Double-rod medium probe nondestructive testing method based on waveguide feed

Abstract

The invention discloses a waveguide feed-based double-rod dielectric probe nondestructive testing method, which comprises the following steps of: constructing a sensing system comprising a specific feed source module, a double-layer dielectric rod and a metal tip, injecting a specific frequency band signal, and performing mode conversion and focusing; enabling a signal to penetrate through a glass window on the surface of the substrate, receiving a reflected signal, and then carrying out phase unwrapping to obtain reference and detection phase data; and calculating a differential phase value and extracting various characteristic parameters, and judging defects and determining positions and sizes of the defects based on a specific threshold criterion. According to the method, signal attenuation is reduced through the design of an exclusive sensing system, the defect recognition accuracy is improved in combination with multi-parameter analysis, and tiny and deep defects can be accurately detected; the system has the advantages of no need of complex pretreatment, compact structure, simplified operation, rapid batch detection, reduction of surface covering layer interference, suitableness for a production line scene, and effective guarantee of the quality of the ceramic substrate.

Description

Technical Field

[0001] This invention relates to the field of ceramic substrate defect detection technology, and in particular to a non-destructive testing method based on a waveguide-fed dual-bar dielectric probe. Background Technology

[0002] In fields such as electronics and new energy, ceramic substrates are widely used in the packaging and support of core components such as power modules and integrated circuits due to their excellent high-temperature resistance, insulation, and thermal conductivity. As these fields continuously raise product performance requirements, the presence of defects such as cracks, voids, and delamination within the ceramic substrate directly affects the operational stability and lifespan of devices. Traditional destructive testing methods damage the substrate, failing to meet the quality screening requirements of mass-produced products. Therefore, non-destructive testing (NDT) technology, capable of accurately identifying internal defects without damaging the substrate, has become a key requirement in the industry for ensuring product quality and reducing production costs. Currently, the industry has placed higher demands on the resolution, detection efficiency, and applicable frequency range of NDT technologies, especially requiring the development of highly adaptable detection solutions that can effectively penetrate the substrate surface and accurately capture internal defect signals, specifically tailored to the dielectric properties of ceramic substrates.

[0003] Existing technologies for non-destructive testing of internal defects in ceramic substrates suffer from two significant drawbacks. Firstly, some testing techniques utilize single-frequency or low-frequency signals, resulting in significant energy attenuation after the signal penetrates the ceramic substrate surface. This makes it difficult to effectively acquire defect information from deeper regions within the substrate, leading to insufficient ability to identify minute or deep defects and an inability to accurately distinguish defect location and size, resulting in low accuracy and reliability of the test results. Secondly, some testing methods rely on complex preprocessing procedures or large-scale testing equipment. These methods are not only cumbersome and inefficient but also require stringent environmental conditions, making them unsuitable for rapid, batch testing of ceramic substrates on production lines. Furthermore, some techniques are susceptible to interference from substrate surface coatings (such as glass windows) during testing, causing signal distortion and further affecting the accuracy of defect identification. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a non-destructive testing method for a waveguide-fed dual-bar dielectric probe.

[0005] The technical solution adopted in this invention is a non-destructive testing method for a waveguide-fed dual-bar dielectric probe, comprising the following steps: S1: Constructing a waveguide-fed dual-bar dielectric sensing system including a WR-90 feed module, a dual-layer dielectric rod, and a metal tip. The WR-90 feed module is used to generate a TE10 mode excitation signal in the 18-24GHz frequency band. The dual-layer dielectric rod is composed of an inner layer of alumina material and an outer layer of polytetrafluoroethylene material, forming a gradient effective refractive index transition region. The metal tip is disposed at the front end of the dual-layer dielectric rod and has a shallow concave structure on its surface; S2: Aligning the metal tip of the sensing system with the glass window covering the ceramic substrate, maintaining a preset distance between the sensing system and the glass window, injecting the TE10 mode signal into the dual-layer dielectric rod through the WR-90 feed module, and converting the signal into a quasi-HE11 mode signal through the gradient effective refractive index transition region of the dual-layer dielectric rod and transmitting it to the metal tip; S3: The metal tip focuses the quasi-HE11 mode signal through the glass window, allowing the signal to act inside the ceramic substrate. Simultaneously, it receives the reflected signal from the ceramic substrate. The reflected signal is transmitted sequentially through the metal tip, the double-layer dielectric rod, and the WR-90 feed module to the signal processing unit. S4: The signal processing unit performs phase unwrapping processing on the reflected signal, acquiring phase reference data of the reflected signal in a defect-free state and phase detection data of the reflected signal in the state to be tested. S5: The differential phase value between the phase detection data and the phase reference data is calculated. The peak value, mean, standard deviation, and energy characteristic parameters of the differential phase value are extracted within the 18-24 GHz frequency band. S6: The extracted characteristic parameters are judged based on the 3σ threshold criterion. When the characteristic parameters exceed the 3σ threshold range, a defect is determined to exist inside the ceramic substrate. Simultaneously, the location and size information of the defect are determined based on the distribution of the characteristic parameters within the frequency band.

[0006] Furthermore, when the double-layer dielectric rod in S2 performs mode switching, the following relationship is satisfied:

[0007] in, The radius of the inner dielectric rod. The radius of the outer dielectric rod, Let be the wavelength of the signal in free space. The relative permittivity of the inner dielectric rod, is the relative permittivity of the outer dielectric rod.

[0008] Furthermore, in S3, when analyzing the signal transmission characteristics through the glass-ceramic stack structure, the following transmission matrix is ​​used:

[0009] in, Let be the transmission matrix of the i-th layer medium. Let be the phase constant of the signal in the i-th layer of the medium. Let be the thickness of the i-th dielectric layer. Let be the characteristic impedance of the i-th dielectric layer. Let be the admittance of the i-th dielectric layer. For free space wavenumber, Let be the relative permittivity of the i-th dielectric layer. This is the free space impedance.

[0010] Furthermore, in step S5, the energy characteristic parameters of the differential phase value are calculated using the following formula:

[0011] in, The energy of the differential phase value, for Number of frequency sampling points within the frequency band This represents the differential phase value corresponding to the i-th frequency sampling point. Let be the frequency value of the i-th frequency sampling point.

[0012] Furthermore, in step S2, the curvature parameters of the shallow recessed structure at the metal tip are determined using the following formula:

[0013] in, Let r be the depth of the concave structure at the radial coordinate r. The desired focal length for the concave structure, The radial coordinates of the edge of the concave structure are: The effective refractive index of the dielectric rod, The phase difference is at the edge of the concave structure. is the free space wavenumber.

[0014] Furthermore, in step S6, when determining the defect detection threshold, the calculation is based on the defect-free baseline residual. The threshold, the formula is as follows:

[0015] in, Standard deviation, This represents the number of frequency sampling points. Let be the differential phase value of the i-th frequency sampling point. The mean of the differential phase values. Let be the frequency value of the i-th frequency sampling point.

[0016] Furthermore, step S2 includes the following sub-steps: S21: Fix the output end of the WR-90 feed module to one end of the double-layer dielectric rod to ensure that there is no signal leakage at the connection. At the same time, adjust the axis of the double-layer dielectric rod to keep it aligned with the signal transmission axis of the WR-90 feed module. S22: Install a metal tip at the other end of the double-layer medium rod, aligning the axis of the metal tip with the axis of the double-layer medium rod, and the shallow concave structure of the metal tip faces the glass window. S23: Start the WR-90 feed module, set the operating frequency band to 18-24GHz, and adjust the output power of the feed module to stabilize the generated TE10 mode signal strength within the preset range; S24: The signal mode at the input end of the double-layer dielectric rod is detected in real time by the signal monitoring equipment to confirm that the purity of the TE10 mode signal meets the preset requirements. If there is noise, the parameters of the feed module are adjusted to suppress the noise.

[0017] Furthermore, step S3 includes the following sub-steps: S31: Adjust the position of the sensing system so that the metal tip and the glass window surface maintain a preset distance, and ensure that the center of the metal tip and the center of the ceramic substrate to be tested are on the same vertical line through the optical positioning device; S32: Control the sensing system to move along a preset path, and continuously emit a focused quasi-HE11 mode signal into the glass window during the movement, while recording the frequency and intensity information of the emitted signal. S33: Receives the reflected signal reflected by the ceramic substrate through the metal tip, transmits the reflected signal to the signal amplification module for signal amplification, and transmits the amplified signal to the filtering module to remove noise; S34: The filtered reflected signal is transmitted to the receiving end of the WR-90 feed module, where the feed module performs preliminary processing on the signal before transmitting it to the signal processing unit.

[0018] Furthermore, step S4 includes the following sub-steps: S41: In the defect-free area of ​​the ceramic substrate, the control sensing system collects multiple sets of reflection signal data, and stores these data as phase reference data in the database of the signal processing unit. S42: For the target area of ​​the ceramic substrate to be tested, according to the same parameter settings as the phase reference data acquisition, acquire multiple sets of reflection signal data as phase detection data; S43: Call the phase untangling algorithm in the signal processing unit to process the phase reference data and phase detection data respectively, and eliminate the tangling phenomenon in the phase data; S44: Align the unwrapped phase reference data and phase detection data to ensure consistency between them in the time and frequency dimensions, laying the foundation for differential calculation.

[0019] Furthermore, step S5 includes the following sub-steps: S51: The differential calculation module in the signal processing unit is used to subtract the aligned phase detection data from the phase reference data at each frequency point to obtain the differential phase value corresponding to each frequency point. S52: Within the 18-24GHz frequency band, scan the differential phase values ​​to determine the maximum value of the differential phase values, and use this maximum value as the peak characteristic parameter; S53: Calculate the arithmetic mean of the differential phase values ​​at all frequency points, use the arithmetic mean as the mean characteristic parameter, and calculate the dispersion of the differential phase values ​​at each frequency point based on the mean to obtain the standard deviation characteristic parameter. S54: By accumulating the squares of the differential phase values ​​at each frequency point through integral operation, the energy characteristic parameters are obtained, and the peak value, mean, standard deviation and energy characteristic parameters are organized into a characteristic parameter set.

[0020] Beneficial Effects: This invention proposes a non-destructive testing method for a waveguide-fed dual-bar dielectric probe. By constructing a specific waveguide-fed dual-bar dielectric sensing system, using a specific frequency band signal and undergoing mode conversion and focusing processing, the energy attenuation of the signal is significantly reduced when penetrating the surface of the ceramic substrate. This effectively acquires information about the deep internal regions of the substrate. Combined with differential phase value calculation and multi-feature parameter extraction, it can accurately identify micro-defects and deep defects, accurately distinguish the location and size of defects, and significantly improve the accuracy and reliability of the detection results. This solves the problem of insufficient detection capability of single or low-frequency signals in existing technologies. This method does not require complex preprocessing procedures, the sensing system used has a compact structure, and the operation steps are simplified. It can move along a preset path to achieve rapid detection and can adapt to batch detection scenarios on production lines. At the same time, through specific signal processing and metal tip structure design, the interference of the substrate surface coating layer on the detection signal is reduced, avoiding signal distortion. This not only improves detection efficiency but also reduces the requirements for the detection environment, overcoming the shortcomings of existing technologies that rely on complex processes, large equipment, and susceptibility to interference. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating the overall method steps of the present invention; Figure 2 This is a flowchart of method step S2 of the present invention; Figure 3 This is a flowchart of method step S3 of the present invention; Figure 4 This is a flowchart of method step S4 of the present invention; Figure 5 This is a flowchart of step S5 of the method of the present invention. Detailed Implementation

[0022] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] like Figure 1 As shown, a non-destructive testing method for a waveguide-fed dual-bar dielectric probe includes the following steps: S1: Construct a waveguide-fed dual-bar dielectric sensing system including a WR-90 feed module, a double-layer dielectric rod, and a metal tip. The WR-90 feed module is used to generate the TE10 mode excitation signal in the 18-24GHz frequency band. The double-layer dielectric rod is composed of an inner layer of alumina material and an outer layer of polytetrafluoroethylene material, forming a gradient effective refractive index transition region. The metal tip is located at the front end of the double-layer dielectric rod and has a shallow concave structure on its surface. Specifically, in step S1, a waveguide-fed dual-bar dielectric sensing system needs to be built. This system includes a WR-90 feed module, a double-layer dielectric rod, and a metal tip. The WR-90 feed module needs to be set to generate a TE10 mode excitation signal in the 18 to 24 GHz frequency band. This frequency band signal can balance penetration depth and detection resolution when penetrating ceramic substrates, making it suitable for most ceramic substrate detection scenarios. The inner layer of the double-layer dielectric rod is made of alumina, and the outer layer is made of polytetrafluoroethylene. The combination of the two forms a gradient effective refractive index transition region, which can reduce signal transmission loss. The inner layer radius is controlled between 1.2 and 1.8 mm, and the outer layer radius is controlled between 3.2 and 3.8 mm. The specific dimensions need to be adjusted according to the thickness of the ceramic substrate to be detected. The metal tip is installed at the front end of the double-layer dielectric rod, and a shallow recessed structure is processed on the surface. The recess depth is set to 0.6 to 0.9 mm, and the recess diameter is consistent with the outer radius of the double-layer dielectric rod. This structure can enhance the signal focusing effect. This step is fundamental to the testing process. By accurately selecting materials and setting parameters, it ensures that the sensing system can stably generate and transmit signals that meet the testing requirements, providing hardware support for subsequent testing procedures.

[0024] S2: Align the metal tip of the sensing system with the glass window covering the ceramic substrate, maintain a preset distance between the sensing system and the glass window, and inject a TE10 mode signal into the double-layer dielectric rod through the WR-90 feed module. The signal is converted into a quasi-HE11 mode signal through the gradient effective refractive index transition zone of the double-layer dielectric rod and transmitted to the metal tip. Specifically, in step S2, the metal tip of the sensing system is first aligned with the glass window covering the ceramic substrate surface, and the distance between them is adjusted to 0.35 to 0.45 mm. This distance avoids direct contact between the metal tip and the glass window, preventing scratches, and also reduces signal attenuation in the air layer. Then, the WR-90 feed module is activated, injecting the pre-set TE10 mode signal into the double-layer dielectric rod. After entering the double-layer dielectric rod, the signal is gradually converted to a quasi-HE11 mode signal through a gradient effective refractive index transition zone. During the conversion process, real-time monitoring with a dedicated signal monitoring device is required to ensure that the purity of the converted quasi-HE11 mode signal is not less than 96% and that there is no significant noise interference. After the signal conversion stabilizes, the quasi-HE11 mode signal is transmitted along the double-layer dielectric rod to the metal tip. This step achieves signal mode conversion and precise positioning, preparing the signal for subsequent penetration through the glass window and action inside the ceramic substrate, directly affecting the quality of the subsequent detection signal and the accuracy of the detection results.

[0025] S3: After the metal tip focuses the HE11 mode signal, it penetrates the glass window, so that the signal acts inside the ceramic substrate. At the same time, it receives the reflected signal reflected back by the ceramic substrate. The reflected signal is transmitted to the signal processing unit in sequence through the metal tip, the double-layer dielectric rod and the WR-90 feed module. Specifically, in step S3, the metal tip focuses the transmitted quasi-HE11 mode signal, increasing the signal energy density by 3.5 to 4.5 times. This effectively penetrates glass windows with a thickness of 0.9 to 1.1 mm, while keeping energy loss during penetration below 9%. The focused signal then acts on the ceramic substrate. If defects such as cracks or voids exist within the substrate, these defects alter the signal propagation path, causing changes in signal reflection characteristics. The metal tip simultaneously receives the reflected signal from the ceramic substrate. The reflected signal is first transmitted to the double-layer dielectric rod and then returns along its original path. During transmission, the gradient structure of the double-layer dielectric rod further reduces losses, ensuring signal loss does not exceed 14%. Subsequently, the reflected signal enters the WR-90 feed module, which performs preliminary filtering to remove high-frequency interference noise before transmitting the processed reflected signal to the signal processing unit. This step completes the signal transmission, penetration, and reception, obtaining raw signal data reflecting the internal state of the ceramic substrate, which is a prerequisite for subsequent data processing and defect determination.

[0026] S4: The signal processing unit performs phase unwrapping processing on the reflected signal to obtain the phase reference data of the reflected signal in the defect-free state of the ceramic substrate and the phase detection data of the reflected signal in the state to be detected. Specifically, in step S4, the signal processing unit first performs phase unwrapping processing on the received reflected signal using the phase gradient method. During this process, a phase difference threshold of 0.1 radians is set to avoid data distortion caused by phase entanglement. First, a defect-free area of ​​the ceramic substrate is selected, and the sensing system continuously collects 25 to 28 sets of reflected signal data in this area. These data are processed using an arithmetic average method to obtain phase reference data, which is stored in the signal processing unit's dedicated database as a benchmark for subsequent comparative analysis. Then, the target detection area of ​​the ceramic substrate to be tested is subjected to the same parameter settings as the phase reference data collection, and 25 to 28 sets of reflected signal data are continuously collected. After the same arithmetic averaging process, phase detection data is obtained. During the acquisition process, the sampling frequency of both the phase reference data and the phase detection data is set to 1000 Hz to ensure consistency in the time dimension between the two sets of data. This step, through standardized data acquisition and processing, yields phase data suitable for comparison, laying the foundation for subsequent calculation of differential phase values ​​and extraction of feature parameters.

[0027] S5: Calculate the differential phase value between the phase detection data and the phase reference data, and extract the peak value, mean value, standard deviation and energy characteristic parameters of the differential phase value in the 18-24GHz frequency band; Specifically, in step S5, the phase detection data and the phase reference data are first aligned to ensure that the alignment error between the two sets of data at frequency and time points does not exceed 0.008 seconds. After alignment, the differential phase values ​​at corresponding points of the two sets of data are calculated by subtracting them point by point. Differential phase values ​​are extracted in the 18 to 24 GHz frequency band at 100 MHz frequency intervals, resulting in a total of 61 differential phase data points. Based on these differential phase data, the peak value, mean, standard deviation, and energy characteristic parameters are calculated. The peak value is the maximum value among the 61 differential phase values. The mean value is obtained by summing the differential phase values ​​at the 61 frequency points and dividing by 61. The standard deviation is obtained by summing the squares of the differences between the differential phase values ​​at each frequency point and the mean value, then averaging and taking the square root. The energy characteristic parameters are obtained by summing the squares of the differential phase values ​​at each frequency point. After the calculation is completed, these characteristic parameters are organized into a characteristic parameter set, and the precision of all parameters is retained to three decimal places. This step uses quantitative calculations to transform the raw phase data into specific indicators that can be used for defect determination, providing data support for accurately determining whether there are defects inside the ceramic substrate.

[0028] S6: The extracted feature parameters are judged based on the 3σ threshold criterion. When the feature parameters exceed the 3σ threshold range, it is determined that there are defects inside the ceramic substrate. At the same time, the location and size information of the defects are determined according to the distribution of the feature parameters in the frequency band.

[0029] Specifically, in step S6, the defect-free ceramic substrate feature parameter data stored in the signal processing unit is first called to calculate the mean and standard deviation of each feature parameter in the defect-free state. Then, according to the 3σ threshold criterion, the judgment threshold range of each feature parameter is calculated, i.e., the mean minus 3 times the standard deviation to the mean plus 3 times the standard deviation. The feature parameters of the ceramic substrate to be tested obtained in step S5 are compared with the corresponding 3σ threshold range. If any feature parameter exceeds the range, it is determined that there is a defect inside the ceramic substrate. At the same time, according to the feature parameters at 18... The defect location is determined by analyzing the distribution of features within the 24 GHz frequency band. If multiple characteristic parameters exhibit abnormal values ​​within a certain frequency band, and these abnormal values ​​are concentrated in one area, then the ceramic substrate region corresponding to that frequency band is the location of the defect. The defect size is determined by the magnitude of the abnormal value of the characteristic parameter. When the peak value exceeds the threshold by more than twice, the defect size is determined to be greater than 0.5 mm; when the peak value exceeds the threshold by one to two times, the defect size is determined to be 0.2 to 0.5 mm; and when the peak value exceeds the threshold by less than one time, the defect size is determined to be less than 0.2 mm. This step completes the defect identification, location, and size estimation, achieving the goal of non-destructive testing of internal defects in the ceramic substrate.

[0030] Preferably, when the double-layer dielectric rod in S2 performs mode switching, the following relationship is satisfied:

[0031] in, The radius of the inner dielectric rod. The radius of the outer dielectric rod, Let be the wavelength of the signal in free space. The relative permittivity of the inner dielectric rod, is the relative permittivity of the outer dielectric rod.

[0032] Specifically, in step S2, the dual-layer dielectric rod achieves mode switching by controlling the dimensional parameters of the inner and outer dielectric rods to ensure stable signal mode switching. During implementation, the radius of the inner dielectric rod needs to be controlled between 1.2 and 1.8 mm, and the radius of the outer dielectric rod needs to be controlled between 3.2 and 3.8 mm. The selected inner alumina material has a relative permittivity of 9.8, and the outer polytetrafluoroethylene material has a relative permittivity of 2.1. The wavelength of the signal in free space is calculated based on the 18 to 24 GHz frequency band, corresponding to a wavelength range of 12.5 to 16.7 mm. By setting these parameters, the inner dielectric rod meets specific size and dielectric constant matching relationships, ensuring stable signal transmission without mode distortion. Simultaneously, the outer dielectric rod size must be controlled within a specific range to prevent signal leakage. Clearly defining the parameter boundaries for mode conversion, and precisely setting the radii and dielectric constants of the inner and outer dielectric rods, ensures efficient conversion of the TE10 mode signal to a quasi-HE11 mode signal. The purity of the converted signal can reach over 96%, providing a stable signal source for subsequent signal focusing and penetration through the glass window. This directly affects the accuracy of subsequent detection, avoiding problems such as excessive signal loss or increased noise due to unstable mode conversion.

[0033] Preferably, in step S3, when analyzing the signal transmission characteristics through the glass-ceramic stack structure, the following transmission matrix is ​​used:

[0034] in, Let be the transmission matrix of the i-th layer medium. Let be the phase constant of the signal in the i-th layer of the medium. Let be the thickness of the i-th dielectric layer. Let be the characteristic impedance of the i-th dielectric layer. Let be the admittance of the i-th dielectric layer. For free space wavenumber, Let be the relative permittivity of the i-th dielectric layer. This is the free space impedance.

[0035] Specifically, step S3 analyzes the signal transmission characteristics through the glass-ceramic stack structure by constructing a transmission matrix to quantify the influence of each dielectric layer on the signal. In implementation, the thickness parameters of the glass window and ceramic substrate must first be determined. The glass window thickness is set to 0.9 to 1.1 mm, and the ceramic substrate thickness is set to 0.5 to 2 mm based on the detection requirements. The relative permittivity of the glass material is 6.8, and the relative permittivity of the ceramic substrate is 9.5. The free-space wavenumber is calculated based on the 18 to 24 GHz frequency band, and the free-space impedance is a fixed standard value. When calculating the transmission matrix of each dielectric layer, the phase constant, characteristic impedance, and admittance of each layer must be calculated separately. The phase constant is determined by the free-space wavenumber and the corresponding relative permittivity of the dielectric. The characteristic impedance is the ratio of the free-space impedance to the square root of the relative permittivity of the dielectric, and the admittance is the reciprocal of the characteristic impedance. The transmission matrix constructed using these parameters allows for precise calculation of signal transmission loss and phase change in each dielectric layer. The signal loss through the glass window is controlled within 9%, and the loss through the ceramic substrate is controlled within 15%. This provides a quantitative method for analyzing transmission characteristics. By clearly defining the parameter settings and calculation logic for each dielectric layer, it ensures accurate evaluation of the signal transmission state in the stacked structure. This provides data support for subsequent adjustments to signal parameters and optimization of detection results, avoiding problems such as insufficient signal strength or excessive phase shift due to the inability to quantify transmission loss.

[0036] Preferably, the following formula is used when calculating the energy characteristic parameters of the differential phase value in step S5:

[0037] in, The energy of the differential phase value, for Number of frequency sampling points within the frequency band This represents the differential phase value corresponding to the i-th frequency sampling point. Let be the frequency value of the i-th frequency sampling point.

[0038] Specifically, in step S5, the process of calculating the differential phase value energy characteristic parameter quantifies the signal energy by accumulating the squares of the differential phase values ​​at each frequency point. In practice, the number of frequency sampling points is set to 61, covering the 18 to 24 GHz frequency band, with a frequency interval of 100 MHz between each sampling point. After collecting phase detection data and phase reference data, the two sets of data are subtracted point by point to obtain the differential phase value corresponding to each frequency point. The precision of the differential phase value is retained to three decimal places. When calculating the energy characteristic parameter, the differential phase value at each frequency point is first squared, and then all squared results are accumulated. During the accumulation process, it must be ensured that data from each frequency point is included in the calculation without omission or duplication. For example, when the differential phase value at a certain frequency is 0.3 radians, its square value is 0.09. The energy characteristic parameter obtained by accumulating the square values ​​of 61 frequency points can reflect the overall strength of the differential phase signal. By clarifying the calculation logic and parameter setting of the energy characteristic parameter, and by fixing the number of sampling points and intervals at a fixed frequency, the comparability of energy parameters under different detection scenarios can be ensured. The numerical range of the energy parameter is usually between 5 and 20. When there are defects in the ceramic substrate, the energy parameter will exceed the 3σ threshold in the defect-free state (the mean is 8, the standard deviation is 2, and the 3σ threshold range is 2 to 14 in the defect-free state). Based on this, it can help determine whether there are defects and improve the reliability of defect determination.

[0039] Preferably, the curvature parameter of the shallow recessed structure at the metal tip in step S2 is determined using the following formula:

[0040] in, Let r be the depth of the concave structure at the radial coordinate r. The desired focal length for the concave structure, The radial coordinates of the edge of the concave structure are: The effective refractive index of the dielectric rod, The phase difference is at the edge of the concave structure. is the free space wavenumber.

[0041] Specifically, in step S2, the required curvature parameters of the shallow recessed structure at the metal tip are determined. By setting parameters such as the focal length and edge radial coordinates of the recessed structure, the signal focusing effect is optimized. During implementation, the desired focal length of the recessed structure is set to 2 to 3 mm. This focal length must match the distance from the metal tip to the glass window (0.35 to 0.45 mm) to ensure that the focused signal can accurately act on the surface of the ceramic substrate. The edge radial coordinates of the recessed structure are consistent with the radius of the outer dielectric rod, controlled within 3.2 to 3.8 mm. The effective refractive index of the dielectric rod is calculated based on the dielectric constants and dimensions of the inner and outer dielectric layers, with a value range of 1.8 to 2.2. The free-space wavenumber is determined based on the 18 to 24 GHz frequency band. When calculating the recess depth, the desired focal length, edge radial coordinates, and effective refractive index must be considered to ensure that the curved surface of the recess structure can focus the quasi-HE11 mode signal onto the designated area. The signal energy density after focusing can be increased by 3.5 to 4.5 times. At the same time, the phase difference at the edge of the recess structure must be calculated. The phase difference must be controlled between 0.5 and 1 radian to avoid signal phase distortion due to excessive phase difference. The parameter design method of the recess structure must be clarified. By accurately setting parameters such as focal length and edge radial coordinates, the metal tip can efficiently focus the signal. The diameter of the focused signal spot is controlled between 0.5 and 1 mm to ensure that the signal can accurately act on the detection area of ​​the ceramic substrate, improve the ability to identify small defects, and avoid problems such as excessive signal coverage or energy dispersion due to poor focusing effect.

[0042] Preferably, when determining the defect detection threshold in S6, it is calculated based on the defect-free baseline residual. The threshold, the formula is as follows:

[0043] in, Standard deviation, This represents the number of frequency sampling points. Let be the differential phase value of the i-th frequency sampling point. The mean of the differential phase values. Let be the frequency value of the i-th frequency sampling point.

[0044] Specifically, the technical solution for determining the defect detection threshold in step S6 involves calculating the 3σ threshold based on the defect-free baseline residual, providing a quantitative standard for defect judgment. During implementation, the number of frequency sampling points is set to 61, covering the 18-24 GHz frequency band. First, multiple sets of reflection signal data from the defect-free ceramic substrate are collected, typically 25 to 28 sets. The differential phase value at each frequency point is calculated (in the defect-free state, the differential phase value is close to 0, with fluctuations between -0.3 and 0.3 radians). Then, the mean and standard deviation of these differential phase values ​​are calculated. The mean is typically around 0.02 radians, and the standard deviation is around 0.08 radians. When calculating the standard deviation, the difference between the differential phase value and the mean at each frequency point is squared first, then the average of all squared values ​​is calculated, and finally the square root is taken to obtain the standard deviation. During the calculation process, it is necessary to ensure that the data is free of outlier interference. If outliers exceeding ±0.5 radians exist, the data needs to be recollected. The 3σ threshold is the mean plus or minus three standard deviations, typically ranging from -0.22 to 0.26 radians. When the differential phase value of the ceramic substrate to be tested exceeds this range, a defect is determined to exist. The specific process and parameter settings for threshold calculation are clearly defined. By fixing the number of sampling points and data acquisition groups at a fixed frequency, the stability and reliability of the threshold are ensured, avoiding missed or false detections due to unreasonable threshold settings. The threshold range under defect-free conditions has been verified by a large number of experiments and can cover 99.7% of defect-free situations, providing a scientific quantitative basis for defect judgment.

[0045] Preferred, such as Figure 2 As shown, step S2 includes the following sub-steps: S21: Fix the output end of the WR-90 feed module to one end of the double-layer dielectric rod to ensure that there is no signal leakage at the connection. At the same time, adjust the axis of the double-layer dielectric rod to keep it aligned with the signal transmission axis of the WR-90 feed module. S22: Install a metal tip at the other end of the double-layer medium rod, aligning the axis of the metal tip with the axis of the double-layer medium rod, and the shallow concave structure of the metal tip faces the glass window. S23: Start the WR-90 feed module, set the operating frequency band to 18-24GHz, and adjust the output power of the feed module to stabilize the generated TE10 mode signal strength within the preset range; S24: The signal mode at the input end of the double-layer dielectric rod is detected in real time by the signal monitoring equipment to confirm that the purity of the TE10 mode signal meets the preset requirements. If there is noise, the parameters of the feed module are adjusted to suppress the noise.

[0046] Specifically, step S2 ensures stable signal mode conversion and meets signal quality standards through step-by-step operations. During implementation, S21 first fixes the output end of the WR-90 feed module to one end of the double-layer dielectric rod, using a metal sealing structure at the connection point, with the sealing gap controlled between 0.05 and 0.1 mm. Simultaneously, a laser positioning device is used to adjust the alignment of the two axes, ensuring a deviation of no more than 0.02 mm to prevent signal leakage. S22 installs a metal tip on the other end of the double-layer dielectric rod, using a threaded connection, with the tightening torque set to 5 to 8 Newtons. For step S23, ensure a secure connection and axis alignment, with the shallow recessed structure of the metal tip facing the glass window. The center of the recess should deviate from the axis of the dielectric rod by no more than 0.03 mm. For step S24, start the feed module, set the operating frequency band to 18 to 24 GHz, and adjust the output power to 10 to 15 dBmW. Observe the power stability using a power monitor, and control the fluctuation range within ±0.5 dBmW. For step S24, use a signal analyzer to detect the signal mode at the input end of the dielectric rod. The purity of the TE10 mode must reach above 96%. If the noise ratio exceeds 4%, adjust the excitation current of the feed module, with the current adjustment range being 50 to 80 mA, until the noise ratio is below 4%. Through standardized step-by-step operation, the parameter requirements of each link are clearly defined to avoid the signal conversion effect being affected by connection deviation, unstable power, or impure mode. This provides a stable foundation for subsequent signal focusing and penetration of the glass window, ensuring that each step of the operation is traceable and controllable, and improving the repeatability and reliability of the detection process.

[0047] Preferred, such as Figure 3 As shown, step S3 includes the following sub-steps: S31: Adjust the position of the sensing system so that the metal tip and the glass window surface maintain a preset distance, and ensure that the center of the metal tip and the center of the ceramic substrate to be tested are on the same vertical line through the optical positioning device; S32: Control the sensing system to move along a preset path, and continuously emit a focused quasi-HE11 mode signal into the glass window during the movement, while recording the frequency and intensity information of the emitted signal. S33: Receives the reflected signal reflected by the ceramic substrate through the metal tip, transmits the reflected signal to the signal amplification module for signal amplification, and transmits the amplified signal to the filtering module to remove noise; S34: The filtered reflected signal is transmitted to the receiving end of the WR-90 feed module, where the feed module performs preliminary processing on the signal before transmitting it to the signal processing unit.

[0048] Specifically, step S3 achieves precise signal transmission, reception, and preliminary processing through step-by-step steps to ensure the quality of the reflected signal. During implementation, S31 uses a three-dimensional displacement platform to adjust the position of the sensing system, maintaining a distance of 0.35 to 0.45 mm between the metal tip and the glass window. The distance deviation is observed using an optical microscope and controlled within ±0.02 mm. Simultaneously, a crosshair is used to ensure that the vertical deviation between the center of the metal tip and the center of the ceramic substrate does not exceed 0.03 mm. S32 sets the sensing system's movement path to a rectangle, with the path side length determined according to the substrate size. The movement speed is controlled at 5 to 8 mm per second. During movement, a signal generator maintains stable signal transmission in quasi-HE11 mode, recording the transmission signal frequency (intervals of 100 MHz) and intensity (fluctuations within ±0.3 dB / mW) in real time. S33, after the metal tip receives the reflected signal, it transmits it to a low-frequency... The noise amplification module has an amplification gain set to 20 to 25 dB and a noise figure controlled to 1.5 to 2 dB. The signal is then transmitted to a bandpass filter with a filtering frequency band of 18 to 24 GHz and a stopband attenuation greater than 60 dB to remove noise outside the frequency band. The reflected signal after S34 filtering is transmitted to the feed module receiver, where the feed module performs amplitude calibration with a calibration error not exceeding 0.5 dB. The signal is then transmitted to the signal processing unit, which refines the parameters of each stage of signal transmission, reception, and preprocessing. By precisely controlling the spacing, moving speed, and signal processing parameters, external interference and signal loss are reduced, ensuring that the reflected signal truly reflects the internal state of the substrate. This provides a high-quality original signal for subsequent phase unwrapping and data comparison, avoiding signal distortion due to operational deviations.

[0049] Preferred, such as Figure 4 As shown, step S4 includes the following sub-steps: S41: In the defect-free area of ​​the ceramic substrate, the control sensing system collects multiple sets of reflection signal data, and stores these data as phase reference data in the database of the signal processing unit. S42: For the target area of ​​the ceramic substrate to be tested, according to the same parameter settings as the phase reference data acquisition, acquire multiple sets of reflection signal data as phase detection data; S43: Call the phase untangling algorithm in the signal processing unit to process the phase reference data and phase detection data respectively, and eliminate the tangling phenomenon in the phase data; S44: Align the unwrapped phase reference data and phase detection data to ensure consistency between them in the time and frequency dimensions, laying the foundation for differential calculation.

[0050] Specifically, the phased data processing flow in step S4 ensures the consistency and accuracy of the phase reference data and the detection data through standardized data acquisition and processing. During implementation, S41 selects a defect-free area at the edge of the ceramic substrate (area not less than 5×5 square millimeters), controls the sensing system to acquire data at a speed of 3 millimeters per second, continuously acquiring 25 to 28 sets of reflection signals. Each set of data acquisition lasts for 100 milliseconds, and the sampling rate is set to 1000 Hz. After acquisition, outliers (data with deviations exceeding 3 times the mean) are removed through data filtering, and the average value is then taken as the phase reference data and stored in the database in binary format with a data compression rate controlled at 1.2 to 1.5 times. In step S42, the area to be detected maintains the same acquisition speed, number of sets, duration, and sampling rate as in S41, acquiring 25 to 28 sets of reflection signals. Outlier removal and averaging are also performed to obtain the phase detection data. 43. The phase unwrapping algorithm in the signal processing unit is invoked, with the unwrapping window size set to 5×5 pixels and the number of iterations controlled between 10 and 15. After unwrapping, the phase data error does not exceed 0.05 radians, eliminating the entanglement phenomenon. S44. The timestamp alignment method is used to align the unwrapped reference data and the detection data in the time dimension, with an alignment error not exceeding 0.008 seconds. At the same time, the corresponding frequency point data is matched at 100 MHz intervals in the frequency dimension. By standardizing the data acquisition, processing and alignment process step by step, the parameter standards of each link are clarified, avoiding the influence of inconsistent acquisition conditions, data abnormalities or alignment deviations on the differential phase calculation results, ensuring that the two sets of data are comparable, and laying a reliable data foundation for subsequent feature parameter extraction and defect judgment.

[0051] Preferred, such as Figure 5 As shown, step S5 includes the following sub-steps: S51: The differential calculation module in the signal processing unit is used to subtract the aligned phase detection data from the phase reference data at each frequency point to obtain the differential phase value corresponding to each frequency point. S52: Within the 18-24GHz frequency band, scan the differential phase values ​​to determine the maximum value of the differential phase values, and use this maximum value as the peak characteristic parameter; S53: Calculate the arithmetic mean of the differential phase values ​​at all frequency points, use the arithmetic mean as the mean characteristic parameter, and calculate the dispersion of the differential phase values ​​at each frequency point based on the mean to obtain the standard deviation characteristic parameter. S54: By accumulating the squares of the differential phase values ​​at each frequency point through integral operation, the energy characteristic parameters are obtained, and the peak value, mean, standard deviation and energy characteristic parameters are organized into a characteristic parameter set.

[0052] Specifically, step S5 completes the differential phase value calculation and feature parameter extraction step by step, providing quantitative indicators for subsequent defect judgment. During implementation, S51 calls the differential calculation module in the signal processing unit, first reading the aligned phase detection data and phase reference data. Both sets of data include phase information from 61 frequency points within the 18-24 GHz band, with a sampling interval of 100 MHz. During the calculation, subtraction is performed one frequency point at a time; that is, the phase detection data at each frequency point is subtracted from the corresponding phase reference data to obtain the differential phase value for each frequency point. The calculation accuracy is retained to three decimal places. Simultaneously, a data verification function ensures no frequency points are missed or calculation errors occur; if data is missing, it is retrieved again. The original data is supplemented with calculations; S52 performs point-by-point scanning of the differential phase values ​​within the 18-24 GHz frequency band, selecting the point with the largest value from 61 frequency points as the peak characteristic parameter. During the scanning process, the frequency points corresponding to the peak values ​​are recorded to facilitate subsequent analysis of defect locations based on frequency band distribution. The reading error of the peak parameter is controlled within ±0.01 radians; S53 sums the differential phase values ​​of the 61 frequency points and divides by 61 to obtain the mean characteristic parameter. After the mean is calculated, the difference between the differential phase value and the mean is calculated for each frequency point, and the squared difference is then calculated. The variance is obtained by dividing the sum by (61-1). The square root of the variance is the standard deviation characteristic parameter. The precision of both the mean and the standard deviation is retained to three decimal places. During the calculation, outlier interference must be eliminated. If the differential phase value at a certain frequency point exceeds the mean ± 3 times, it is judged as an outlier and removed before recalculation. S54 performs a square operation on the differential phase value at each frequency point, and then sums all the squared results to obtain the energy characteristic parameter. During the summation process, it must be ensured that the squared value of each frequency point is accurately included without duplication or omission. The precision of the energy parameter is retained to one decimal place. Finally, the peak value, The mean, standard deviation, and energy characteristic parameters are organized into a characteristic parameter set in frequency order and stored in the temporary database of the signal processing unit. This prepares the data for defect determination in step S6. Through standardized step-by-step operations, the calculation methods and accuracy requirements of each characteristic parameter are clarified, ensuring the accuracy and consistency of parameter extraction and avoiding distortion of characteristic parameters due to non-standard calculation processes. This provides a reliable quantitative basis for subsequent defect determination based on the 3σ threshold criterion. At the same time, by recording the frequency points corresponding to the peak values, the foundation for determining the defect location is laid, improving the scientific nature and repeatability of the entire detection process.

[0053] A non-destructive testing method based on a waveguide-fed dual-bar dielectric probe is proposed. This method utilizes a dedicated waveguide-fed dual-bar dielectric sensing system, coupled with a specific frequency band signal. Through mode conversion and focusing processing, the energy attenuation of the signal penetrating the surface of the ceramic substrate is significantly reduced, ensuring the acquisition of defect information in deep regions within the substrate. Furthermore, by employing differential phase value calculation and extraction of multiple feature parameters, the method can accurately identify both minute and deep defects, and precisely distinguish the location and size of defects. This significantly improves the accuracy and reliability of the detection results, solving the problem of insufficient identification capability for minute and deep defects in existing technologies that rely on single or low-frequency signals.

[0054] This method boasts significant advantages in detection efficiency and scenario adaptability. It eliminates the need for complex preprocessing, employs a compact sensing system, simplifies operation, and enables rapid detection along a pre-defined path, making it suitable for batch inspection of ceramic substrates in production lines. Furthermore, through specific signal processing methods and a metal tip structure design, interference from the substrate surface coating on the detection signal is reduced, preventing signal distortion. This not only improves detection efficiency but also lowers environmental requirements, overcoming the shortcomings of existing technologies that rely on complex processes, bulky equipment, and susceptibility to surface coating interference during detection.

[0055] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A non-destructive testing method for a waveguide-fed dual-rod dielectric probe, characterized in that, Includes the following steps: S1: Construct a waveguide-fed dual-bar dielectric sensing system including a WR-90 feed module, a double-layer dielectric rod, and a metal tip. The WR-90 feed module is used to generate the TE10 mode excitation signal in the 18-24GHz frequency band. The double-layer dielectric rod is composed of an inner layer of alumina material and an outer layer of polytetrafluoroethylene material, forming a gradient effective refractive index transition region. The metal tip is located at the front end of the double-layer dielectric rod and has a shallow concave structure on its surface. S2: Align the metal tip of the sensing system with the glass window covering the ceramic substrate, maintain a preset distance between the sensing system and the glass window, and inject a TE10 mode signal into the double-layer dielectric rod through the WR-90 feed module. The signal is converted into a quasi-HE11 mode signal through the gradient effective refractive index transition zone of the double-layer dielectric rod and transmitted to the metal tip. S3: After the metal tip focuses the HE11 mode signal, it penetrates the glass window, so that the signal acts inside the ceramic substrate. At the same time, it receives the reflected signal reflected back by the ceramic substrate. The reflected signal is transmitted to the signal processing unit in sequence through the metal tip, the double-layer dielectric rod and the WR-90 feed module. S4: The signal processing unit performs phase unwrapping processing on the reflected signal to obtain the phase reference data of the reflected signal in the defect-free state of the ceramic substrate and the phase detection data of the reflected signal in the state to be detected. S5: Calculate the differential phase value between the phase detection data and the phase reference data, and extract the peak value, mean value, standard deviation and energy characteristic parameters of the differential phase value in the 18-24GHz frequency band; S6: The extracted feature parameters are judged based on the 3σ threshold criterion. When the feature parameters exceed the 3σ threshold range, it is determined that there are defects inside the ceramic substrate. At the same time, the location and size information of the defects are determined according to the distribution of the feature parameters in the frequency band.

2. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, When the double-layer dielectric rod in S2 performs mode switching, the following relationship is satisfied: in, The radius of the inner dielectric rod. The radius of the outer dielectric rod, Let be the wavelength of the signal in free space. The relative permittivity of the inner dielectric rod, is the relative permittivity of the outer dielectric rod.

3. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, When analyzing the signal transmission characteristics through the glass-ceramic stack structure in S3, the following transmission matrix is ​​used: in, Let be the transmission matrix of the i-th layer medium. Let be the phase constant of the signal in the i-th layer of the medium. Let be the thickness of the i-th dielectric layer. Let be the characteristic impedance of the i-th dielectric layer. Let be the admittance of the i-th dielectric layer. For free space wavenumber, Let be the relative permittivity of the i-th dielectric layer. This is the free space impedance.

4. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, The following formula is used when calculating the energy characteristic parameters of the differential phase value in S5: in, The energy of the differential phase value, for Number of frequency sampling points within the frequency band This represents the differential phase value corresponding to the i-th frequency sampling point. Let be the frequency value of the i-th frequency sampling point.

5. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, When determining the curvature parameters of the shallow concave structure at the metal tip in S2, the following formula is used: in, Let r be the depth of the concave structure at the radial coordinate r. The desired focal length for the concave structure, The radial coordinates of the edge of the concave structure are: The effective refractive index of the dielectric rod, The phase difference is at the edge of the concave structure. is the free space wavenumber.

6. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, When determining the defect detection threshold in step S6, it is calculated based on the defect-free baseline residual. The threshold, the formula is as follows: in, Standard deviation, This represents the number of frequency sampling points. Let be the differential phase value of the i-th frequency sampling point. The mean of the differential phase values. Let be the frequency value of the i-th frequency sampling point.

7. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, S2 includes the following steps: S21: Fix the output end of the WR-90 feed module to one end of the double-layer dielectric rod to ensure that there is no signal leakage at the connection. At the same time, adjust the axis of the double-layer dielectric rod to keep it aligned with the signal transmission axis of the WR-90 feed module. S22: Install a metal tip at the other end of the double-layer medium rod, aligning the axis of the metal tip with the axis of the double-layer medium rod, and the shallow concave structure of the metal tip faces the glass window. S23: Start the WR-90 feed module, set the operating frequency band to 18-24GHz, and adjust the output power of the feed module to stabilize the generated TE10 mode signal strength within the preset range; S24: The signal mode at the input end of the double-layer dielectric rod is detected in real time by the signal monitoring equipment to confirm that the purity of the TE10 mode signal meets the preset requirements. If there is noise, the parameters of the feed module are adjusted to suppress the noise.

8. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, S3 includes the following steps: S31: Adjust the position of the sensing system so that the metal tip and the glass window surface maintain a preset distance, and ensure that the center of the metal tip and the center of the ceramic substrate to be tested are on the same vertical line through the optical positioning device. S32: Control the sensing system to move along a preset path, and continuously emit a focused quasi-HE11 mode signal into the glass window during the movement, while recording the frequency and intensity information of the emitted signal. S33: Receives the reflected signal reflected by the ceramic substrate through the metal tip, transmits the reflected signal to the signal amplification module for signal amplification, and transmits the amplified signal to the filtering module to remove noise; S34: The filtered reflected signal is transmitted to the receiving end of the WR-90 feed module, where the feed module performs preliminary processing on the signal before transmitting it to the signal processing unit.

9. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, S4 includes the following steps: S41: In the defect-free area of ​​the ceramic substrate, the control sensing system collects multiple sets of reflection signal data, and stores these data as phase reference data in the database of the signal processing unit. S42: For the target area of ​​the ceramic substrate to be tested, according to the same parameter settings as the phase reference data acquisition, acquire multiple sets of reflection signal data as phase detection data; S43: Call the phase untangling algorithm in the signal processing unit to process the phase reference data and phase detection data respectively, and eliminate the tangling phenomenon in the phase data; S44: Align the unwrapped phase reference data and phase detection data to ensure consistency between them in the time and frequency dimensions, laying the foundation for differential calculation.

10. The non-destructive testing method for a waveguide-fed dual-rod dielectric probe according to claim 1, characterized in that, S5 includes the following steps: S51: The differential calculation module in the signal processing unit is used to subtract the aligned phase detection data from the phase reference data at each frequency point to obtain the differential phase value corresponding to each frequency point. S52: Within the 18-24GHz frequency band, scan the differential phase values ​​to determine the maximum value of the differential phase values, and use this maximum value as the peak characteristic parameter; S53: Calculate the arithmetic mean of the differential phase values ​​at all frequency points, use the arithmetic mean as the mean characteristic parameter, and calculate the dispersion of the differential phase values ​​at each frequency point based on the mean to obtain the standard deviation characteristic parameter. S54: By accumulating the squares of the differential phase values ​​at each frequency point through integration, the energy characteristic parameters are obtained, and the peak value, mean value, standard deviation, and energy characteristic parameters are organized into a set of characteristic parameters.

Images