A method and system for visual inspection of quality defects in a racing boat hull curing process

Abstract

The present application relates to a kind of racing boat hull curing process quality defect visual inspection method and system.The method comprises: the first thermal excitation is applied, the first thermal excitation has preset energy and duration time;After the first thermal excitation is applied, the first temperature change data of the hull surface in a period of time is obtained;When the thermal effect generated by the first thermal excitation has not completely dissipated, the second thermal excitation is applied to the hull surface, and the second temperature change data of the hull surface in a period of time is obtained, and the energy or duration time of the second thermal excitation is different from the first thermal excitation;First temperature change data and second temperature change data are compared and processed to generate signal reflecting the difference between the two;According to difference signal, the structural anomaly in the hull material is identified.The present application overcomes the problem that the existing pulse thermal imaging system appears in the detection of coating functional coating hull.

Description

Technical Field

[0001] This application relates to the field of nondestructive testing technology, and in particular to a visual inspection method and system for quality defects in the curing process of racing boat hulls. Background Technology

[0002] In the manufacturing process of high-performance racing boat hulls, the composite material hull, after being precisely laid with multiple layers of fibers and impregnated with high-performance resin, is placed in a curing oven or autoclave for a long period of curing under a preset temperature and pressure profile. This curing process determines the final mechanical properties, durability, and fatigue resistance of the hull. To assess the quality of the hull after curing, it undergoes a quality inspection process after being removed from the mold.

[0003] To enhance the hydrodynamic performance, UV resistance, and resistance to marine organism adhesion of racing boats, the cured hull surface is typically coated with multiple layers of high-performance functional coatings, including a primer, topcoat, drag-reducing coating, and antifouling coating. Traditional quality inspection primarily relies on visible light imaging systems, which can identify visible problems such as surface bubbles, fiber wrinkles, scratches, or dents. However, visible light imaging cannot penetrate to detect hidden defects such as micro-delamination between composite material layers, internal voids, localized soft spots caused by incomplete resin curing, and poor bonding between fibers and resin. To address these hidden defects, pulsed thermal imaging technology has been introduced. This technology uses a high-intensity, short-pulse flash lamp to instantaneously heat the hull surface, and a high-sensitivity infrared thermal imager continuously captures the surface temperature decay curve. Internal structural problems are then inferred from the abnormal heat diffusion caused by internal material defects.

[0004] However, when pulsed thermal imaging systems inspect hulls coated with functional coatings, the thermophysical properties of the coating itself (such as thermal conductivity, specific heat capacity, and surface emissivity) differ from those of the underlying composite matrix. Furthermore, due to the complex and continuously varying curved geometry of the hull, coupled with limitations in the spraying process, it is difficult to maintain uniform coating thickness across different areas of the hull, with variations ranging from tens to hundreds of micrometers. These differences in coating thickness and thermophysical property inhomogeneity lead to localized variations in heat absorption, conduction, and radiation processes, producing characteristics on the surface temperature decay curve similar to internal defects. The thermal response in thicker areas may simulate the decay mode of tiny voids, causing false alarms; conversely, rapid heat dissipation in thinner areas may mask underlying defects, leading to omissions.

[0005] This confusion between genuine and false signals significantly reduces the reliability of pulsed thermal imaging systems in real-world production environments. How to distinguish between thermal artifacts caused by coating effects and temperature anomalies caused by actual defects within the composite material on racing boat hulls coated with multiple layers of functional coatings of uneven thickness, in order to improve the accuracy and reliability of detection, is a pressing technical problem in this field. Summary of the Invention

[0006] This application provides a visual inspection method and system for quality defects in the curing process of racing boat hulls, which at least solves the problem that pulsed thermal imaging systems cannot distinguish between thermal artifacts caused by coating effects and abnormal temperature signals caused by real defects inside the composite material on racing boat hulls with multiple layers of functional coatings and uneven coating thickness.

[0007] In a first aspect, this application provides a visual inspection method for quality defects in the curing process of a racing boat hull, comprising the following steps: A first thermal excitation is applied, the first thermal excitation having a preset energy and duration; After the first thermal excitation is applied, the first temperature change data of the hull surface over a period of time is obtained; Before the thermal effect generated by the first thermal excitation has completely dissipated, a second thermal excitation is applied to the surface of the hull, and second temperature change data of the hull surface over a period of time is obtained. The energy or duration of the second thermal excitation differs from that of the first thermal excitation. The first temperature change data and the second temperature change data are compared and processed to generate a signal that reflects the difference between the two. Based on the difference signals, structural anomalies within the hull material are identified.

[0008] Secondly, this application provides a visual inspection system for quality defects in the curing process of a racing boat hull, the system comprising: A first thermal excitation application module is used to apply a first thermal excitation, wherein the first thermal excitation has a preset energy and duration; The first temperature data acquisition module is used to acquire the first temperature change data of the hull surface over a period of time after the first thermal excitation is applied. The secondary excitation acquisition module is used to apply a second thermal excitation to the surface of the hull before the thermal effect generated by the first thermal excitation has completely dissipated, and to acquire second temperature change data of the hull surface over a period of time. The energy or duration of the second thermal excitation differs from that of the first thermal excitation. The data comparison and processing module is used to compare and process the first temperature change data and the second temperature change data to generate a signal reflecting the difference between the two. The structural anomaly identification module is used to identify structural anomalies inside the hull material based on the difference signals.

[0009] Compared with related technologies, the visual inspection method and system for quality defects in the curing process of racing boat hulls provided in this application have at least the following technical advantages: By applying two thermal excitations with differences in energy or duration, and applying the second thermal excitation before the thermal effect of the first thermal excitation has completely dissipated, the unique sensitivity of internal material defects to continuous thermal response and the relative consistency of surface coating inhomogeneity to the response of the two thermal excitations are utilized. By comparing the differences in the two temperature change data, thermal artifacts caused by uneven coating thickness or differences in thermophysical properties can be effectively distinguished from the actual structural anomalies inside the composite material. This overcomes the problem of confusion between true and false signals in the detection of functionally coated ship hulls by existing pulsed thermal imaging systems.

[0010] Details of one or more embodiments of this application are set forth in the following drawings and description to make other features, objects and advantages of this application more readily apparent. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a visual inspection method for quality defects in the curing process of a racing boat hull, according to an exemplary embodiment. Figure 2 This is a flowchart illustrating step S41 according to an exemplary embodiment; Figure 3 This is a flowchart illustrating step S411 according to an exemplary embodiment; Figure 4 This is a flowchart illustrating step S51 according to an exemplary embodiment; Figure 5 This is a flowchart illustrating step S551 according to an exemplary embodiment; Figure 6 This is a flowchart illustrating step S5321 according to an exemplary embodiment; Figure 7 This is a block diagram illustrating a visual inspection system for quality defects in the curing process of a racing boat hull, according to another exemplary embodiment. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the application is described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments provided in this application without inventive effort are within the scope of protection of this application.

[0013] Obviously, the accompanying drawings described below are merely some examples or embodiments of this application. Those skilled in the art can apply this application to other similar scenarios based on these drawings without any inventive effort. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, any changes to design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as insufficient disclosure of the content of this application.

[0014] In this application, the reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment that is mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0015] Unless otherwise defined, the technical or scientific terms used in this application shall have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms “a,” “an,” “an,” “the,” and similar words used in this application do not indicate quantity limitation and may indicate singular or plural. The terms “comprising,” “including,” “having,” and any variations thereof used in this application are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units not listed, or may include other steps or units inherent to these processes, methods, products, or devices. The terms “connected,” “linked,” “coupled,” and similar words used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. “Multiple” used in this application refers to two or more. “And / or” describes the relationship between related objects, indicating that three relationships may exist; for example, “A and / or B” can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following objects are in an "or" relationship. The terms "first," "second," and "third" used in this application are merely to distinguish similar objects and do not represent a specific ordering of the objects.

[0016] Example 1 Embodiment 1 of this application provides a visual inspection method for quality defects in the curing process of a racing boat hull. Figure 1 This is a flowchart illustrating a visual inspection method for quality defects in the curing process of a racing boat hull, according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps: S1, apply a first thermal excitation, the first thermal excitation having a preset energy and duration; In this step, the first thermal excitation is emitted by a high-intensity, short-pulse flash lamp array, which instantaneously and uniformly heats the area to be tested on the surface of the racing boat hull. The pulse duration is determined based on the type and thickness of the hull composite material, typically ranging from several milliseconds to tens of milliseconds, with a heat flux density sufficient to produce a measurable temperature change on the surface without damaging the coating. The purpose of the first thermal excitation is to establish an initial temperature field on the hull surface, stimulating heat conduction processes within the material, which diffuse inward at a rate determined by the material's thermophysical properties.

[0017] S2, after applying the first thermal excitation, acquire the first temperature change data of the hull surface over a period of time; In this step, a high-frame-rate, high-resolution infrared thermal imager is used to continuously acquire a sequence of temperature decay images of the hull surface over a period of time after the first thermal excitation is applied. The thermal imager frame rate is no less than 100Hz, and the temperature resolution is better than 0.02°C to capture subtle temperature changes during the rapid decay process. The acquisition duration is determined based on the typical thermal diffusion time constant of the hull material to ensure a complete record of the temperature decay from its peak to near ambient levels. The acquired first temperature change data reflects the baseline thermal response behavior of the defect-free area under the current coating and environmental conditions, while also including the relaxation characteristics after the coating property transformation.

[0018] S3, while the thermal effect generated by the first thermal excitation has not completely dissipated, a second thermal excitation is applied to the surface of the hull, and second temperature change data of the hull surface over a period of time is obtained, wherein the energy or duration of the second thermal excitation differs from that of the first thermal excitation. In this step, after a delay before the thermal effect of the first thermal excitation has completely dissipated, a second thermal excitation is applied to the same region. The energy of the second thermal excitation... Set as the first thermal excitation energy 50% to 80%, that is Where k1 is the energy proportionality coefficient, ranging from 0.5 to 0.8, preferably approximately 70%. The purpose of applying the second thermal excitation at this time point is that the leading edge of the heat wave generated by the first thermal excitation has reached the typical depth of the internal defects in the hull material, and the heat wave from the second thermal excitation will interfere with the residual thermal field of the first thermal excitation at the defect. The infrared thermal imager synchronously acquires the second set of temperature change image sequences at the same frame rate and resolution as the first set.

[0019] S4, compare and process the first temperature change data and the second temperature change data to generate a signal that reflects the difference between the two; In this step, the image processing unit in the industrial computer first performs precise spatial registration on the two sets of temperature change image sequences acquired by the first and second thermal excitations, ensuring that each pixel corresponds to the same physical location on the hull surface in both acquisitions. For each registered pixel, its temperature decay curve in the first set of thermal response data is extracted. In the second group The core mechanism of the comparative processing lies in the fact that the differences in the thermophysical properties and thickness inhomogeneity of the surface functional coating, under two closely connected thermal pulses with slightly different parameters, exhibit a high degree of consistency in their influence on the surface temperature decay curve, belonging to a relatively linear background effect. However, structural anomalies such as delamination or voids actually existing within the material, under two thermal pulses of different energies or durations, show significant nonlinear differences in their influence on surface temperature decay due to altered heat diffusion paths. The system calculates the difference or characteristic parameter difference between the two sets of temperature decay curves at each time point to generate a differential thermal response signal. The amplitude and time-varying characteristics of this differential signal are related to the type, size, and depth of the anomaly.

[0020] S5. Based on the difference signal, identify the structural anomaly inside the hull material.

[0021] In this step, spatially, the system performs statistical analysis on the generated differential thermal response signals. Based on the statistical distribution of the differential signals in defect-free areas (e.g., the average value plus three standard deviations), a dynamic discrimination threshold is calculated. Any pixel whose differential signal value exceeds this threshold is marked as a candidate anomalous pixel. Connectivity analysis is performed on the candidate anomalous pixels, aggregating spatially adjacent pixels exceeding the threshold into candidate defect regions. Temporally, the system analyzes the change pattern of the differential signal over time for each candidate defect region: layered defects, due to hindering heat transfer to deeper layers, maintain a high positive value and decay slowly over a longer timescale; void defects show an earlier thermal response peak and decay relatively quickly; resin-rich areas exhibit a characteristic temperature hysteresis pattern caused by differences in heat conduction capacity. Combining the amplitude threshold and temporal characteristic patterns of the differential signals, signal regions deviating from the normal fluctuation range and conforming to the known temporal characteristics of defects are identified as internal structural anomalies. The identification results are mapped to the coordinate system of the ship's 3D CAD model through pre-calibrated geometric transformation relationships, outputting the spatial location of the anomaly, the coordinates of the projected area on the ship's surface, and a preliminary estimate of the anomaly type and severity.

[0022] In the above embodiment, by applying two thermal excitations with different energy or duration to the surface of the hull, and applying the second thermal excitation before the first thermal effect has completely dissipated, the difference signal is generated by comparing and processing the two sets of temperature change data, thereby identifying structural anomalies inside the hull material. This achieves effective suppression of coating effect interference and reliable detection of internal defects during the curing process of the racing boat hull.

[0023] In one possible design, Figure 2 This is a flowchart illustrating step S4 according to an exemplary embodiment. (Refer to the attached document.) Figure 2 In step S4, the step of comparing the first temperature change data and the second temperature change data includes: S41, after applying the first thermal excitation, continuously monitor the temperature decay process of each local area on the hull surface; In this step, for each pixel on the hull surface or a sub-region pre-divided according to geometric features, the system extracts the decay curve of the surface temperature of each region over time after the first thermal excitation from the first set of thermal response image sequences. A Savitzky-Golay filter was used to smooth and denoise the original temperature curve to suppress high-frequency noise from the infrared thermal imager. Then, first-order derivative analysis was performed on the smoothed curve to calculate the temperature recovery rate at each time point. The recovery rate directly reflects the thermal diffusion capability of the region under the current coating condition and environmental conditions: a faster recovery rate indicates that heat dissipates smoothly into the material interior and surrounding environment, possibly corresponding to areas with thinner coatings or materials with good thermal conductivity; a slower recovery rate indicates heat retention, possibly caused by thicker coatings, structural anomalies within the material that hinder thermal diffusion (such as delamination or voids), or abnormal material heat capacity (such as resin-rich areas). The distribution of recovery rates in each region provides a quantitative basis for subsequently determining the optimal delay time for applying the second thermal excitation and identifying candidate abnormal regions.

[0024] S42, Based on the temperature decay process, analyze the recovery rate and trend of temperature in each local area of ​​the hull surface; In this step, the temperature decay curves of each local region are fitted using an exponential decay model: where is the ambient temperature baseline for that region, is the initial temperature rise, and is the thermal relaxation time constant. The thermal relaxation time constant characterizes the time scale of heat diffusion and dissipation from the surface to the interior, and is related to the material's thermal conductivity, specific heat capacity, and local thickness—the value is usually greater in defective regions than in normal regions due to longer heat diffusion paths or abnormal heat capacity. The system extracts the and decay characteristic parameters of each region, classifying the detection area into three categories—fast response region, normal response region, and slow response region—based on the recovery rate, and establishes a recovery trend map for each region. The recovery rate trend information provides a quantitative decision-making basis for determining the delay time for applying the second thermal excitation in different types of response regions.

[0025] S43, determine the delay time for applying the second thermal excitation based on the recovery rate and trend; In this step, the optimal delay time point is determined based on the recovery rate distribution of the temperature decay curves in each local region. This time point satisfies the following: in, To delay time, Let be the material's thermal relaxation time constant. At this time point, the residual temperature of the first thermal excitation's thermal effect is still higher than the ambient temperature in most regions, the thermal diffusion process within the material is still ongoing, and the differences in thermal response caused by internal defects are at their maximum expression stage. Delay times are determined for different types of response regions to accommodate the differences in thermophysical properties in different regions of the hull structure.

[0026] S44, after the delay time, a second thermal excitation is applied to the surface of the hull, and the second temperature change data is acquired; In this step, a second thermal excitation is triggered after a predetermined delay. The energy of the second thermal excitation is set to 50% to 80% of the energy of the first thermal excitation to ensure that the thermal effect can penetrate to the depth of the defect without causing surface overheating. An infrared thermal imager synchronously acquires a second set of temperature change image sequences at the same frame rate and resolution as the first set, recording the temperature decay process of each pixel on the hull surface for a period of time after the second thermal excitation.

[0027] S45, compare and process the first temperature change data and the second temperature change data to generate a signal that reflects the difference between the two; In this step, the first set of temperature change data for each pixel is... Second set of temperature change data Perform time-point comparisons and calculate the difference between the two sets of temperature decay curves. Sum and ratio : In the formula, the amplitude of the difference signal characterizes the degree of change in thermal response at that location under the first and second thermal stimuli; the greater the difference, the more likely a structural anomaly exists in the region. The generated difference signal is presented in the form of a two-dimensional image, and the signal intensity reflects the probability and severity of the internal structural anomaly.

[0028] In the technical solution of the above embodiment, by continuously monitoring the temperature decay process after the first thermal excitation, analyzing the recovery rate and trend of each local area, determining the optimal delay time for applying the second thermal excitation, and applying the second thermal excitation after the delay time for comparison processing, the difference signal between the two thermal response data is maximized, thereby improving the sensitivity of subsequent defect signal identification.

[0029] In one example, when inspecting a region of a racing boat hull, after applying the first thermal excitation, the infrared thermal imager captured that the surface temperature of that region decayed exponentially. The system analyzed the decay curves of each sub-region and found that sub-region A had a faster temperature recovery rate, while sub-region B had a slower recovery rate. Based on this, the system determined a shorter delay time for sub-region A and a slightly longer delay time for sub-region B, and then applied a second thermal excitation for differential analysis to maximize the difference in thermal response data between the two sets.

[0030] In one possible design, Figure 3 This is a flowchart illustrating step S45 according to an exemplary embodiment. (Refer to the attached diagram.) Figure 3 In step S45, the step of comparing the first temperature change data and the second temperature change data to generate a signal reflecting the difference between the two includes: S451, Analyze the first temperature change data and extract specific features that reflect the relaxation process after the coating properties change; In this step, a thorough analysis is conducted on the normal regions in the initial temperature change data that are unaffected by defects, extracting relaxation features that reflect the transformation of the coating's properties after curing. These specific features include the initial slope of the temperature decay curve, the temperature and time point at which the decay process reaches an inflection point, and the temperature integral value within a specific time window. The initial slope reflects the thermal diffusivity of the material surface, the inflection point is related to the glass transition process of the coating's property transformation, and the temperature integral value comprehensively characterizes the overall heat capacity of the material in the current coating state. These features provide crucial input for the subsequent parameter calibration of the thermal diffusion model.

[0031] S452, Based on the specific characteristics, adjust the parameters of the thermal diffusion model to generate a thermal response baseline curve; In this step, the relaxation features extracted from the first temperature change data are used to calibrate the parameters of the preset finite element thermal diffusion model. The adjusted parameters include the material's thermal conductivity, specific heat capacity, and density, ensuring that the temperature decay curve output by the model best matches the actual measured temperature change data in the normal region. The calibrated model then generates a thermal response baseline curve. This curve represents the expected behavior of the surface temperature decay over time in this defect-free hull region under the current coating condition and environmental conditions. The accuracy of the baseline curve directly affects the accuracy of subsequent differential signals.

[0032] S453, compare the second temperature change data with the thermal response reference curve, calculate the difference, and generate a dynamic compensation differential thermal response signal; In this step, the actual temperature change data obtained after the second thermal excitation will be used. Compared with thermal response reference curve Perform point-by-point comparisons and calculate the dynamic compensation differential thermal response signal using the following formula: In this formula, since the baseline curve already includes the contributions of coating characteristics and environmental factors, the subtraction operation effectively eliminates background noise caused by non-defect factors such as coating effects, ambient temperature fluctuations, and differences in excitation parameters. The generated dynamic compensation differential thermal response signal is highly pure, prominently reflecting the true thermal response deviation caused by abnormalities in the internal structure of the hull material.

[0033] S454, Based on the dynamic compensation differential thermal response signal, identify structural anomalies inside the hull material.

[0034] In this step, threshold analysis and spatial continuity determination are performed on the dynamically compensated differential thermal response signal. A signal amplitude threshold is set, and pixels exceeding the threshold are marked as candidate anomalous pixels. Connectivity analysis is performed on the candidate anomalous pixels, and spatially adjacent pixels exceeding the threshold are aggregated into candidate defect regions. Based on the spatial size, shape characteristics, and signal intensity distribution of the candidate regions, the defect type and size are preliminarily determined. Different types of internal defects exhibit distinguishable response patterns in the compensated signal; for example, layered defects typically exhibit a planar distribution, while void defects appear as point-like high-signal areas.

[0035] In the above embodiment, the thermal response reference curve is generated by dynamically calibrating the thermal diffusion model parameters by extracting the relaxation characteristics after the coating property transformation. The second temperature change data is compared with the reference curve to obtain the dynamically compensated differential thermal response signal, which effectively eliminates the interference of coating effect and environmental factors on defect signal and improves the accuracy of internal defect identification.

[0036] In one example, when inspecting the mid-section of the starboard hull plating, analysis of the first temperature change data extracted the characteristic attenuation constant of the coating relaxation. Based on this, the thermal conductivity parameters of the thermal diffusion model were fine-tuned to match the actual attenuation curve, generating a thermal response baseline curve. Comparing the second temperature change data for this area with the baseline curve revealed an anomaly at approximately 1.5 meters from the bow, with a signal peak duration of approximately 120 milliseconds and a peak amplitude approximately 2.3 times that of the surrounding area. This anomaly was ultimately identified as a microscopic delamination defect located between the third and fourth layers of carbon fiber fabric, measuring approximately 8 mm × 5 mm.

[0037] In one possible design, Figure 4 This is a flowchart illustrating step S454 according to an exemplary embodiment. (Refer to the attached document.) Figure 4 In step S454, the step of extracting time-domain features from the dynamically compensated differential thermal response signal includes: S4541, perform time-domain feature extraction on the dynamic compensation differential thermal response signal to obtain the peak occurrence time, peak amplitude, attenuation rate and signal duration of the signal; In this step, temporal features are extracted from the dynamic compensation differential thermal response signal of each candidate defect region. The peak occurrence time refers to the moment the signal reaches its most significant value, reflecting the time delay of the heat wave propagating from the surface to the defect and returning; it is positively correlated with the defect depth. The peak amplitude is the maximum signal intensity, related to the defect size and depth. The decay rate describes the rate at which the signal decreases from the peak to half-peak, influenced by the thermal diffusion characteristics of the defect region. The signal duration is the time span from the start to recovery to the baseline level, related to the defect's geometry and thermal capacity. These temporal features collectively constitute the temporal dimension criteria for defect identification.

[0038] S4542, Frequency domain feature extraction is performed on the dynamic compensation differential thermal response signal to obtain the energy distribution and main frequency component of the signal; In this step, a Fast Fourier Transform (FFT) is performed on the dynamically compensated differential thermal response signal to convert the time-domain signal to a frequency-domain representation. Parameters such as the peak frequency, peak amplitude, half-power bandwidth, and spectral energy concentration of the dominant frequency component are extracted from the spectrum. The frequency-domain characteristics reflect the modulation effect of defects on the frequency components of the thermal wave; different types and depths of defects have different enhancement or attenuation effects on specific frequency components. For example, deeper defects tend to affect low-frequency components, while shallower defects mainly affect high-frequency components. The frequency-domain characteristics and time-domain characteristics complement each other, together constituting a multi-dimensional input for defect feature matching.

[0039] S4543, compare the time-domain features and the frequency-domain features with a preset defect feature library; In this step, the extracted time-domain feature vector and frequency-domain feature vector are concatenated into a joint feature vector, which is then compared one by one with the defect type templates in the pre-set defect feature library. The defect feature library is pre-established using historical inspection data and standard laboratory defect samples, and includes statistical distribution parameters for common defect types such as delamination, voids, resin-rich areas, fiber breakage, and poor interfacial bonding across various feature dimensions. The comparison uses weighted Euclidean distance to calculate the similarity score between the feature vector to be identified and the cluster centers of each type. The weights of each feature dimension are determined based on their contribution to classification accuracy in historical inspections.

[0040] S4544, Based on the comparison results, determine the type of the internal structural anomaly; In this step, the highest similarity score among all types is selected as the initial identification result. When the highest similarity score is lower than the preset confidence threshold, it indicates that the signal feature is significantly different from all known defect types in the database. This is then marked as an unknown defect type, triggering the subsequent dynamic update process of the defect database. The confidence threshold is set based on historical false positive rate statistics to achieve a balance between identification accuracy and the ability to discover unknown types.

[0041] S4545, assess the severity of the internal structural anomaly based on its type and size.

[0042] In this step, a quantitative assessment is performed based on the identified defect type and size information, combined with a pre-defined defect severity evaluation model. The evaluation model considers the different degrees of impact of defect types on structural strength, the proportion of defect size to wall thickness, and the stress characteristics of the defect area. It outputs a quantified severity level, providing a basis for subsequent maintenance decisions and quality control.

[0043] In the technical solution of the above embodiments, by jointly extracting time-domain and frequency-domain features from the dynamic compensation differential thermal response signal and comparing it with a preset defect feature library, the accurate identification of defect types and severity assessment are achieved, forming a complete process from signal detection to qualitative and quantitative analysis of defects.

[0044] In one possible design, Figure 5 This is a flowchart illustrating step S4545 according to an exemplary embodiment. (Refer to the attached document.) Figure 5 In step S4545, the step of assessing the severity of the internal structural anomaly based on its type and size includes: S45451, Obtain the geometric shape data and material distribution information of the area to be detected on the hull; In this step, the geometric data includes structural data such as the radius of curvature distribution, wall thickness distribution, and stiffener locations of the area to be inspected, obtained through laser 3D scanning or CAD model import. Material distribution information includes the number of carbon fiber fabric layers, the fiber orientation angles of each layer, the resin type, and its thermophysical parameters. This data provides an accurate geometric and material model foundation for subsequent finite element stress analysis, ensuring the accuracy of stress calculations.

[0045] S45452, Based on the geometric data and material distribution information, and combined with the load conditions borne by the hull in actual use, calculate the stress distribution in the hull area; In this step, the acquired geometric data and material distribution information are imported into finite element analysis software, and boundary conditions of water pressure loads and bending loads are applied to the racing boat during high-speed sailing and turning. The stress distribution of the area to be tested under various typical working conditions is calculated, and high-stress concentration areas and low-stress areas are identified in the form of stress cloud maps. Even small defects located in high-stress areas pose a much greater potential hazard than defects of the same size in low-stress areas.

[0046] S45453, The severity of the internal structural anomaly is assessed based on its type, size, and calculated stress distribution.

[0047] In this step, the defect type and size information identified in the previous step are overlaid with the calculated stress distribution data for analysis. Defects located in high stress concentration areas, even if small in size, are assessed as high-severity; defects located in low-stress areas are assessed according to conventional criteria. By integrating structural mechanics analysis with defect detection results, quantitative and personalized defect risk assessment is achieved.

[0048] In the technical solution of the above embodiments, by acquiring the geometric shape data and material distribution information of the area to be inspected on the hull, and combining it with the actual load conditions to perform finite element stress analysis, the physical characteristics of the defect are combined with the actual stress conditions of the hull, and the severity of the defect is quantitatively assessed, avoiding the limitations of relying solely on the type and size of the defect to determine the severity.

[0049] In one example, a void with a diameter of approximately 5 millimeters was discovered in a certain area of ​​the ship's bottom plate. By obtaining geometric data such as the thickness and curvature of the bottom plate in this area, as well as material information such as the carbon fiber laminate structure, and combining this with finite element analysis, it was calculated that the void was located in a region subjected to high tensile stress over a long period of time, with a high stress concentration factor. Although the void was small in size, it was reclassified as high-risk based on the stress assessment results.

[0050] In one possible design, Figure 6 This is a flowchart illustrating step S4541 according to an exemplary embodiment. (Refer to the attached document.) Figure 6 In step S4541, the step of extracting time-domain features from the dynamically compensated differential thermal response signal includes: S45411, Obtain the geometric shape data and material distribution information of the area to be detected on the hull; In this step, the geometric data includes the surface curvature radius and wall thickness distribution of each sub-region of the area to be detected, and the material distribution information includes the number of carbon fiber fabric layers, the fiber orientation angle of each layer, and the interlayer thickness data. This information provides the necessary spatial reference and material reference for the normalization processing of the time-domain signal, enabling signals under different geometric and material conditions to be compared on a uniform scale.

[0051] S45412, Based on the geometric data and the material distribution information, perform local region normalization processing on the dynamic compensation differential thermal response signal; In this step, based on geometric data and material distribution information, the detection area is divided into several sub-regions with similar geometric and material characteristics. The dynamic compensation differential thermal response signal within each sub-region is normalized and scaled according to local statistical characteristics (such as mean and standard deviation). For sub-regions with larger material thickness or curvature, the normalization coefficients are adjusted accordingly to compensate for their inherent signal amplitude differences. Normalization processing unifies signals under different geometric and material conditions to a comparable scale, eliminating systematic biases in temporal characteristics caused by geometric and material differences.

[0052] S45413, performs time-domain feature extraction on the normalized dynamic compensation differential thermal response signal to obtain the peak occurrence time, peak amplitude, decay rate and signal duration of the signal.

[0053] In this step, time-domain features are extracted from the normalized signal, including identifying the time point of signal peak occurrence, measuring the change in peak amplitude relative to the normalized baseline, calculating the rate at which the signal decays from the peak to half-peak, and the total duration from the start to recovery to the baseline level. The time-domain features after normalization eliminate interference from geometric and material differences, more accurately reflecting the time-domain response characteristics of the defect itself.

[0054] In the technical solution of the above embodiments, by acquiring geometric shape data and material distribution information, the dynamic compensation differential thermal response signal is subjected to local region normalization processing, which eliminates the interference of hull geometric complexity and material distribution differences on time-domain feature extraction and improves the accuracy and comparability of time-domain features.

[0055] In one possible design, step S4542, the step of extracting frequency domain features from the dynamically compensated differential thermal response signal, includes: S45421, Obtain the geometric shape data and material distribution information of the area to be detected on the hull; In this step, the geometric data includes a three-dimensional surface model of the area to be tested and the local wall thickness distribution. The material distribution information includes the number of carbon fiber fabric layers, the fiber orientation of each layer, and the thermal diffusivity of the resin matrix. The spatial variation of wall thickness and curvature has a systematic impact on the spectral characteristics of the thermal response. These data provide a reference for the thermophysical characteristics of each sub-region for frequency domain calibration.

[0056] S45422, Based on the geometric data and the material distribution information, perform local region frequency domain calibration processing on the dynamic compensation differential thermal response signal; In this step, the frequency domain amplitude and phase compensation coefficients of each sub-region are calculated based on its geometric characteristics (radius of curvature, wall thickness) and the material's thermal diffusion time constant. For regions with large curvature or thick walls, their inherent spectral characteristics generate additional energy distribution in the low-frequency band. Frequency domain compensation eliminates this spectral deviation introduced by geometric and material factors. The calibrated signal eliminates interference from non-defect factors in the frequency domain, allowing the frequency domain characteristics to more accurately reflect the impact of defects on heat wave propagation.

[0057] S45423 performs frequency domain feature extraction on the dynamically compensated differential thermal response signal after frequency domain calibration to obtain the energy distribution and main frequency components of the signal.

[0058] In this step, a Fast Fourier Transform is performed on the frequency-domain calibrated signal to obtain the calibrated power spectral density distribution. Frequency domain characteristic parameters such as peak frequency, peak amplitude, 3dB bandwidth, and spectral energy concentration of the dominant frequency component are extracted. The calibrated frequency domain features eliminate interference from geometric and material differences, more accurately reflecting the influence of internal material defects on heat wave propagation. Together with the time-domain features, they constitute a multi-dimensional feature space for defect identification.

[0059] In the technical solution of the above embodiments, the dynamic compensation differential thermal response signal is calibrated in the local region by acquiring the geometric shape data and material distribution information of the hull, which further eliminates the interference of material differences and geometric complexity on frequency domain feature extraction, and complements the time domain normalization.

[0060] In one example, when inspecting a ship hull with complex curved surfaces and varying thickness regions, laser 3D scanning was used to acquire geometric data, which was then combined with ultrasonic thickness measurement results to establish material distribution information. Before frequency domain feature extraction, the system performed frequency domain amplitude and phase compensation on each sub-region to eliminate the dominant frequency shift introduced by geometric differences. The calibrated spectrum analysis identified an abnormal frequency component at 1.2 kHz, which matched a typical spectral pattern of poor interface bonding, thus identifying a defect in the fiber-resin interface bonding.

[0061] In one possible design, step S4543, which compares the time-domain features and the frequency-domain features with a preset defect feature library, includes: S45431, Perform a preliminary comparison of the time-domain features and the frequency-domain features to identify signals with low pattern matching degree in the preset defect feature library; In this step, the extracted joint time-domain and frequency-domain feature vectors are compared one by one with the templates of each type in the defect feature library, and the weighted Euclidean distance is calculated as the similarity score. Signals with a total similarity score lower than the preset matching threshold are marked as signals with low matching degree, indicating that they may belong to unknown defect types or variations of known defects, and proceed to the subsequent in-depth analysis process.

[0062] S45432, For signals with low matching degree, analyze the combination of their time domain features and frequency domain features, and identify features that differ from the patterns in the preset defect feature library; In this step, a feature-dimensional deviation analysis is performed on signals with low matching degrees, comparing their deviation direction and magnitude from the nearest neighbor defect mode in dimensions such as peak time, peak amplitude, decay rate, dominant frequency component, and spectral energy distribution. A differential feature vector is constructed to identify the feature dimensions with the most significant deviations and their magnitudes, forming a differentiated feature fingerprint for the signal.

[0063] S45433, Based on the characteristics of the differences, combined with the physical properties and manufacturing process of the hull material, the formation mechanism of the defect is inferred; In this step, the difference feature vectors are correlated with the material physical property database and the manufacturing process knowledge base. For example, if the difference is mainly manifested in abnormal energy attenuation in the high-frequency band, combined with the material batch information and layup process parameters in that region, it can be inferred whether it is caused by poor fiber-resin interface bonding or insufficient resin impregnation. The material-process correlation reasoning establishes a mapping relationship based on historical manufacturing data and known defect cases.

[0064] S45434, Based on the inferred defect formation mechanism, a new defect feature pattern is constructed, and the new defect feature pattern is expanded to the preset defect feature library; In this step, based on the inferred defect formation mechanism, a feature pattern template for this novel defect is defined in the time and frequency domain feature spaces, including the center value and allowable deviation range of each feature dimension. The constructed new pattern and its metadata (formation mechanism, material association information, and first discovery time) are written into the defect feature library and indexed, enabling the library to identify this type of novel defect.

[0065] S45435, using the expanded defect feature library, the signals with low matching degree in the initial comparison are compared again to determine the type of internal structural anomaly.

[0066] In this step, the expanded defect feature library is used to perform a second round of comparison on signals with low matching scores in the initial comparison. The newly added defect patterns cover the signal feature space regions that could not be matched before. The similarity scores of each type are recalculated, and the defect type is determined for signals that reach the confidence threshold. The final identification result is then output.

[0067] In the technical solution of the above embodiments, by identifying the difference features of signals with low matching degree, inferring the defect formation mechanism by combining material and process information, constructing new defect feature patterns and expanding them to the defect feature library, the ability to automatically discover and identify new defects is realized, so that the defect feature library can continuously evolve and improve with the accumulation of detection data.

[0068] In one possible design, step S45433, the step of inferring the formation mechanism of the defect based on the characteristics of the difference, combined with the physical properties and manufacturing process of the hull material, includes: S454331, Obtain material batch information and manufacturing area identification; In this step, the material batch information includes the batch number of the carbon fiber prepreg, the resin formulation version, and the curing process parameters. The manufacturing area identifier indicates the layup station and operating team to which the hull area belongs during the manufacturing process. This information is retrieved from the Manufacturing Execution System (MES) and provides a data foundation for subsequent material-process correlation analysis and defect cause tracing.

[0069] S454332, Based on the material batch information and the manufacturing area identifier, obtain the corresponding material features and process features from the preset material feature library; In this step, the material feature library records the model, supplier, mechanical property parameters, and thermophysical parameters of different batches of fibers and resins. The process feature library records process parameters such as layup sequence, curing temperature profile, and molding pressure in different manufacturing areas. By indexing and matching batch numbers and area identifiers, the relevant material and process background information of the defect location is extracted.

[0070] S454333, perform correlation analysis between the characteristics of the difference and the material characteristics and the process characteristics, and infer the formation mechanism of the defect by combining the manufacturing process flow.

[0071] In this step, the signal's differential characteristics are correlated with extracted material and process features using a multi-dimensional correlation analysis. For example, if the difference is mainly manifested as abnormal attenuation in a specific frequency range, and there are records of resin flowability deviations in the prepreg batch used in that area, combined with the layup station and curing temperature profiles, it can be inferred that the defect is caused by localized resin deficiency due to insufficient resin flow. The correlation analysis results provide a traceability basis for manufacturing process improvement.

[0072] In the technical solution of the above embodiments, by obtaining material batch information and manufacturing area identifiers, corresponding material and process features are extracted from a preset feature library, and the difference features are correlated with material and process information, thus realizing the traceability from defect signals to formation mechanisms and providing data support for manufacturing process improvement.

[0073] In summary, the visual inspection method for quality defects in the curing process of racing boat hulls provided in Embodiment 1 of this application suppresses coating effect interference by applying two thermal excitations and performing dynamic compensation differential analysis, accurately identifies defect types by jointly extracting time-domain and frequency-domain features and comparing them with a defect feature library, assesses the severity of defects by combining the stress distribution of the hull, eliminates geometric difference interference by local normalization and frequency domain calibration, and adapts to new defect patterns by dynamically updating the defect library and material-process correlation traceability, thus forming a full-process quality inspection scheme from thermal excitation design to defect cause analysis.

[0074] Example 2 Embodiment 2 of this application provides a visual inspection system for quality defects in the curing process of a racing boat hull. Figure 7 This is a block diagram illustrating a visual inspection system for quality defects in the curing process of a racing boat hull, according to another exemplary embodiment. Figure 7 As shown, the system includes: The first thermal excitation application module 01 is used to apply a first thermal excitation, which has a preset energy and duration. In this embodiment, the first thermal excitation application module 01 includes a high-energy xenon flash lamp array and a high-voltage capacitor bank, and can output a first thermal excitation with an energy range of 1kJ to 10kJ and a pulse duration of 1ms to 20ms.

[0075] The first temperature data acquisition module 02 is used to acquire the first temperature change data of the hull surface over a period of time after the first thermal excitation is applied. In this embodiment, the first temperature data acquisition module 02 includes a high frame rate and high resolution infrared thermal imager, which is connected to the data processing unit via an industrial Ethernet to collect and store the surface temperature change image sequence after the first thermal excitation in real time.

[0076] The secondary excitation acquisition module 03 is used to apply a second thermal excitation to the surface of the hull before the thermal effect generated by the first thermal excitation has completely dissipated, and to acquire second temperature change data of the hull surface over a period of time. The energy or duration of the second thermal excitation differs from that of the first thermal excitation. In this embodiment, the secondary excitation acquisition module 03 triggers the second thermal excitation and acquires a second set of temperature change data after the delay time of the first thermal excitation.

[0077] The data comparison and processing module 04 is used to compare and process the first temperature change data and the second temperature change data to generate a signal reflecting the difference between the two. In this embodiment, the data comparison and processing module 04 analyzes and processes the two sets of temperature change data, extracts the coating relaxation features, and generates a dynamic compensation differential thermal response signal.

[0078] The structural anomaly identification module 05 is used to identify structural anomalies inside the hull material based on the difference signal. In this embodiment, the structural anomaly identification module 05 performs time-domain and frequency-domain feature extraction based on the dynamic compensation differential thermal response signal, compares it with the defect feature library, and evaluates the severity of the defect.

[0079] In summary, the visual inspection system for quality defects in the curing process of racing boat hull provided in Embodiment 2 of this application includes a first thermal excitation application module 01, a first temperature data acquisition module 02, a secondary excitation acquisition module 03, a data comparison and processing module 04, and a structural anomaly identification module 05, which work together to complete the entire process of inspection tasks from the application of dual-pulse thermal excitation, temperature data acquisition and comparison processing to structural anomaly identification.

[0080] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0081] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A visual inspection method for quality defects in the curing process of a racing boat hull, characterized in that, Includes the following steps: A first thermal excitation is applied, the first thermal excitation having a preset energy and duration; After the first thermal excitation is applied, the first temperature change data of the hull surface over a period of time is obtained; Before the thermal effect generated by the first thermal excitation has completely dissipated, a second thermal excitation is applied to the surface of the hull, and second temperature change data of the hull surface over a period of time is obtained. The energy or duration of the second thermal excitation differs from that of the first thermal excitation. The first temperature change data and the second temperature change data are compared and processed to generate a signal that reflects the difference between the two. Based on the difference signals, structural anomalies within the hull material are identified.

2. The method according to claim 1, characterized in that, The step of comparing the first temperature change data and the second temperature change data includes: After the first thermal excitation is applied, the temperature decay process of each local area on the hull surface is continuously monitored; Based on the temperature decay process, analyze the recovery rate and trend of temperature in each local area of ​​the hull surface; Based on the recovery rate and trend, determine the delay time for applying the second thermal excitation; After the delay time, the second thermal excitation is applied to the surface of the hull, and the second temperature change data is acquired; The first temperature change data and the second temperature change data are compared and processed to generate a signal that reflects the difference between the two.

3. The method according to claim 2, characterized in that, The step of comparing the first temperature change data and the second temperature change data to generate a signal reflecting the difference between the two includes: Analyze the first temperature change data to extract specific features that reflect the relaxation process after the coating properties change; Based on the specific characteristics, the parameters of the thermal diffusion model are adjusted to generate a thermal response baseline curve, which represents the expected behavior of the surface temperature of the defect-free area decaying over time under the current coating state. The second temperature change data is compared with the thermal response reference curve, and the difference is calculated to generate a dynamically compensated differential thermal response signal. Based on the dynamic compensation differential thermal response signal, structural anomalies inside the hull material are identified.

4. The method according to claim 3, characterized in that, The step of identifying structural anomalies inside the hull material based on the dynamically compensated differential thermal response signal includes: The dynamic compensation differential thermal response signal is subjected to time-domain feature extraction to obtain the peak occurrence time, peak amplitude, decay rate and signal duration of the signal; Frequency domain feature extraction is performed on the dynamically compensated differential thermal response signal to obtain the energy distribution and dominant frequency components of the signal; The time-domain features and the frequency-domain features are compared with a preset defect feature library, which contains typical time-domain and frequency-domain feature patterns of different types of internal structural anomalies in dynamic compensation differential features. Based on the comparison results, the type of the internal structural anomaly is determined; The severity of the internal structural anomaly is assessed based on its type and size.

5. The method according to claim 4, characterized in that, The step of assessing the severity of the internal structural anomaly based on its type and size includes: Obtain the geometric shape data and material distribution information of the area to be detected on the hull; Based on the geometric data and material distribution information, and combined with the load conditions borne by the hull in actual use, the stress distribution in the hull area is calculated. The severity of the internal structural anomaly is assessed based on its type, size, and calculated stress distribution.

6. The method according to claim 4, characterized in that, The step of extracting time-domain features from the dynamically compensated differential thermal response signal includes: Obtain the geometric shape data and material distribution information of the area to be detected on the hull; Based on the geometric data and the material distribution information, the dynamic compensation differential thermal response signal is subjected to local region normalization processing to eliminate the influence of local region material thickness, curvature or internal structure depth differences on temporal characteristics. Time-domain features are extracted from the normalized dynamic compensation differential thermal response signal to obtain the peak occurrence time, peak amplitude, decay rate, and signal duration of the signal.

7. The method according to claim 4, characterized in that, The step of extracting frequency domain features from the dynamically compensated differential thermal response signal includes: Obtain the geometric shape data and material distribution information of the area to be detected on the hull; Based on the geometric data and the material distribution information, the dynamic compensation differential thermal response signal is subjected to local region frequency domain calibration processing to eliminate the influence of local region material thickness, curvature or internal structure depth differences on frequency domain characteristics; Frequency domain features are extracted from the dynamically compensated differential thermal response signal after frequency domain calibration to obtain the energy distribution and dominant frequency components of the signal.

8. The method according to claim 4, characterized in that, The step of comparing the time-domain features and the frequency-domain features with a preset defect feature library includes: A preliminary comparison is performed on the time-domain features and the frequency-domain features to identify signals with low pattern matching degree in the preset defect feature library; For signals with low matching degree, analyze the combination of their time domain features and frequency domain features to identify features that differ from the patterns in the preset defect feature library; Based on the characteristics of the differences, combined with the physical properties of the hull material and the manufacturing process, the formation mechanism of the defects is inferred; Based on the inferred defect formation mechanism, a new defect feature pattern is constructed, and the new defect feature pattern is expanded to the preset defect feature library; Using the expanded defect feature library, signals with low matching degree in the initial comparison are compared again to determine the type of internal structural anomaly.

9. The method according to claim 8, characterized in that, The step of inferring the formation mechanism of the defect based on the characteristics of the difference, combined with the physical properties of the hull material and the manufacturing process, includes: Obtain material batch information and manufacturing area identification; Based on the material batch information and the manufacturing area identifier, the corresponding material features and process features are obtained from the preset material feature library; By correlating the characteristics of the differences with the material characteristics and the process characteristics, and combining this with the manufacturing process flow, the formation mechanism of the defects can be inferred.

10. A visual inspection system for quality defects in the curing process of a racing boat hull, characterized in that, The system includes: A first thermal excitation application module is used to apply a first thermal excitation, wherein the first thermal excitation has a preset energy and duration; The first temperature data acquisition module is used to acquire the first temperature change data of the hull surface over a period of time after the first thermal excitation is applied. The secondary excitation acquisition module is used to apply a second thermal excitation to the surface of the hull before the thermal effect generated by the first thermal excitation has completely dissipated, and to acquire second temperature change data of the hull surface over a period of time. The energy or duration of the second thermal excitation differs from that of the first thermal excitation. The data comparison and processing module is used to compare and process the first temperature change data and the second temperature change data to generate a signal reflecting the difference between the two. The structural anomaly identification module is used to identify structural anomalies inside the hull material based on the difference signals.

Images