motion speed and the sampling frequency of the line scanning device, the spatiotemporal deviation values ​​of the temperature gradient distribution data and phase offset data are calculated.

[0024] Using the spatiotemporal deviation value, the timestamps of the temperature gradient distribution data are interpolated and compensated to generate a target dataset. The spatial coordinates of the defect classification identifiers are curvature compensated to generate a target identifier set. The target dataset includes the temperature gradient amplitude.

[0025] The target dataset and the target identifier set are coupled and analyzed to calculate the rate of change of temperature gradient magnitude and the divergence of phase offset;

[0026] Based on the rate of change of the temperature gradient amplitude and the divergence of the phase offset, reference parameters are generated;

[0027] Based on the spatial distribution weights of the benchmark parameters, combined with the temperature gradient amplitude, offset difference data, and defect classification identifier, the target defect region is extracted from the anomaly candidate region.

[0028] Optionally, based on the spatial distribution weights of the reference parameters, combined with the temperature gradient amplitude, offset difference data, and the defect classification identifier, the target defect region is extracted from the anomaly candidate region, including:

[0029] Based on the spatial distribution weights of the reference parameters and the surface curvature parameters of the bogie, the mesh resolution parameters of the anomaly candidate region are calculated.

[0030] Based on the grid resolution parameters, the anomaly candidate region is divided into grids to generate multiple grid regions;

[0031] Based on the temperature gradient amplitude, offset difference data and defect classification identifier of each grid region, the feature similarity of adjacent grid regions is calculated, and adjacent grid regions whose feature similarity exceeds a preset merging threshold are merged into an initial defect region.

[0032] Based on the crack morphology template in the bogie material fatigue property database, regions in the initial defect region whose centroid offset exceeds a preset deformation threshold are removed to obtain the target defect region.

[0033] Optionally, temperature gradient distribution data is extracted from the temperature field distribution data, and anomaly candidate regions are generated based on the temperature gradient distribution data, including:

[0034] Based on the bogie's motion speed and the scanning parameters of the line scanning device, the spacing compensation amount between adjacent scanning lines is generated.

[0035] The slope of temperature change along the bogie's direction of motion within each scan line is extracted from the temperature field distribution data, and combined with the spacing compensation amount, temperature gradient distribution data is generated.

[0036] The absolute value of the temperature difference between adjacent scan lines in the temperature gradient distribution data is calculated, and noise suppression processing is performed on the absolute value of the temperature difference based on the thermal diffusivity of the bogie and the thickness of the surface oxide layer to generate target absolute value of temperature difference data.

[0037] Select the target adjacent scan lines whose absolute temperature difference exceeds a preset temperature difference value from the target absolute temperature difference data;

[0038] The adjacent scan lines of the target are subjected to region extension and closing operations along the movement direction of the bogie to generate abnormal candidate regions.

[0039] Secondly, this application provides a surface flaw detection system for parts based on line scanning, comprising:

[0040] The acquisition module is used to scan the surface of the bogie using a line scanning device and acquire a scan line dataset of the bogie surface, and extract temperature field distribution data from the scan line dataset.

[0041] The generation module is used to extract temperature gradient distribution data from the temperature field distribution data and generate anomaly candidate regions based on the temperature gradient distribution data.

[0042] The acquisition module is used to acquire phase offset data of eddy current signals within the abnormal candidate region, and assign defect classification identifiers to the abnormal candidate region based on the phase offset data.

[0043] The detection module is used to analyze the temperature gradient distribution data and the spatial coordinates of the defect classification identifier to generate benchmark parameters. Based on the spatial distribution weight of the benchmark parameters, the target defect area is extracted from the abnormal candidate area to realize the flaw detection of the bogie surface.

[0044] Thirdly, this application provides a computing device including a processor and a memory, wherein the memory stores a computer program, and the processor is configured to run the computer program to perform a line scan-based part surface flaw detection method as described in any of the first aspects.

[0045] Fourthly, this application provides a computer storage medium storing computer program instructions thereon, which, when executed by a processor, implement the line-scan-based flaw detection method for part surfaces as described in any one of the first aspects.

[0046] The beneficial effects of this application are:

[0047] This application provides a line-scan-based surface flaw detection method for parts. It scans the bogie surface using a line-scanning device, collects scan line datasets, and extracts temperature field distribution data. This overcomes the penetration limitations of traditional detection techniques on surface coverings, effectively capturing thermal response information of the part surface and near-surface. Based on temperature gradient distribution data, it generates anomaly candidate regions, achieving precise preliminary identification of suspicious areas and reducing invalid detection ranges. It acquires phase offset data of eddy current signals within the anomaly candidate regions and assigns defect classification labels, supplementing feature information in the defect depth direction and avoiding information loss due to a single data dimension. By analyzing the spatial coordinates of temperature gradient distribution data and defect classification labels to generate benchmark parameters, and combining the spatial distribution weights of the benchmark parameters, it extracts target defect regions from the anomaly candidate regions. This achieves collaborative analysis of thermal and electromagnetic field data, comprehensively assessing the overall risk of defects, and solving the technical problems of existing solutions that struggle to simultaneously consider surface and deep features, are prone to missed detections, and cannot comprehensively assess complex defects.

[0048] Furthermore, by collecting raw phase data of eddy current signals at different frequency bands within the anomaly candidate area, and generating phase offset data through noise suppression and baseline correction, the phase offset data is compensated by combining the bogie's material property parameters, surface oxide layer thickness, and geometric feature parameters of the anomaly candidate area to obtain offset difference data, thereby assigning defect classification labels to the anomaly candidate area. Through multi-band eddy current signal acquisition and noise reduction correction, the reliability of the raw signal is improved. Combined with compensation processing of material, oxide layer, and geometric parameters, the interference of non-defect factors on the detection results is eliminated, effectively uncovering the essential characteristics of defects, enriching the basis for defect classification, and solving the problem of insufficient information acquisition and inaccurate classification caused by the influence of surface coverings and the properties of the parts themselves in existing solutions. This provides reliable feature support for the subsequent accurate extraction of target defect areas. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating a surface flaw detection method for parts based on line scanning, provided for an embodiment of this application;

[0051] Figure 2 A schematic diagram of a part surface flaw detection system based on line scanning provided in an embodiment of this application;

[0052] Figure 3 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0053] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0054] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0055] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] Figure 1 This application provides a flowchart of a method for detecting surface defects in parts based on line scanning, as shown in the embodiments. Figure 1 As shown, the method includes:

[0057] To address the shortcomings of existing bogie inspection technologies, such as insufficient penetration of surface coverings, limited information acquisition leading to missed detections, and the difficulty of comprehensively assessing the risk of compound defect expansion through single data analysis, this application utilizes a line scanning device to scan and collect a dataset of scan lines on the bogie surface and extract temperature field distribution data. Temperature gradient distribution data is then extracted from the temperature field distribution data to generate anomaly candidate regions, achieving preliminary screening of suspicious areas. Subsequently, phase offset data of eddy current signals within the anomaly candidate regions is acquired and assigned defect classification labels, supplementing defect depth direction feature information to enrich the data dimensions. Finally, benchmark parameters are generated by analyzing the spatial coordinates of the temperature gradient distribution data and defect classification labels. Based on the spatial distribution weights of the benchmark parameters, target defect regions are extracted from the anomaly candidate regions, forming a multi-dimensional data collaborative analysis detection logic. This overcomes the limitations of single data analysis and comprehensively covers the detection needs of surface and deep defect features.

[0058] Based on this, this application provides a method for surface flaw detection of parts based on line scanning, such as... Figure 1 ,include:

[0059] Step 101: Use a line scanning device to scan the surface of the bogie and collect a scan line dataset of the bogie surface, and extract temperature field distribution data from the scan line dataset.

[0060] In this step, the temperature field distribution data refers to the temperature information at various locations on the bogie surface collected by the line scanning device. This data includes the temperature values ​​corresponding to different spatial points on the bogie surface.

[0061] In this embodiment, a line scanning device is used to move at a constant speed along a preset path on the bogie surface, and the built-in temperature sensing module collects the raw temperature data of the bogie surface point by point. All the collected raw data are then arranged in the scanning order to form a scan line dataset. The scan line dataset is then traversed row by row and column by column to extract the temperature value corresponding to each data point. The data points are then arranged according to their actual spatial positions to finally generate temperature field distribution data.

[0062] Step 102: Extract temperature gradient distribution data from the temperature field distribution data, and generate anomaly candidate regions based on the temperature gradient distribution data.

[0063] In this step, temperature gradient distribution data refers to a physical quantity calculated based on the temperature difference between adjacent spatial points in the temperature field distribution data. This temperature gradient distribution data includes the rate of temperature change of each spatial point along different directions.

[0064] Anomaly candidate regions refer to bogie surface areas with abnormal temperature changes identified based on temperature gradient distribution data.

[0065] In this embodiment, firstly, multiple adjacent spatial points are selected around each spatial point in the temperature field distribution data. Then, the temperature difference between the spatial point and each adjacent spatial point is calculated. Next, each temperature difference is divided by the straight-line distance between the corresponding adjacent spatial point and the spatial point to obtain the temperature change rate in each direction. The temperature change rates in all directions of all spatial points are integrated to form temperature gradient distribution data. Subsequently, a temperature gradient threshold suitable for bogie detection is set, and the temperature gradient distribution data is traversed. Spatial points with temperature change rates exceeding the threshold are marked. Then, adjacent marked spatial points are connected and integrated to finally form anomaly candidate regions.

[0066] Step 103: Obtain the phase offset data of the eddy current signal within the abnormal candidate region, and assign a defect classification identifier to the abnormal candidate region based on the phase offset data.

[0067] In this step, the phase offset data refers to the data obtained after processing the electromagnetic response signal in the abnormal candidate region acquired by eddy current detection technology, which is used to reflect the changes in the internal structure of the material in that region.

[0068] Defect classification identifiers refer to markers used to distinguish defect types based on the spatial distribution differences of phase offset data. These identifiers include specific identifiers set for different structural change characteristics.

[0069] Step 104: Analyze the spatial coordinates of the temperature gradient distribution data and the defect classification identifier to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, extract the target defect region from the abnormal candidate region to achieve flaw detection on the bogie surface.

[0070] In this step, the benchmark parameter refers to the comprehensive parameter obtained by collaborative analysis based on temperature gradient distribution data and the spatial coordinates of defect classification markers. The benchmark parameter includes quantitative indicators that reflect the degree of spatial correlation between temperature change characteristics and defect markers.

[0071] Spatial distribution weight refers to the weight value assigned to different locations within the abnormal candidate region based on the baseline parameters, which is used to reflect the priority of each location becoming a real defect region.

[0072] This application embodiment achieves comprehensive collection of bogie surface temperature information, providing basic data for subsequent inspection; it can accurately locate temperature anomaly areas, reducing the scope of invalid inspections; it enriches the inspection dimensions by acquiring and classifying defect-related electromagnetic features; and it improves the comprehensiveness and accuracy of inspection by fusing multi-dimensional data to extract real defect areas, solving the problems of insufficient information acquisition and incomplete evaluation in traditional technologies.

[0073] This application provides a specific embodiment. Step 103 involves obtaining phase offset data of the eddy current signal within the abnormal candidate region, and assigning a defect classification identifier to the abnormal candidate region based on the phase offset data. This specifically includes the following steps:

[0074] Step 301: Collect the original phase data of eddy current signals in different frequency bands within the abnormal candidate region, perform noise suppression and baseline correction processing on the original phase data, and generate phase offset data.

[0075] In this step, the raw phase data refers to the phase information obtained by collecting electromagnetic induction signals at various locations within the anomaly candidate area through eddy current sensors at different preset frequency bands. This data includes the phase value of each spatial point corresponding to each frequency band.

[0076] In this embodiment, multiple non-overlapping electromagnetic signal frequency bands are first set, and the emission parameters of the eddy current sensor are adjusted so that it operates in each frequency band in sequence. Then, the sensor is controlled to scan point by point along the surface of the abnormal candidate region and collect the original electromagnetic signal at each location in each frequency band.

[0077] Next, the phase values ​​of each original electromagnetic signal are extracted and stored according to frequency band and spatial location to form the original phase data. Then, a moving average method is used for noise suppression, replacing each phase value with the average of the phase values ​​of a preset number of neighboring points to eliminate random interference. Then, a known defect-free standard area on the bogie is selected, and the phase data of this standard area in the same frequency band is collected as baseline data. The baseline correction is completed by subtracting the corresponding frequency band baseline data value from each value in the original phase data, and finally generating the phase offset data.

[0078] Step 302: Based on the bogie's material property parameters, surface oxide layer thickness, and geometric feature parameters of the abnormal candidate region, the phase offset data is compensated to obtain offset difference data.

[0079] In this step, geometric feature parameters refer to the spatial structure information of the abnormal candidate regions extracted based on the scan data. These parameters include the shape of the region, boundary size, surface curvature, and surface roughness.

[0080] Offset difference data refers to the phase offset data obtained after compensation and correction by material property parameters, surface oxide layer thickness and geometric feature parameters. This data includes the corrected phase offset values ​​at each position in each frequency band, which is used to accurately reflect the differences in electromagnetic response caused by the defect itself.

[0081] In this embodiment, material property parameters, including the conductivity and permeability of the material, are first retrieved from a preset bogie parameter database. Then, a specific frequency signal is emitted to the bogie using an eddy current sensor, and the thickness of the surface oxide layer of the bogie is calculated based on the signal reflection intensity. Next, by analyzing the scan coordinate data of the abnormal candidate region, the boundary range and shape characteristics of the abnormal candidate region are determined, and the curvature and unevenness of each point within the abnormal candidate region are calculated to obtain geometric feature parameters.

[0082] Next, a phase correction coefficient is calculated based on the differences between the material's conductivity and permeability and the standard material parameters. The phase offset data is multiplied by this phase correction coefficient to obtain the corrected value, thus eliminating the influence of the material's inherent properties. Then, the oxide layer influence value is calculated based on the signal attenuation law corresponding to the surface oxide layer thickness. The oxide layer influence value is subtracted from the corrected value to obtain the secondary corrected value, thus eliminating the interference of the oxide layer. Finally, based on the electromagnetic signal distortion compensation rules corresponding to the curvature and unevenness of each point in the geometric feature parameters, the secondary corrected value at each position is fine-tuned to compensate for the error caused by the surface structure difference, ultimately obtaining the offset difference data.

[0083] Step 303: Assign defect classification identifiers to the abnormal candidate regions based on the offset difference data.

[0084] Optionally, step 303, assigning defect classification labels to the anomaly candidate regions based on the offset difference data, specifically includes the following steps:

[0085] Step 311: Generate depth amplitude decay curves based on the bogie material fatigue property database and historical load data.

[0086] In this step, the material fatigue characteristics database refers to the collection that stores data such as fatigue life, crack propagation rate, and mechanical property decay law of the materials used in the bogie under different load conditions. This database includes fatigue characteristic parameters of the material under various stress states.

[0087] Historical load data refers to the load information recorded during the bogie's past operation. This data includes the load magnitude, load frequency, load direction, and duration at different operating stages.

[0088] The depth amplitude decay curve is a curve established based on material fatigue characteristics and historical load data, which describes the decay relationship between defect depth and electromagnetic response signal amplitude. This curve reflects the signal amplitude change trend corresponding to different defect depths.

[0089] In this embodiment, fatigue characteristic parameters of bogie materials are first retrieved from a preset bogie material fatigue characteristic database; then, load records of past bogie operations are collected, and data such as load magnitude, frequency, and direction are organized in chronological order to form historical load data; then, the degree of fatigue damage of materials under different load conditions in the historical load data is statistically analyzed, and combined with the material fatigue characteristic parameters, the variation law of electromagnetic response signal amplitude corresponding to different defect depths is determined; finally, a depth amplitude attenuation curve is plotted with defect depth as the abscissa and signal amplitude as the ordinate.

[0090] Step 312: Based on the depth amplitude attenuation curve, perform intensity conversion and feature fusion processing on the offset difference data to obtain the vector synthesis result.

[0091] In this step, the vector synthesis result refers to the vector data obtained by fusing different directional features after intensity conversion of the offset difference data. The result includes the signal intensity magnitude and directional information at each location, which is used to comprehensively reflect the depth characteristics and spatial expansion trend of the defect.

[0092] In this embodiment of the application, the values ​​at each position in the offset difference data are first traversed, and the defect depth corresponding to each value is found according to the depth amplitude attenuation curve. The values ​​of the offset difference data are then converted into intensity values ​​that match the defect depth to complete the intensity conversion.

[0093] The geometric feature parameters of the candidate anomaly region are then analyzed to determine the possible defect propagation directions, including surface propagation direction and depth propagation direction. The converted intensity value is then combined with the corresponding propagation direction to form a vector data containing intensity magnitude and propagation direction for each location. Finally, the vector data from all locations are integrated to obtain the vector synthesis result.

[0094] Step 313: Based on the vector synthesis results, assign defect classification labels to the abnormal candidate regions.

[0095] Optionally, step 313, based on the vector synthesis result, assigns a defect classification label to the anomaly candidate region, specifically including the following steps:

[0096] Step 321: Set the proportional threshold based on the anisotropy coefficient and stress concentration factor in the bogie material fatigue property database.

[0097] In this step, the anisotropy coefficient refers to the parameter characterizing the degree of difference in the mechanical and physical properties of the bogie material in different directions. This coefficient includes the difference coefficients of the material's electrical conductivity, magnetic permeability, strength, and other properties in the longitudinal, transverse, and thickness directions.

[0098] The stress concentration factor is a parameter that characterizes the degree of stress concentration at abrupt changes in the surface or internal structure of a bogie. This factor includes the ratio of the maximum stress to the average stress at the abrupt change in structure, and is used to quantify the impact of stress concentration on the formation and propagation of defects.

[0099] The proportional threshold refers to the judgment threshold calculated based on the anisotropy coefficient and stress concentration factor, which is used to set the intensity proportional standard for defect classification.

[0100] In this embodiment, the anisotropy coefficient of the bogie material is first retrieved from the material fatigue property database; then, based on the structural design data of the bogie, the parts where stress concentration may exist are analyzed, and the stress concentration factor of each part is calculated; then, the anisotropy coefficient and the stress concentration factor are multiplied to obtain the comprehensive influence coefficient; finally, based on the distribution range of the coefficient, a proportional threshold applicable to defect classification is set.

[0101] Step 322: Based on the surface curvature parameters and motion speed of the bogie, calculate the cumulative energy value in the direction of the defect depth in the vector synthesis result to generate a cumulative lower limit value.

[0102] In this step, the surface curvature parameter refers to the data related to the degree of curvature at various locations on the bogie surface. This parameter includes the radius of curvature, magnitude of curvature, and direction of curvature at each point on the surface.

[0103] The cumulative energy value along the defect depth direction refers to the superposition value of signal intensity along the defect depth direction in the vector synthesis result. This cumulative energy value includes the value obtained by summing the intensity values ​​of the same depth layer after decomposing the vector intensity at each location along the defect depth direction.

[0104] The cumulative lower limit refers to the minimum energy accumulation standard required to determine a valid defect, calculated based on surface curvature parameters and motion velocity.

[0105] In this embodiment, surface curvature parameters and bogie movement speed are first retrieved from the bogie geometry parameter database. Then, a curvature influence coefficient is calculated based on the surface curvature parameters. This coefficient is obtained by calculating the ratio of the actual curvature value in the surface curvature parameters to a preset standard curvature value, where the preset standard curvature value refers to the curvature reference value of the defect-free flat area of ​​the bogie. A time weight is calculated based on the movement speed, obtained by calculating the ratio of a preset standard speed to the actual movement speed of the bogie, where the preset standard speed refers to the reference operating speed under normal bogie inspection conditions.

[0106] Next, the depth-direction intensity value at each location is multiplied by the corresponding curvature influence coefficient and time weight to obtain the product result for all locations. Then, the layers are divided according to the depth direction, and the product result of all locations at the same depth layer is set with a cumulative lower limit value based on the energy requirement corresponding to the minimum detectable defect in the bogie material.

[0107] Step 323: Based on the ratio threshold and the cumulative lower limit, determine the vector synthesis result and assign a defect classification label to the abnormal candidate region.

[0108] In this embodiment, the vector data at each position in the vector synthesis result is first traversed to extract the intensity and direction information of each vector; then the intensity of each vector is compared with the ratio threshold to filter out vector data whose intensity exceeds the ratio threshold; then the cumulative energy value in the defect depth direction corresponding to these filtered vector data is checked to determine whether it has reached the lower limit of the cumulative value.

[0109] Specifically, for locations that simultaneously meet the conditions of having an intensity exceeding the proportional threshold and an energy accumulation value reaching the lower limit of the accumulation value, the defect type corresponding to the location is determined based on the defect expansion trend corresponding to its vector direction and in conjunction with the preset defect classification rules, such as cracks, corrosion, and dents. Finally, adjacent locations with the same defect type are integrated, and a unique defect classification identifier is assigned to each continuous defect region, thus completing the assignment of defect classification identifiers for abnormal candidate regions.

[0110] This application embodiment obtains reliable phase data through multi-band acquisition and noise reduction correction, and obtains accurate offset difference data through material, oxide layer and geometric parameter compensation. Then, it generates depth amplitude attenuation curves by combining material fatigue characteristics and historical loads. Vector synthesis results are obtained through intensity conversion and feature fusion. Finally, defect classification is achieved based on proportional threshold and cumulative lower limit. It fully considers the influencing factors such as material properties, operating history and surface structure, makes up for the limitations of traditional single data classification technology, improves the accuracy and pertinence of defect classification, and can solve the problem of misclassification caused by surface coverings, material differences and stress concentration.

[0111] This application provides a specific embodiment. Step 104 involves analyzing the spatial coordinates of the temperature gradient distribution data and the defect classification identifier to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, the target defect region is extracted from the anomaly candidate region. This specifically includes the following steps:

[0112] Step 401: Calculate the spatiotemporal deviation values ​​of temperature gradient distribution data and phase offset data based on the bogie's motion speed and the sampling frequency of the line scanning device.

[0113] In this step, the spatiotemporal deviation value refers to the difference between the temperature gradient distribution data calculated based on the bogie movement speed and the sampling frequency of the line scanning device and the phase offset data in terms of acquisition time and spatial location.

[0114] In this embodiment, the bogie's motion speed and the sampling frequency of the line scanning device are first obtained, and the time interval between adjacent sampling points of the line scanning device is calculated. Then, the motion speed is multiplied by the time interval to obtain the spatial distance between adjacent sampling points. Subsequently, the acquisition timestamp and spatial coordinates of the motion speed and sampling frequency are extracted, and the acquisition time difference at the same spatial location and the spatial coordinate difference at the same acquisition time are calculated. The acquisition time difference and spatial coordinate difference are integrated according to a preset ratio to generate a spatiotemporal deviation value.

[0115] Step 402: Using the spatiotemporal deviation value, interpolate and compensate the timestamps of the temperature gradient distribution data to generate a target dataset. Perform curvature compensation on the spatial coordinates of the defect classification identifier to generate a target identifier set. The target dataset includes the temperature gradient amplitude.

[0116] In this step, the target dataset refers to the dataset obtained after interpolating and compensating the timestamps of the temperature gradient distribution data. The target dataset includes the temperature gradient magnitude and the corrected timestamps.

[0117] The target identifier set refers to the set of identifiers obtained after curvature compensation of the spatial coordinates of the defect classification identifiers. The target identifier set includes the corrected spatial coordinates and the corresponding defect classification identifiers.

[0118] Temperature gradient magnitude refers to the quantized value obtained by dividing the temperature difference between adjacent spatial points by the distance.

[0119] In this embodiment, based on the time deviation in the spatiotemporal deviation value, linear interpolation is used to fill the missing data corresponding to discontinuous timestamps in the temperature gradient distribution data, so that each time point has corresponding temperature gradient data, and the data is integrated to generate a target dataset containing the temperature gradient amplitude. At the same time, combined with the bogie surface curvature parameters, the deviation of the original spatial coordinates of the defect classification identifier caused by surface curvature is corrected to generate a target identifier set so that the identifier position is consistent with the actual defect position.

[0120] Step 403: Perform coupled analysis on the target dataset and the target identifier set to calculate the rate of change of temperature gradient magnitude and the divergence of phase offset.

[0121] In this step, the divergence of the phase offset refers to the degree of dispersion of the phase offset data associated with the target identifier set at different spatial locations within the anomaly candidate region. This divergence is obtained based on the discreteness of the data at each location and is used to reflect the spatial uniformity of the defect's electromagnetic response.

[0122] In this embodiment of the application, firstly, the spatial coordinates corresponding to each defect classification identifier are extracted from the target identifier set. Based on these spatial coordinates, the corresponding phase offset data is matched. Then, the mean of all phase offset data is calculated, and the deviation value between each phase offset data and the mean is calculated one by one. Each deviation value is squared, and all squared deviations are accumulated to obtain the deviation sum of squares. The deviation sum of squares is divided by the data volume of the phase offset data to obtain a quantitative value reflecting the degree of data dispersion. This quantitative value is the divergence of the phase offset.

[0123] Next, the temperature gradient amplitude is extracted from the target dataset, and the amplitude difference between adjacent time points or spatial locations is calculated. Then, the amplitude difference is divided by the corresponding time interval or spatial distance to obtain the rate of change of the temperature gradient amplitude.

[0124] Step 404: Generate reference parameters based on the rate of change of the temperature gradient amplitude and the divergence of the phase offset.

[0125] In this embodiment, the weighting coefficients of the rate of change of temperature gradient amplitude and the divergence of phase offset are first set based on the importance of the two types of data to defect detection. Then, the rate of change of temperature gradient amplitude and the divergence of phase offset are weighted and summed with the corresponding weighting coefficients to obtain the comprehensive quantitative index of each spatial location. Finally, the comprehensive quantitative indexes of all locations are integrated to generate the benchmark parameters.

[0126] Step 405: Based on the spatial distribution weight of the benchmark parameters, combined with the temperature gradient amplitude, offset difference data and the defect classification identifier, extract the target defect region from the abnormal candidate region.

[0127] Optionally, step 405, based on the spatial distribution weights of the reference parameters, combined with the temperature gradient amplitude, offset difference data, and the defect classification identifier, extracts the target defect region from the anomaly candidate region, specifically including the following steps:

[0128] Step 411: Calculate the mesh resolution parameters of the anomaly candidate region based on the spatial distribution weights of the reference parameters and the surface curvature parameters of the bogie.

[0129] In this step, the mesh resolution parameter refers to the mesh density standard calculated based on the spatial distribution weight of the baseline parameter and the bogie surface curvature parameter. This parameter includes the mesh edge length or area.

[0130] In this embodiment, the spatial distribution weights of the reference parameters are added to the surface curvature parameters of the bogie, and the average value is taken. Then, based on the rule that regions with higher weight values ​​have higher grid resolution and regions with greater curvature have higher grid resolution, a negative correlation mapping relationship is established between the average value and the grid side length. This average value is then converted into the corresponding grid side length to obtain the grid resolution parameters. It should be noted that this embodiment does not limit the magnitude of the converted values; the specific values ​​can be set according to actual conditions.

[0131] Step 412: Based on the grid resolution parameters, divide the anomaly candidate region into grids to generate multiple grid regions.

[0132] In this step, the grid region refers to multiple small regions obtained after dividing the anomaly candidate region based on the grid resolution parameter. Each grid region contains a preset spatial location and corresponding detection data.

[0133] In this embodiment of the application, the grid side length is determined according to the grid resolution parameter, and the entire anomaly candidate region is divided into multiple small regions of the same size or dynamically adjusted according to the resolution, with the boundary of the anomaly candidate region as the range. Each small region is a grid region.

[0134] Step 413: Calculate the feature similarity of adjacent grid regions based on the temperature gradient amplitude, offset difference data and defect classification identifier corresponding to each grid region, and merge adjacent grid regions whose feature similarity exceeds a preset merging threshold into an initial defect region.

[0135] In this step, feature similarity refers to the quantified value of the degree of feature matching between adjacent grid regions, calculated based on the temperature gradient magnitude, offset difference data, and defect classification identifier of each grid region.

[0136] The preset merging threshold refers to the preset feature similarity standard for determining whether adjacent grid regions should be merged.

[0137] The initial defect region refers to the continuous region obtained by merging adjacent grid regions whose feature similarity exceeds a preset merging threshold.

[0138] In this embodiment, for each grid region, the temperature gradient amplitude of all locations within the grid region is first extracted and the mean is calculated. The offset difference data of all locations within the grid region is extracted and the mean is calculated. The target defect classification identifier with the highest proportion within the grid region is statistically analyzed. Then, the first difference of the mean temperature gradient amplitude between adjacent grid regions, the second difference of the mean offset difference data, and the binary difference value for whether the target defect classification identifiers are consistent are calculated. When the target defect classification identifiers are consistent, the binary difference value is 0. When the target defect classification identifiers are inconsistent, the binary difference value is 1.

[0139] Subsequently, a first value for the average temperature gradient magnitude of all grid regions is determined. The absolute value of the first difference between adjacent grid regions is divided by the first value to obtain the difference value in the dimension of temperature gradient magnitude. This first value is the maximum value minus the minimum value among the average temperature gradient magnitudes of all grid regions. Then, a second value for the average offset difference data of all grid regions is determined. The absolute value of the second difference between adjacent grid regions is divided by the second value to obtain the difference value in the dimension of offset difference data. This second value is the maximum value minus the minimum value among the average offset difference data of all grid regions.

[0140] The binary difference value is directly used as the difference value of the target defect classification identifier dimension. The average of the difference values ​​of the above three dimensions is taken, and the feature similarity is obtained by subtracting the average value from 1. Finally, the feature similarity is compared with the preset merging threshold, and adjacent grid areas with feature similarity exceeding the preset merging threshold are merged to generate the initial defect area.

[0141] Step 414: Based on the crack morphology template in the bogie material fatigue property database, remove the regions in the initial defect region whose centroid offset exceeds the preset deformation threshold to obtain the target defect region.

[0142] In this step, the crack morphology template refers to the standard morphology data of different types of defects stored in the material fatigue property database. The template includes defect shape, size ratio, edge features, etc.

[0143] Centroid offset refers to the straight-line distance between the centroid coordinates of the initial defect region and the centroid coordinates of the crack morphology template.

[0144] The preset deformation threshold refers to the upper limit of the centroid offset for determining whether the initial defect area is a valid defect.

[0145] The target defect area refers to the area obtained after removing the initial defect area whose centroid offset exceeds the preset deformation threshold. It is used to accurately locate the range of the real defect.

[0146] In this embodiment, a crack morphology template corresponding to the initial defect region is retrieved from a material fatigue property database. The centroid coordinates of the initial defect region are obtained by calculating the mean of the horizontal and vertical coordinates of all spatial positions within the initial defect region. Simultaneously, the centroid coordinates of the crack morphology template are acquired, and the straight-line distance between the two types of centroid coordinates is calculated to obtain the centroid offset. Then, the centroid offset is compared with a preset deformation threshold, and initial defect regions with centroid offsets exceeding the threshold are removed. Finally, the initial defect regions that do not exceed the preset deformation threshold are integrated to obtain the target defect region.

[0147] This application embodiment achieves synchronization of thermal and electromagnetic field data through spatiotemporal deviation correction, generates comprehensive benchmark parameters through coupling analysis, refines the mesh division by weight and curvature, and accurately extracts target defect areas through feature similarity merging and morphology template screening. By integrating multi-dimensional data, it makes up for the shortcomings of traditional technologies such as data asynchrony and coarse feature extraction, solves the problem of missed detection and misjudgment in single data detection, comprehensively covers defect features, and improves the accuracy and reliability of bogie flaw detection.

[0148] This application provides a specific embodiment. Step 102 involves extracting temperature gradient distribution data from the temperature field distribution data and generating anomaly candidate regions based on the temperature gradient distribution data. This specifically includes the following steps:

[0149] Step 201: Based on the bogie's movement speed and the scanning parameters of the line scanning device, generate the spacing compensation amount between adjacent scanning lines.

[0150] In this step, the spacing compensation amount refers to the amount calculated based on the bogie movement speed and the scanning parameters of the line scanning device. This spacing compensation amount includes the spacing correction value and the direction compensation information.

[0151] In this embodiment, the scanning parameters of the line scanning device are first obtained, including the scanning frequency, the preset spacing of the scanning lines, and the scanning direction. Then, the time interval between two adjacent scans of the line scanning device is calculated, and the actual scanning line spacing is obtained by multiplying the bogie's movement speed by the time interval. The actual scanning line spacing is compared with the preset spacing, and the difference between the two is calculated. This difference is the spacing compensation amount. If the actual spacing is greater than the preset spacing, the compensation amount is positive; otherwise, it is negative.

[0152] Step 202: Extract the slope of temperature change along the bogie's movement direction within each scan line from the temperature field distribution data, and combine it with the spacing compensation amount to generate temperature gradient distribution data.

[0153] In this step, the temperature change slope refers to the rate at which the temperature changes with spatial position along the direction of bogie movement within each scan line.

[0154] In this embodiment, firstly, all temperature data points corresponding to each scan line are extracted from the temperature field distribution data and sorted according to the bogie movement direction; then, the temperature difference between every two adjacent temperature data points is calculated, and the difference is divided by the actual spatial distance corresponding to the adjacent data points to obtain the temperature change rate of each adjacent interval, and the temperature change rates of all adjacent intervals within each scan line are integrated to form the temperature change slope of the scan line.

[0155] Next, the temperature change slope of each scan line is multiplied by the correction coefficient corresponding to the spacing compensation amount to obtain the corrected temperature change slope, in order to adjust the slope error caused by the spacing deviation. When the spacing compensation amount is positive, the correction coefficient is greater than 1, and when it is negative, the correction coefficient is between 0 and 1. The corrected temperature change slopes of all scan lines are integrated to generate temperature gradient distribution data.

[0156] Step 203: Calculate the absolute value of the temperature difference between adjacent scan lines in the temperature gradient distribution data, and perform noise suppression processing on the absolute value of the temperature difference according to the thermal diffusivity of the bogie and the thickness of the surface oxide layer to generate target absolute value of temperature difference data.

[0157] In this step, the target temperature difference absolute value data refers to the data obtained after noise suppression processing of the absolute value of the temperature difference between adjacent scan lines.

[0158] In this embodiment of the application, firstly, the temperature gradient values ​​corresponding to the spatial positions of two adjacent scan lines are extracted from the temperature gradient distribution data, and the difference between the temperature gradient values ​​at the same spatial position is calculated. The absolute value of the difference is then taken to obtain the absolute value of the temperature difference.

[0159] Next, the thermal diffusivity is retrieved from the bogie parameter database, and the surface oxide layer thickness is obtained through eddy current detection to determine the degree of interference of the oxide layer on the temperature signal. Then, based on the thermal diffusivity and surface oxide layer thickness obtained above, the noise suppression weight is calculated. The noise suppression weight is obtained by dividing the thermal diffusivity by the surface oxide layer thickness. The larger the thermal diffusivity and the thinner the oxide layer, the greater the weight, thereby quantifying the noise suppression intensity under different conditions.

[0160] Then, each absolute temperature difference is multiplied by the corresponding noise suppression weight to obtain the product result of a single location. This product result is then summed with the product results of a preset number of adjacent locations and averaged to obtain the target absolute temperature difference of a single location, thus eliminating random noise interference. Finally, the target absolute temperature difference values ​​of all spatial locations are integrated according to their corresponding spatial coordinates to generate target absolute temperature difference data.

[0161] Step 204: Select the target adjacent scan line whose absolute temperature difference exceeds the preset temperature difference value from the target absolute temperature difference data.

[0162] In this step, the preset temperature difference value refers to the judgment criterion set based on the thermal conductivity of the bogie material and the maximum temperature rise rate in the historical load spectrum.

[0163] In this embodiment, firstly, the thermal conductivity of the bogie material is retrieved from the material parameter database, and the maximum temperature rise rate under different load conditions is analyzed based on the historical load data of the bogie. Then, the product of the thermal conductivity and the maximum temperature rise rate is calculated according to a preset ratio to obtain a preset temperature difference value. The target temperature difference absolute value data is traversed, and each temperature difference absolute value is compared with the preset temperature difference value. Adjacent scan lines whose absolute temperature difference exceeds the preset temperature difference value are selected as target adjacent scan lines.

[0164] Step 205: Perform region extension and closing operations on the adjacent scan lines of the target along the movement direction of the bogie to generate abnormal candidate regions.

[0165] In this embodiment, the target spatial location range corresponding to the adjacent scan lines of the target is first determined. Then, the target spatial location range is extended to both sides by a preset distance along the bogie movement direction to cover the associated area that may have defects, thus obtaining the extended region. Next, the extended region is subjected to a closing operation. Specifically, the tiny gaps and holes in the extended region are first filled, and the broken parts in the extended region caused by local noise are connected to obtain the closed region. Finally, all closed regions are integrated and the overlapping parts are removed to form a clear, continuous and complete abnormal candidate region.

[0166] The embodiments of this application can eliminate scanning deviation, noise and interference from surface coverings, accurately locate suspicious areas, and solve the problems of traditional technologies being sensitive to surface interference and having unclear area locking.

[0167] Figure 2 This application provides a schematic diagram of a line-scan-based surface flaw detection system for parts, as shown in the embodiment. Figure 2 As shown, the system includes:

[0168] The acquisition module 21 is used to scan the surface of the bogie using a line scanning device and acquire a scan line dataset of the bogie surface, and extract temperature field distribution data from the scan line dataset.

[0169] The generation module 22 is used to extract temperature gradient distribution data from the temperature field distribution data and generate anomaly candidate regions based on the temperature gradient distribution data.

[0170] The acquisition module 23 is used to acquire the phase offset data of the eddy current signal in the abnormal candidate region, and assign a defect classification identifier to the abnormal candidate region based on the phase offset data.

[0171] The detection module 24 is used to analyze the temperature gradient distribution data and the spatial coordinates of the defect classification identifier to generate benchmark parameters. Based on the spatial distribution weight of the benchmark parameters, the target defect area is extracted from the abnormal candidate area to realize the flaw detection of the bogie surface.

[0172] Figure 2 The aforementioned line-scan based part surface flaw detection system can perform... Figure 1 The implementation principle and technical effects of the line-scan-based surface flaw detection method for parts described in the illustrated embodiment will not be repeated here. The specific operation methods of each module and unit in the line-scan-based surface flaw detection system for parts described in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0173] In one possible design, Figure 2 The line-scan-based part surface flaw detection system of the embodiment shown can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0174] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0175] The processing component 32 is used to: scan the bogie surface using a line scanning device and collect a scan line dataset of the bogie surface; extract temperature field distribution data from the scan line dataset; extract temperature gradient distribution data from the temperature field distribution data; generate anomaly candidate regions based on the temperature gradient distribution data; obtain phase offset data of eddy current signals within the anomaly candidate regions; assign defect classification labels to the anomaly candidate regions based on the phase offset data; analyze the spatial coordinates of the temperature gradient distribution data and the defect classification labels to generate benchmark parameters; and extract target defect regions from the anomaly candidate regions according to the spatial distribution weights of the benchmark parameters, thereby achieving flaw detection on the bogie surface.

[0176] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as one or more application-specific integrated circuits, digital signal processors, digital signal processing devices, programmable logic devices, field-programmable gate arrays, controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described method.

[0177] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory, electrically erasable programmable read-only memory, erasable programmable read-only memory, programmable read-only memory, read-only memory, magnetic storage, flash memory, magnetic disk, or optical disk.

[0178] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0179] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0180] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0181] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0182] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a surface flaw detection method for parts based on line scanning.

[0183] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0184] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0185] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application 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 do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for surface flaw detection of parts based on line scanning, characterized in that, include: The surface of the bogie is scanned using a line scanning device, and a scan line dataset of the bogie surface is collected. Temperature field distribution data is then extracted from the scan line dataset. Temperature gradient distribution data is extracted from the temperature field distribution data, and anomaly candidate regions are generated based on the temperature gradient distribution data. The phase offset data of the eddy current signal within the abnormal candidate region is obtained, and a defect classification identifier is assigned to the abnormal candidate region based on the phase offset data. The spatial coordinates of the temperature gradient distribution data and the defect classification identifier are analyzed to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, the target defect region is extracted from the anomaly candidate region to achieve flaw detection on the bogie surface. Acquire phase offset data of eddy current signals within the anomaly candidate region, and assign defect classification labels to the anomaly candidate region based on the phase offset data, including: The raw phase data of eddy current signals in different frequency bands within the anomaly candidate region are collected, and noise suppression and baseline correction are performed on the raw phase data to generate phase offset data. The phase offset data is compensated based on the bogie's material properties, surface oxide layer thickness, and geometric feature parameters of the abnormal candidate region to obtain offset difference data. Based on the offset difference data, a defect classification identifier is assigned to the abnormal candidate region; Based on the offset difference data, defect classification labels are assigned to the abnormal candidate regions, including: Based on the bogie's material fatigue property database and historical load data, a depth amplitude decay curve is generated. Based on the depth amplitude attenuation curve, the offset difference data is subjected to intensity conversion and feature fusion processing to obtain the vector synthesis result; Based on the vector synthesis results, defect classification identifiers are assigned to the anomaly candidate regions; Based on the vector synthesis results, defect classification labels are assigned to the anomaly candidate regions, including: Based on the anisotropy coefficient and stress concentration factor in the bogie material fatigue property database, a proportional threshold is set; Based on the surface curvature parameters and motion speed of the bogie, the cumulative energy value in the direction of the defect depth in the vector synthesis result is calculated to generate a cumulative lower limit value; Based on the aforementioned proportional threshold and cumulative lower limit, the vector synthesis result is judged to assign a defect classification identifier to the abnormal candidate region.

2. The method for surface flaw detection of parts based on line scanning according to claim 1, characterized in that, The spatial coordinates of the temperature gradient distribution data and the defect classification identifier are analyzed to generate benchmark parameters. Based on the spatial distribution weights of the benchmark parameters, the target defect region is extracted from the anomaly candidate region, including: Based on the bogie's motion speed and the sampling frequency of the line scanning device, the spatiotemporal deviation values ​​of the temperature gradient distribution data and phase offset data are calculated. Using the spatiotemporal deviation value, the timestamps of the temperature gradient distribution data are interpolated and compensated to generate a target dataset. The spatial coordinates of the defect classification identifiers are curvature compensated to generate a target identifier set. The target dataset includes the temperature gradient amplitude. The target dataset and the target identifier set are coupled and analyzed to calculate the rate of change of temperature gradient magnitude and the divergence of phase offset; Based on the rate of change of the temperature gradient amplitude and the divergence of the phase offset, reference parameters are generated; Based on the spatial distribution weights of the benchmark parameters, combined with the temperature gradient amplitude, offset difference data, and defect classification identifier, the target defect region is extracted from the anomaly candidate region.

3. The method for surface flaw detection of parts based on line scanning according to claim 2, characterized in that, Based on the spatial distribution weights of the benchmark parameters, combined with the temperature gradient amplitude, offset difference data, and defect classification identifier, the target defect region is extracted from the anomaly candidate region, including: Based on the spatial distribution weights of the reference parameters and the surface curvature parameters of the bogie, the mesh resolution parameters of the anomaly candidate region are calculated. Based on the grid resolution parameters, the anomaly candidate region is divided into grids to generate multiple grid regions; Based on the temperature gradient amplitude, offset difference data and defect classification identifier of each grid region, the feature similarity of adjacent grid regions is calculated, and adjacent grid regions whose feature similarity exceeds a preset merging threshold are merged into an initial defect region. Based on the crack morphology template in the bogie material fatigue property database, regions in the initial defect region whose centroid offset exceeds a preset deformation threshold are removed to obtain the target defect region.

4. The method for surface flaw detection of parts based on line scanning according to claim 1, characterized in that, Temperature gradient distribution data is extracted from the temperature field distribution data, and anomaly candidate regions are generated based on the temperature gradient distribution data, including: Based on the bogie's motion speed and the scanning parameters of the line scanning device, the spacing compensation amount between adjacent scanning lines is generated. The slope of temperature change along the bogie's direction of motion within each scan line is extracted from the temperature field distribution data, and combined with the spacing compensation amount, temperature gradient distribution data is generated. The absolute value of the temperature difference between adjacent scan lines in the temperature gradient distribution data is calculated, and noise suppression processing is performed on the absolute value of the temperature difference based on the thermal diffusivity of the bogie and the thickness of the surface oxide layer to generate target absolute value of temperature difference data. Select the target adjacent scan lines whose absolute temperature difference exceeds a preset temperature difference value from the target absolute temperature difference data; The adjacent scan lines of the target are subjected to region extension and closing operations along the movement direction of the bogie to generate abnormal candidate regions.

5. A line-scan-based surface flaw detection system for parts, used in the line-scan-based surface flaw detection method for parts as described in any one of claims 1 to 4, characterized in that, include: The acquisition module is used to scan the surface of the bogie using a line scanning device and acquire a scan line dataset of the bogie surface, and extract temperature field distribution data from the scan line dataset. The generation module is used to extract temperature gradient distribution data from the temperature field distribution data and generate anomaly candidate regions based on the temperature gradient distribution data. The acquisition module is used to acquire phase offset data of eddy current signals within the abnormal candidate region, and assign defect classification identifiers to the abnormal candidate region based on the phase offset data. The detection module is used to analyze the temperature gradient distribution data and the spatial coordinates of the defect classification identifier to generate benchmark parameters. Based on the spatial distribution weight of the benchmark parameters, the target defect area is extracted from the abnormal candidate area to realize the flaw detection of the bogie surface.

6. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement a line scan-based part surface flaw detection method as described in any one of claims 1 to 4.

7. A computer storage medium, characterized in that, The device contains a computer program that, when executed by a computer, implements a line-scan-based surface flaw detection method for parts as described in any one of claims 1 to 4.

Images