Intelligent image enhancement method for internal flaw detection of power plant equipment

Abstract

The invention relates to the technical field of image enhancement, in particular to an intelligent image enhancement method for internal flaw detection of power plant equipment, which comprises the following steps: acquiring an original image acquired by flaw detection equipment, and separating a base material signal layer and an oxide layer interference signal in the original image based on an equipment material high-temperature oxidation kinetic model; frequency domain distribution of oxide layer interference signals is reconstructed according to material phase change temperature points, band elimination filtering is executed in a characteristic frequency interval, decoupled oxide layer signals are generated, a base material signal layer and the decoupled oxide layer signals are input into a self-adaptive compensation network, fusion weight is adjusted according to defect form gradient, and an enhanced flaw detection image is output. According to the method, a signal decoupling mechanism based on material phase change behavior modeling is introduced, so that effective separation of an oxide layer interference signal and a base material defect signal is realized, a tiny sound velocity abrupt change region caused by phase change can be accurately identified, enhanced expression of microcrack features is realized, and the image signal-to-noise ratio and the defect boundary resolution are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of image enhancement technology, and in particular to a smart image enhancement method for internal flaw detection in power plant equipment. Background Technology

[0002] In large energy equipment such as thermal power plants and nuclear power plants, internal structural materials (such as P91 steel and 12Cr1MoV alloy) are prone to oxidation corrosion and microcrack accumulation during long-term high-temperature and high-pressure service, which seriously threatens the safety of equipment operation. Although ultrasonic phased array, infrared thermal imaging and eddy current detection are widely used for internal defect detection, they still face significant image degradation problems in the imaging stage. In particular, the superposition of oxide layer interference signals seriously affects the identification of defect boundaries and size quantification. Most existing image enhancement technologies adopt a unified filtering strategy or a single image fusion method, which cannot simultaneously preserve the high-frequency crack features and suppress the low-frequency redundant signals of the oxide layer. This results in low defect detection rate, high false positive rate, and the enhanced image is disconnected from the physical properties of the material, affecting the reliability of the project.

[0003] Furthermore, existing enhancement methods lack effective modeling of material acoustic constraints and defect morphology characteristics, often ignoring the effects of material thickness variations, phase transition-induced sound velocity changes, and local curvature on signal propagation paths and amplitudes, thus failing to achieve physical consistency control in image enhancement. Commonly used frequency domain filtering-based denoising methods struggle to adaptively adjust stopband width and lack signal source differentiation capabilities. Traditional fusion methods cannot automatically adjust fusion weights based on defect gradients, easily leading to edge blurring and artifact residue. Therefore, there is an urgent need for an intelligent enhancement method with phase transition decoupling capabilities, frequency domain adaptive filtering mechanisms, and acoustic impedance physical constraint models to achieve a realistic, clear, and measurable perceptual representation of equipment material defect images. Summary of the Invention

[0004] This invention provides a smart image enhancement method for internal flaw detection in power plant equipment.

[0005] A smart image enhancement method for internal flaw detection of power plant equipment includes the following steps: S1: Acquire the original image collected by the flaw detection equipment, and separate the substrate signal layer and oxide layer interference signal in the original image based on the high-temperature oxidation kinetic model of the equipment material; S2: Reconstruct the frequency domain distribution of the oxide layer interference signal based on the material phase transition temperature point, perform band-stop filtering in the characteristic frequency range, and generate the decoupled oxide layer signal; S3: Input the substrate signal layer and the decoupled oxide layer signal into the adaptive compensation network, adjust the fusion weights according to the defect morphology gradient, and output the final flaw detection image.

[0006] Optionally, S1 includes: S11: Obtain the original flaw detection image inside the equipment through a multi-frequency eddy current flaw detection probe, and simultaneously collect the real-time surface temperature and historical operating temperature curve of the target area; S12: Obtain oxidation kinetic parameters, including oxidation rate constant, activation energy and reaction index, from the material database based on the equipment material grade, and calculate the oxide layer thickness distribution; S13: Construct an acoustic impedance matching model to separate the signal layer of the substrate and the interference signal of the oxide layer.

[0007] Optionally, S12 includes: S121: Obtain the oxidation kinetic parameters of the relevant materials, including oxidation rate constant, activation energy and reaction index, by querying the material grade of the equipment; S122: Based on the obtained oxidation kinetic parameters, calculate the oxide layer thickness distribution at different locations.

[0008] Optionally, S13 includes: S131: By calculating the acoustic impedance ratio of the substrate and the oxide layer, the gray value of the substrate signal layer is obtained, and then corrected by combining the thickness and attenuation coefficient of the oxide layer to separate the substrate signal. S132: Obtain oxide layer interference signal by removing substrate signal from original image.

[0009] Optionally, S2 includes: S21: Obtain the phase transition temperature point and sound velocity change rate of the equipment substrate; S22: Determine the stopband center frequency based on oxide layer thickness and sound velocity abrupt change rate; S23: Perform frequency domain processing on oxide layer interference signals; S24: The decoupled oxide layer signal is obtained by performing an inverse Fourier transform on the frequency domain processed signal.

[0010] Optionally, S23 includes: S231: Perform Fourier transform on the oxide layer interference signal to convert the time-domain signal into a frequency-domain signal; S232: Construct the transfer function of the band-stop filter based on the calculated stopband center frequency; S233: Use the constructed filter transfer function to filter the frequency domain signal; S234: Perform an inverse Fourier transform on the frequency domain processed signal to convert it back from the frequency domain to the time domain, restoring the original time domain form of the signal, thus completing the decoupling of the signal and the removal of interference signals.

[0011] Optionally, S3 includes: S31: Perform multi-scale morphological gradient calculation on the substrate signal layer; S32: Generate pixel-level fusion weights based on the gradient field of defect morphology; S33: The substrate signal layer and the decoupled oxide layer signal are fused according to weights to construct an enhanced flaw detection image; S34: Apply material acoustic impedance boundary constraints to the fused image to limit the grayscale range of the final image and output the final flaw detection image.

[0012] Optionally, S33 includes: S331: Based on the obtained fusion weights, the substrate signal image and the oxide layer signal image are weighted and fused. The substrate signal undergoes fidelity transformation to preserve details, and the oxide layer signal is enhanced in areas of weak interference through a compensation function. S332: Combine the two types of signals after processing according to the fusion weight to obtain an enhanced flaw detection image with defect significance and interference suppression capability.

[0013] Optionally, S34 includes: S341: Based on the acoustic impedance characteristics of the coupling medium, substrate and oxide layer, calculate the upper and lower boundaries of the grayscale values ​​of the fused image, including the lowest measurable value (related to the probe impedance ratio) and the highest confidence value (related to the substrate-oxide layer impedance ratio). S342: Intensity limiting is applied to the intermediate fused image to restrict the grayscale values ​​of all pixels within the acoustic impedance range, thus obtaining the final flaw detection image.

[0014] The beneficial effects of this invention are: This invention achieves effective separation of oxide layer interference signals and substrate defect signals by introducing a signal decoupling mechanism based on material phase transition behavior modeling. In particular, in frequency domain processing, an adaptive band-stop filtering method is adopted. By dynamically adjusting the center frequency and bandwidth of the filter and combining the oxide layer thickness and reference sound velocity ratio parameters, the energy peaks of interference frequency bands under different oxidation states are effectively suppressed, solving the problem that traditional frequency domain processing methods cannot adapt to oxide layers of multiple thicknesses. The enhancement and dynamic weight function construction technology can accurately identify the small sound velocity change regions caused by phase transition, realize the enhanced expression of microcrack features, and significantly improve the image signal-to-noise ratio and defect boundary resolution.

[0015] This invention proposes an adaptive fusion model based on defect morphology gradients, integrating multi-scale morphological analysis, curvature weight adjustment, and acoustic impedance physical boundary control mechanisms. This effectively ensures the dual consistency of the enhancement results in visual clarity and physical measurability. The fusion weights are jointly driven by gradient intensity, curvature change, and oxide layer thickness, adaptively distinguishing the dominant signal source. This ensures the preservation of high-frequency textures in microcrack-dominant regions and reduces redundant background in oxide interference regions. Simultaneously, acoustic impedance constraints are used to set the grayscale range of the enhanced image, preventing intensity saturation and acoustic distortion. This invention outperforms existing technologies in crack detection rate, false positive rate, and image restoration accuracy, demonstrating good engineering adaptability and promotional value. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention 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 only for this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart of a method according to an embodiment of the present invention; Figure 2 This is a frequency domain processing diagram according to an embodiment of the present invention. Detailed Implementation

[0018] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. It should also be noted that, to make the embodiments more comprehensive, the following embodiments are the best and preferred embodiments, and those skilled in the art can use other alternative methods to implement some well-known technologies; moreover, the accompanying drawings are only for more specific description of the embodiments and are not intended to specifically limit the present invention.

[0019] It should be noted that the use of terms such as "an embodiment," "an embodiment," "an exemplary embodiment," and "some embodiments" in the specification indicates that the described embodiment may include a specific feature, structure, or characteristic, but not every embodiment necessarily includes that specific feature, structure, or characteristic. Furthermore, when a specific feature, structure, or characteristic is described in connection with an embodiment, implementing such a feature, structure, or characteristic in conjunction with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the art.

[0020] Generally, terms can be understood at least partly from their use in context. For example, depending at least partly on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in a singular sense, or a combination of features, structures, or characteristics in a plural sense. Additionally, the term "based on" can be understood not necessarily to convey an exclusive set of factors, but rather, alternatively, depending at least partly on the context, to allow for the presence of other factors that are not necessarily explicitly described.

[0021] like Figures 1-2 As shown, a smart image enhancement method for internal flaw detection of power plant equipment includes the following steps: S1: Acquire the original image collected by the flaw detection equipment, and separate the substrate signal layer and oxide layer interference signal in the original image based on the high-temperature oxidation kinetic model of the equipment material; S1 specifically includes: S11: Obtain raw flaw detection images of the equipment's interior using a multi-frequency eddy current flaw detection probe. Simultaneously collect real-time surface temperature of the target area and historical operating temperature curve ,in Image pixel coordinates, As a variable for equipment operating time, the historical temperature curve of the equipment during long-term operation provides the temperature changes of the equipment at different time periods. By analyzing historical data, we can understand the temperature change pattern of the equipment and provide a reference for the equivalent temperature calculation in oxide layer modeling. Historical temperature data is crucial for estimating the heat treatment process of materials, the formation and aging process of oxide layer, and helps to conduct accurate oxidation kinetic simulation. S12: Obtain oxidation kinetic parameters, including oxidation rate constant, from the material database based on the equipment material grade. ,activation energy and reaction index Calculate the oxide layer thickness distribution , is represented as: ; in, It represents the oxide layer thickness, expressed as an exponential function that incorporates the effects of oxidation rate, temperature, and time. This allows for an accurate description of the oxidation process of equipment under different operating conditions. The oxidation rate constant is a constant that describes the rate of oxidation of a material. It is affected by the material's chemical composition, temperature, and other environmental factors. Specific values ​​are obtained by querying a database to more accurately model the oxidation behavior of different materials. It is the minimum energy required for a material to undergo an oxidation reaction. It is closely related to temperature, reflecting the temperature sensitivity of the oxidation process. The activation energy directly affects the oxidation rate of a material; a higher activation energy means that the material oxidizes at a lower temperature. The gas constant represents the proportionality coefficient in the gas law. It is a universal constant that does not change with the material and takes the standard value. , The equivalent temperature is a comprehensive temperature value calculated based on historical temperature curves and a location-related temperature weighting function. It reflects the actual temperature influence at a given location. Equivalent temperature, by considering the temperature changes over time and space, reflects the true temperature conditions at different locations of the equipment and is a core parameter in oxide layer thickness calculation. This is a location-dependent temperature weighting function, reflecting the degree of influence of temperature at different locations on the oxidation process. For effective high-temperature operation time, i.e., temperatures above 400°C... When the cumulative time the equipment is actually in a high-temperature state is less than 400... The oxidation process almost stops at this point, so only values ​​above 400 are counted. The time period should be determined to avoid overestimating the oxide layer thickness. S13: Construct an acoustic impedance matching model to separate the signal layer, represented as: ; in, This is the grayscale value of the substrate signal layer, representing the signal intensity reflected or transmitted by the substrate in the original image. It is used to separate the substrate signal from the original image, and its value ranges from 0 to 255, representing the pixel value range of a standard grayscale image. This is the grayscale value of the interference signal caused by the oxide layer, representing the intensity of the signal interference caused by the oxide layer. The value ranges from 0 to 255. The interference signal from the oxide layer is usually weak, resulting in a low corresponding grayscale value. Furthermore, as the oxide layer thickness increases, the degree of interference may increase, thereby increasing the grayscale value. This is the acoustic impedance of the substrate, representing the resistance to sound wave propagation within the substrate. Its value ranges from 10 to 20. The acoustic impedance of the substrate reflects the acoustic properties of its material. Metallic materials typically have higher acoustic impedance. This value range is based on the acoustic properties of common industrial materials. This refers to the acoustic impedance of the oxide layer, which represents the resistance to sound wave propagation within the oxide layer. Its value ranges from 5 to 15. The acoustic impedance of the oxide layer is lower than that of the substrate because the oxide layer is typically looser and has a lower density. The attenuation coefficient is... The value ranges from 1 to 10. The attenuation coefficient is directly related to the signal frequency and the sound velocity in the oxide layer. The thicker the oxide layer, the greater the signal attenuation. Therefore, it is necessary to adjust the attenuation coefficient to accurately simulate the signal propagation process. The center frequency of the flaw detection wave represents the frequency characteristics of the signal, ranging from 1 to 10. Higher frequencies result in higher signal resolution but also lead to increased signal attenuation. Different frequencies are suitable for different flaw detection requirements. It is the sound velocity of the oxide layer, with a value ranging from 1000 to 3000; The grayscale value of the substrate signal layer is obtained by calculating the acoustic impedance ratio of the substrate and the oxide layer. The difference between the acoustic impedance of the substrate and the oxide layer determines the propagation and attenuation of the signal. The signal in the original image is multiplied by the acoustic impedance ratio and adjusted in conjunction with the thickness and attenuation coefficient of the oxide layer to separate the substrate signal layer. The oxide layer interference signal is obtained by subtracting the substrate signal layer from the original image. The intensity of the oxide layer signal is related to the thickness, acoustic impedance, and attenuation coefficient of the oxide layer. This separation method can effectively reduce the interference of the oxide layer on the final flaw detection image and improve the clarity and accuracy of the signal.

[0022] S2: Reconstruct the frequency domain distribution of the oxide layer interference signal based on the material phase transition temperature point, perform band-stop filtering in the characteristic frequency range, and generate the decoupled oxide layer signal; S2 specifically includes: S21: Obtain the phase transition temperature of the equipment substrate. and the rate of change of the speed of sound, expressed as: ; ; in, This refers to the phase transformation temperature of a material, which is the temperature at which a material transforms from a low-temperature phase (such as ferrite) to a high-temperature phase (such as austenite). The value is retrieved from a material database and ranges from 730 to 810. yes The speed of sound at high temperatures (in the austenitic phase) ranges from 3000 to 5000. The speed of sound in the austenitic phase is higher because its atomic structure is more loose, leading to faster sound wave propagation. yes The velocity of sound at low temperatures (in the low-temperature ferrite phase) ranges from 1000 to 2500. The velocity of sound in the low-temperature ferrite phase is relatively low because its compact structure and small interatomic spacing result in a relatively slow propagation speed of sound waves. It is the reference speed of sound. , It is the rate of change of sound velocity, which represents the intensity of the change in sound velocity near the phase transition temperature of a material. The abrupt change rate of sound velocity reflects the degree of change in sound velocity between high-temperature and low-temperature phases of a material, and depends on the specific phase transition behavior of the material. Phase transition temperature of the substrate This refers to the temperature at which a material transforms from a low-temperature ferrite phase to a high-temperature austenite phase. At this temperature, the material's physical properties change significantly, especially the velocity of sound. The abrupt change in velocity of sound caused by the phase transition is crucial for signal processing of oxide layers, and it is necessary to calculate the change in velocity of sound. and the rate of change of sound speed This change will affect the propagation characteristics of the signal, and accurate calculation is necessary for subsequent frequency domain decoupling processing; S22: Based on oxide layer thickness and the rate of change of sound speed The stopband center frequency is determined as follows: ; in, This is the Poisson's ratio (obtained from a database), the ratio of lateral strain to longitudinal strain of a material under stress. It represents the ratio of lateral strain to longitudinal strain when the material is under stress, and its value ranges from 0.2 to 0.4. This is the elastic modulus of a material, representing its stiffness under tension or compression. Its value ranges from 100 to 210. The elastic modulus is related to the material's stiffness and affects the accuracy of frequency calculations. It is a geometric correction factor. , For pipe diameter, The geometric correction factor is used to correct the effect of the pipe curvature on signal propagation. The smaller the curvature, the higher the frequency. Therefore, it needs to be adjusted according to the pipe structure. Based on oxide layer thickness and the rate of change of sound speed The center frequency of the stopband, the thickness of the oxide layer, and the abrupt change rate of the sound velocity are important factors affecting the frequency characteristics of a filter. A thicker oxide layer results in more significant signal attenuation, leading to changes in the frequency range and Poisson's ratio. and elastic modulus It is also a parameter that affects the frequency characteristics of a signal, and the geometric correction factor Adjustments are made based on the geometry of the equipment to ensure that the frequency calculation matches the actual equipment structure; S23: Interference signal to oxide layer Perform frequency domain processing, specifically including: (1) Perform a Fourier transform on the oxide layer interference signal to obtain its frequency domain representation, which is expressed as: ; in, It is the frequency domain representation obtained after the oxide layer interference signal is Fourier transformed. The frequency domain representation helps to accurately identify and locate the interference components in a specific frequency band, which makes it easier for the filter to accurately suppress them at the frequency level. (2) Based on the calculated stopband center frequency Construct the transfer function of the band-stop filter to remove oxide interference signals in the frequency range, and determine the stopband width. Based on oxide layer thickness and reference oxide layer thickness Dynamic adjustment, represented as: ; in, It is the transfer function of the band-stop filter, which determines the degree to which each frequency component is retained. Spatial frequency, representing the frequency distribution of a signal in the frequency domain. , These are the frequency coordinates after the Fourier transform. Spatial frequency describes the periodic characteristics of a signal and is related to the spatial distribution of an image. It is used to distinguish different frequency components in the frequency domain. It is a stopband width control parameter. , It is the stopband center frequency. The stopband width control parameter is related to the oxide layer thickness. The greater the thickness, the narrower the filter bandwidth, allowing for more precise suppression of interference signals. It is the reference oxide layer thickness. As a standard reference value, it is used to standardize parameters so that the width of the filter can be adaptively adjusted under different oxide layer thicknesses; (3) By adaptively adjusting the width of the filter, the interference signal of the oxide layer can be accurately filtered out, and the filtering process can be expressed as follows: ; in, The frequency domain signal obtained after applying a band-stop filter (with oxide layer frequency band energy removed) is used to remove oxide layer interference components, retain substrate-related frequencies, and facilitate subsequent reconstruction of the real structure signal. (4) Inverse Fourier Transform: ; in, It is the decoupled image after inverse Fourier transform, that is, the time domain expression of the oxide layer interference signal after frequency domain filtering. It is used to output the image signal in which the interference components are effectively removed, so as to provide a purer interference layer for subsequent fusion with the substrate image. S24: Perform an inverse Fourier transform on the frequency domain processed signal to output the decoupled oxide layer signal, represented as: ; in, This is a mask for the oxide layer region, with values ​​ranging from 1 to 0. The intensity of the oxide layer interference signal... Greater than the preset intensity threshold At that time, the oxide layer region mask The value is 1, indicating the intensity of the oxide layer interference signal. Less than or equal to the preset intensity threshold At that time, the oxide layer region mask The value is 0, which is the preset intensity threshold. Setting a fixed value of 95 can effectively separate the substrate from the oxidation interference signal. It is the signal of the decoupled oxide layer; Perform an inverse Fourier transform on the frequency domain signal to reconstruct the time domain signal. The reconstructed signal... This still includes the interference portion of the oxide layer, therefore a mask is needed based on the oxide layer region. To further process the signal, only the effective signal area is retained. This mask ensures that only the signal in the oxide layer area is preserved, while the substrate signal is effectively filtered out.

[0023] S3: Input the substrate signal layer and the decoupled oxide layer signal into the adaptive compensation network, adjust the fusion weight according to the defect morphology gradient, and output the final flaw detection image; S3 specifically includes: S31: To extract edge features of defect areas and enhance the recognition of microcracks and macroscopic contours, the substrate signal layer... Multi-scale morphological gradient calculation is performed by introducing three structuring elements of different sizes to conduct morphological dilation and erosion operations on the image, calculating the gradient response at each scale, and then weighting and fusing them with scale weights to obtain the final defect morphological gradient field, expressed as: ; in, It is the first Scale-based structural elements; multi-scale structures adapt to defects of different sizes. Corresponding dimensions , These are scale weight coefficients, satisfying... ,and Emphasizing small-scale features (such as intergranular microcracks) while minimizing large-scale background. These represent morphological dilation and erosion operations, used to extract edge information. It is a normalized defect morphology gradient field, with values ​​ranging from... This provides scale-independent input for subsequent weight calculations, enhancing the flexibility of fusion. S32: To achieve accurate weighted fusion of signals from different regions (such as the substrate-dominant region and the oxidation interference region), based on the defect morphology gradient field... Constructing pixel-level fusion weight function This weighting function combines the normalized gradient with local image curvature information and controls the transition bandwidth through the sigmoid function, enabling the fusion strategy to dynamically adjust. It is expressed as: ; in, It is a pixel-level fusion weight function. This is the defect gradient threshold, used to delineate high / low morphological response regions, with a value ranging from 0.1 to 0.2. This is a morphological bias factor, with a fixed value of 1.0. In the crack region, it is set to a fixed value of 1.2 to improve the weighted response of the fusion function to high morphological intensity regions and prevent the loss of crack features. It is the transition steepness coefficient. The value ranges from 4 to 16. The greater the oxide layer thickness, the steeper the signal transition, which helps prevent interference from penetrating into the substrate area. It is the oxide layer signal curvature weight. The value range is 0.1-1.0. It is a two-dimensional Laplacian operator used to extract micro curvature changes in an image, enhance the response to weak oxide layer perturbations near the edges, and improve the clarity of the fusion boundary; S33: To ensure that the fusion result retains both the true defect features and the smoothness of the image and the authenticity of the signal, two transformation functions are introduced to enhance the substrate signal with fidelity and to perform amplitude compensation on the oxide layer signal, thus decoupling the substrate signal layer from the decoupled oxide layer signal. The enhanced flaw detection image is constructed by weighted fusion and represented as follows: ; in, This is an enhanced flaw detection image. It is a substrate signal fidelity transformation. , ,use Enhance low-grayscale details while maintaining overall dynamic range without expansion. It is an oxide layer signal compensation transform function, used to enhance the oxide layer signal. This smooths and enhances the weak signal in the oxide layer, preventing saturation and also preventing the denominator from being zero. It is to prevent division by zero constant. To ensure numerical stability; S34: To ensure the physical interpretability of the grayscale values ​​of the fused image and to match the acoustic impedance response of the actual flaw detection system, material acoustic impedance boundary constraints are applied to the fused image. Upper and lower boundary limits are constructed using the acoustic impedance of the probe, substrate, and oxide layer. The fused image is then cropped to this effective range, and the final flaw detection image is output, represented as follows: ; in, It is the two-dimensional grayscale distribution of the final flaw detection image output. By limiting the image grayscale to a physically reasonable acoustic impedance grayscale range that the flaw detection equipment can perceive, it ensures that the grayscale value of each pixel in the image has true physical meaning and avoids distortion phenomena such as oversaturation or over-attenuation. It is the minimum grayscale value of the output image. To prevent image distortion or information loss due to excessively low grayscale values, It is the lower limit scaling factor, used to set the minimum physical boundary for enhancing grayscale values ​​in an image. It represents an engineering balance between preventing image distortion and ensuring the visualization of weak signals. It is the maximum grayscale value of the output image. Limit the grayscale value to the upper limit of physical reflectivity to avoid distortion or false detection of bright areas. It is an upper limit control factor used to control the maximum output range of image grayscale values. It represents the best compromise between ensuring high reflectivity contrast, avoiding overexposure and saturation, and enhancing defect discernibility. This refers to the probe impedance. The coupling medium used in the probe is generally a low-impedance material such as water or gel, and a fixed value of 1.5 is taken. This is the acoustic impedance of the substrate, used to determine the sound energy penetration and reflection ratio, affecting the upper limit of the grayscale range, and is set to a fixed value of 45. It is the acoustic impedance of the oxide layer. The oxide layer density and sound velocity determine the reflected energy ratio, which affects the lower limit of low grayscale in the image. The value range is 25-35.

[0024] This invention encompasses any substitutions, modifications, equivalent methods, and solutions made within the spirit and scope of this invention. To provide the public with a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments; however, those skilled in the art will fully understand the invention even without these details. Furthermore, to avoid unnecessary misunderstanding of the essence of this invention, well-known methods, processes, procedures, components, and circuits are not described in detail.

[0025] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A smart image enhancement method for internal flaw detection of power plant equipment, characterized in that, Includes the following steps: S1: Acquire the original image collected by the flaw detection equipment, and separate the substrate signal layer and oxide layer interference signal in the original image based on the high-temperature oxidation kinetic model of the equipment material; S2: Reconstruct the frequency domain distribution of the oxide layer interference signal based on the material phase transition temperature point, perform band-stop filtering in the characteristic frequency range, and generate the decoupled oxide layer signal; S3: Input the substrate signal layer and the decoupled oxide layer signal into the adaptive compensation network, adjust the fusion weights according to the defect morphology gradient, and output the final flaw detection image.

2. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 1, characterized in that, S1 includes: S11: Obtain the original flaw detection image inside the equipment through a multi-frequency eddy current flaw detection probe, and simultaneously collect the real-time surface temperature and historical operating temperature curve of the target area; S12: Obtain oxidation kinetic parameters, including oxidation rate constant, activation energy and reaction index, from the material database based on the equipment material grade, and calculate the oxide layer thickness distribution; S13: Construct an acoustic impedance matching model to separate the signal layer of the substrate and the interference signal of the oxide layer.

3. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 2, characterized in that, S12 includes: S121: Obtain the oxidation kinetic parameters of the relevant materials, including oxidation rate constant, activation energy and reaction index, by querying the material grade of the equipment; S122: Based on the obtained oxidation kinetic parameters, calculate the oxide layer thickness distribution at different locations.

4. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 2, characterized in that, S13 includes: S131: By calculating the acoustic impedance ratio of the substrate and the oxide layer, the gray value of the substrate signal layer is obtained, and then corrected by combining the thickness and attenuation coefficient of the oxide layer to separate the substrate signal. S132: Obtain oxide layer interference signal by removing substrate signal from original image.

5. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 4, characterized in that, S2 includes: S21: Obtain the phase transition temperature point and sound velocity change rate of the equipment substrate; S22: Determine the stopband center frequency based on oxide layer thickness and sound velocity abrupt change rate; S23: Perform frequency domain processing on oxide layer interference signals; S24: The decoupled oxide layer signal is obtained by performing an inverse Fourier transform on the frequency domain processed signal.

6. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 5, characterized in that, S23 includes: S231: Perform Fourier transform on the oxide layer interference signal to convert the time-domain signal into a frequency-domain signal; S232: Construct the transfer function of the band-stop filter based on the calculated stopband center frequency; S233: Use the constructed filter transfer function to filter the frequency domain signal; S234: Perform an inverse Fourier transform on the frequency domain processed signal to convert it back from the frequency domain to the time domain, restoring the original time domain form of the signal.

7. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 6, characterized in that, S3 includes: S31: Perform multi-scale morphological gradient calculation on the substrate signal layer; S32: Generate pixel-level fusion weights based on the gradient field of defect morphology; S33: The substrate signal layer and the decoupled oxide layer signal are fused according to weights to construct an enhanced flaw detection image; S34: Apply material acoustic impedance boundary constraints to the fused image to limit the grayscale range of the final image and output the final flaw detection image.

8. The intelligent image enhancement method for internal flaw detection of power plant equipment according to claim 8, characterized in that, S33 includes: S331: Based on the obtained fusion weights, the substrate signal image and the oxide layer signal image are weighted and fused; S332: Combine the two types of signals after processing according to the fusion weight to obtain an enhanced flaw detection image with defect significance and interference suppression capability.

9. A smart image enhancement method for internal flaw detection of power plant equipment according to claim 8, characterized in that, S34 includes: S341: Based on the acoustic impedance characteristics of the coupling medium, substrate and oxide layer, calculate the upper and lower boundaries of the grayscale values ​​of the fused image, including the lowest measurable value and the highest confidence value. S342: Intensity limiting is applied to the intermediate fused image to restrict the grayscale values ​​of all pixels within the acoustic impedance range, thus obtaining the final flaw detection image.

Images