Differential value analysis method for microwave detection result evaluation

Abstract

A difference analysis method for microwave detection result evaluation comprises the following steps: carrying out microwave detection on a defect-free sample to obtain a reference S parameter; performing microwave detection on the to-be-detected sample to obtain S parameters of the same type; carrying out wavelet transform de-noising processing on the obtained two groups of S parameters to eliminate the influence of local fluctuation points on calculation; performing mean square error MSE calculation on the two groups of de-noising processing S parameters; evaluating the defect condition of the to-be-detected sample according to the value of the mean square error MSE; the larger the mean square error MSE value is, the larger the S parameter difference of the to-be-detected sample and the reference defect sample is, and the higher the defect degree of the to-be-detected sample is. And continuing to detect other samples to be detected, and carrying out detection and difference analysis to complete defect evaluation of the samples to be detected. Defect evaluation is carried out by utilizing the difference value analysis method provided by the invention, the frequency band does not need to be manually selected, and the whole frequency band is defaulted, so that the detection complexity is greatly reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of microwave detection, and particularly relates to a difference analysis method for evaluating microwave detection results. Background Art

[0002] As an efficient and convenient non-destructive testing method, microwave detection has been gradually popularized in engineering sites for detecting internal defects of various materials. The conventional method of microwave detection is as follows: after performing microwave detection on a defect-free reference sample and a sample to be tested respectively, comparing the S-parameter frequency-domain curves of the two, and evaluating the defect state by analyzing the differences in the spectra. The above process often needs to be completed manually. In most cases, the S-parameter frequency-domain curves obtained from the detection of different defective samples have no rules to follow. Therefore, this work requires high experience of the evaluation personnel and takes a lot of time. Summary of the Invention

[0003] Aiming at the deficiencies of the traditional defect evaluation method for microwave detection, the present invention provides a difference analysis method for evaluating microwave detection results based on the error analysis method. By calculating the mean square error MSE between the frequency-domain curves of the sample to be tested and the defect-free sample, the rapid evaluation of defects is realized; thereby improving the overall detection efficiency and saving manpower input.

[0004] The technical solution adopted by the present invention is as follows:

[0005] A difference analysis method for evaluating microwave detection results, comprising the following steps:

[0006] Step 1: Perform microwave detection on a defect-free sample to obtain reference S parameters;

[0007] Step 2: Perform microwave detection on the sample to be tested and obtain S parameters of the same type as in Step 1;

[0008] Step 3: Perform wavelet transform denoising processing on the S parameters obtained in Step 1 and Step 2 respectively to eliminate the influence of local fluctuation points on the calculation. The denoising process is shown in the following formula:

[0009] f denoised (t) = IDWT(Threshold(DWT(f(t))))

[0010] Where, f(t) is the input noisy signal, specifically referring to the S-parameters obtained in Step 1 and Step 2; DWT is the discrete wavelet transform, which converts the noisy S-parameter signal from the time domain to the wavelet domain for facilitating subsequent noise processing. Threshold is the threshold processing, that is, in the wavelet domain, a threshold operation is performed on the coefficients after the discrete wavelet transform, setting the wavelet coefficients less than the set threshold to zero or performing other processing to remove or weaken the noise-related coefficients; IDWT is the inverse discrete wavelet transform, which converts the wavelet coefficients after the threshold processing back to the time domain to obtain the denoised signal f denoised (t), that is, the denoised S-parameters.

[0011] For the discrete signal x[n], the common formula for the discrete wavelet transform (DWT) based on the Mallat algorithm is as follows:

[0012]

[0013] The formula for the inverse discrete wavelet transform (IDWT) is as follows:

[0014]

[0015] Where, j represents the decomposition level; c j [n] represents the approximation coefficient (low-frequency part) of the j-th level; d j+1 [n] represents the detail coefficient (high-frequency part) of the (j + 1)-th level; h0[n] and h1[n] are the low-pass and high-pass filter coefficients respectively, which are determined by the selected wavelet function; n represents the sampling point position in the original discrete signal or the approximation coefficient sequence of the previous level; k represents the position of the approximation coefficient c j+1 [k] and the detail coefficient d j+1 [k] in the sequence.

[0016] Step 4: Calculate the mean square error (MSE) for the two groups of denoised S-parameters in Step 3. The calculation formula is as follows:

[0017]

[0018] Where, x i is the reference S-parameter at each frequency point, y i is the S-parameter to be detected at each frequency point, and n is the total number of microwave detection frequency points, that is, n points are evenly selected within the frequency band for detection.

[0019] Step 5: Evaluate the defect situation of the sample to be measured according to the value of the mean square error (MSE):

[0020] The larger the mean square error (MSE) value, the greater the difference in the S-parameters between the sample to be measured and the reference defective sample, indicating that the defect degree of the sample to be measured is higher.

[0021] Step 6: Continue to detect the remaining samples to be tested and perform detection and difference analysis to complete their defect assessment.

[0022] In Step 1, microwave detection is performed on defect-free samples, and the measured data includes: transmission parameter S 21 amplitude, transmission parameter S 21 phase, reflection parameter S 11 amplitude, reflection parameter S 11 phase frequency domain curve, etc. Generally, one of them can be selected according to the detection method used. This data serves as the benchmark for whether there are defects in the detection object, and the remaining data to be tested is compared with this data to achieve defect assessment.

[0023] The differential analysis method for microwave detection result evaluation of the present invention has the following technical effects:

[0024] 1) The present invention uses the error analysis method for analysis and calculation. By calculating the mean square error MSE between S parameters and comparing it with a reasonable range, the simplified processing of S parameters and the rapid evaluation of defects are realized. MSE can reflect the average gap between two sets of data.

[0025] 2) The present invention adopts the wavelet denoising algorithm, which can eliminate local abnormal points of S parameter amplitude, avoiding the distortion of the final result due to local amplitude abnormality, resulting in a high mean square error MSE and thus causing misjudgment.

[0026] 3) The present invention changes the data processing and evaluation method that has been used in microwave detection, greatly simplifying the workload; and the method of the present invention does not rely on manual work and can be used as the theoretical support for the automatic evaluation algorithm of microwave detection in the future. Brief Description of the Drawings

[0027] Figure 1 It is the flowchart of the method of the present invention.

[0028] Figure 2(a) is a schematic diagram of a defective sample;

[0029] Figure 2(b) is a partial enlarged view of Figure 2(a).

[0030] Figure 3 It is the detection result of the microwave transmission method.

[0031] Figure 4 It is the detection result after denoising.

[0032] Figure 5(a) is a schematic diagram of the microwave transmission method;

[0033] Figure 5(b) is a schematic diagram of the microwave reflection method. Detailed Embodiment

[0034] A differential analysis method for evaluating microwave detection results, which realizes the rapid evaluation of defects by calculating the mean square error MSE between the frequency-domain curves of the sample to be measured and the defect-free sample. The method includes the following steps:

[0035] Step (1): Build a microwave detection platform, perform microwave detection on the defect-free sample, and obtain the reference S parameters;

[0036] The measured data includes the transmission parameter S 21 amplitude, the transmission parameter S 21 phase, the reflection parameter S 11 amplitude, the reflection parameter S 11 phase frequency-domain curves, etc.

[0037] As shown in Figures 5(a) and 5(b), they are schematic diagrams of the microwave transmission method and the microwave reflection method respectively. Among them, transmission means that the microwave signal emitted by the transmitting antenna penetrates the sample to be measured and is received by the receiving antenna, and reflection means that the microwave signal emitted by the transmitting antenna acts on the sample to be measured and is reflected and then received by the transmitting antenna again. Let the incident wave, transmitted wave, and reflected wave signals be a1, b2, and b1 respectively. Then the transmission coefficient S 21 and the reflection coefficient S 11 are:

[0038] S 21 = b2 / a1

[0039] S 11 = b1 / a1

[0040] Converting the unit to the dB form, we can get:

[0041] S 21 = 10lg(b2 / a1) 2 = 20lg(b2 / a1)

[0042] S 11 = 10lg(b1 / a1) 2 = 20lg(b1 / a1)

[0043] Among them, each parameter is a vector. The amplitude represents the ratio between the transmitted wave (reflected wave) and the incident wave, and the phase represents the phase difference between the transmitted wave (reflected wave) and the incident wave.

[0044] Each of the above S-parameter data contains the sample dielectric distribution information. Therefore, generally, one of them can be selected according to the detection method used.

[0045] This data is used as the reference for whether the detection object has defects, and the remaining data to be measured is compared with this data to achieve defect evaluation.

[0046] Step (2): Measure the S parameters of the sample to be measured;

[0047] According to step (1), perform microwave detection on the sample to be tested and obtain S-parameters of the same type.

[0048] Step (3): Perform wavelet transform denoising on the two sets of S-parameters to eliminate the influence of local fluctuation points on the calculation.

[0049] Step (4): Calculate the mean square error between the above two sets of data. The calculation formula is as follows:

[0050]

[0051] In the formula, x i is the reference S-parameter at each frequency point, y i is the S-parameter to be detected at each frequency point, and n is the total number of frequency points for microwave detection, that is, n points are evenly selected within the frequency band for detection.

[0052] Step (5): Evaluate the defect situation according to the value of MSE;

[0053] The larger the MSE value, the greater the difference in S-parameters between the sample to be tested and the reference sample, indicating a higher degree of defect in the sample to be tested. Generally, it is considered that a sample with MSE greater than 1 contains defects, and this value will vary according to different detection objects;

[0054] Step (6): Continue to detect the remaining samples to be tested and perform detection and difference analysis to complete their defect evaluation.

[0055] Specific example:

[0056] Taking the defect detection of 10 kV suspension composite insulators as an example. Set crack defects at the junction of the umbrella skirt and the sheath of the insulator. The number of cracks is 1 - 6, and the setting method is shown in Fig. 2(a) and Fig. 2(b). The crack diameter is 2 mm.

[0057] Perform microwave detection on the defect-free sample (reference sample) and the samples with different numbers of cracks (0 - 6). The detection frequency is 5 - 18 GHz, and the number of frequency points is 201. The detection results are as Figure 3 shown. Since microwave detection is highly sensitive to frequency, the detection effect can only be observed in some frequency bands: as the number of cracks increases, the S 21 amplitude shows an upward trend. Perform wavelet denoising on the data, and the results are as Figure 4 shown: The denoising process does not affect the original law of the S-parameters, but makes the curve smoother and eliminates some large local errors. Perform difference analysis on the above S-parameters, and the obtained MSE is shown in Table 1.

[0058] Table 1 Results of difference analysis

[0059]

[0060]

[0061] It can be clearly seen from the data in Table 1 that as the number of cracks increases, the overall MSE shows an increasing trend, and the presence of defects has an obvious increasing effect on MSE compared with the defect-free samples. In addition, using the difference analysis method proposed in the present invention for defect evaluation, there is no need to manually select frequency bands, and the overall frequency band can be defaulted, which greatly reduces the detection complexity. For the detection object studied in this example, it can be considered that there are defects when MSE is greater than 1. The present invention has a high accuracy for defect detection evaluation.

Claims

1. A differential analysis method for evaluating microwave detection results, characterized in that It includes the following steps: Step 1: Perform microwave detection on a defect-free sample to obtain S-parameters; Step 2: Perform microwave detection on the sample to be tested and obtain S-parameters of the same type as in Step 1; Step 3: Perform wavelet transform denoising on the S-parameters obtained in Step 1 and Step 2 respectively to eliminate the influence of local fluctuation points on the calculation. The denoising process is shown in the following formula: f denoised (t) = IDWT(Threshold(DWT(f(t)))) In the formula, f(t) is the input noisy signal, specifically referring to the S parameter; DWT is the discrete wavelet transform; IDWT is the inverse discrete wavelet transform, which converts the wavelet coefficients after threshold processing into the time domain to obtain the denoised signal f denoised (t), that is, the denoised S parameter; Step 4: Calculate the mean square error MSE of the two groups of denoised S-parameters in Step 3. The calculation formula is shown as follows: where x i is the reference S-parameter at each frequency point, y i is the S-parameter to be detected at each frequency point, and n is the total number of microwave detection frequency points, that is, n points are uniformly selected within the frequency band for detection; Step 5: Evaluate the defect condition of the sample to be tested according to the value of the mean square error MSE: Step 6: Continue to detect the remaining samples to be tested and perform detection and difference analysis to complete their defect evaluation.

2. The differential analysis method for microwave detection result evaluation according to claim 1, characterized in that: In the said step 1, microwave detection is performed on the defect-free samples, and the measured data includes: transmission parameter S 21 amplitude, transmission parameter S 21 phase, reflection parameter S 11 amplitude, reflection parameter S 11 phase frequency domain curve.

3. The differential analysis method for microwave detection result evaluation according to claim 2, wherein: Transmission means that the microwave signal emitted by the transmitting antenna penetrates the sample to be tested and is received by the receiving antenna, and reflection means that the microwave signal emitted by the transmitting antenna acts on the sample to be tested and is reflected and then received by the transmitting antenna again.

4. The differential analysis method for microwave detection result evaluation according to claim 3, characterized in that: Let the incident wave, transmitted wave and reflected wave signals be \(a_1\), \(b_2\) and \(b_1\) respectively, and the transmission coefficient \(S\) can be obtained 21 and the reflection coefficient \(S\ 11 : S 21 = b2 / a1 S 11 = b1 / a1 Converting the unit to the dB form gives: S 21 = 10 lg(b2 / a1) 2 = 20 lg(b2 / a1) S 11 = 10 lg(b1 / a1) 2 = 20 lg(b1 / a1) Among them, each parameter is a vector. The amplitude represents the ratio between the transmitted wave or the reflected wave and the incident wave, and the phase represents the phase difference between the transmitted wave or the reflected wave and the incident wave; Each of the above S-parameter data contains sample dielectric distribution information, so one of them can be selected according to the detection method used.

5. The differential analysis method for microwave detection result evaluation according to claim 1, characterized in that: In Step 3, for the discrete signal x[n], the discrete wavelet transform DWT formula based on the Mallat algorithm is as follows: The inverse discrete wavelet transform IDWT formula is as follows: Where j represents the decomposition level; c j [n] represents the approximation coefficient of the j-th layer, which is the low-frequency part; d j+1 [n] represents the detail coefficient of the (j + 1)-th layer, which is the high-frequency part; h0[n] and h1[n] are the low-pass and high-pass filter coefficients respectively, determined by the selected wavelet function; n represents the sampling point position in the original discrete signal or the approximation coefficient sequence of the previous layer; k represents the position of the approximation coefficient c j+1 [k] and the detail coefficient d j+1 [k] in the sequence.

6. The differential analysis method for microwave detection result evaluation according to claim 1, characterized in that: In Step 5, the larger the mean square error MSE value, the greater the difference in S-parameters between the sample to be tested and the reference defect sample, which indicates that the defect degree of the sample to be tested is higher.

Images