阅读说明：本技术 一种双极化气象雷达差分相位的质量控制方法 (Quality control method for dual-polarization meteorological radar differential phase ) 是由 闫贺 徐星 周晔 汪玲 于 2021-09-09 设计创作，主要内容包括：本发明公开了一种双极化气象雷达差分相位的质量控制方法,步骤依次为,对双极化雷达实测差分相位进行预处理；利用差分相位,水平反射率因子和差分反射率因子之间的关系,去除实测差分相位值中存在的后向差分相位误差；根据同一径向上差分相位的变换情况,进行LSTM处理,从而达到对差分相位的滤波效果；遍历双极化雷达射线全方位扫描所有径向的实测差分相位值,进行LSTM处理；基于处理之后的降水差分相位值,利用反射率衰减订正算法进行反射率的衰减订正。本发明可以有效提高反射率衰减订正中存在的精度问题,提高衰减订正的稳定性。(The invention discloses a quality control method of a dual-polarization meteorological radar differential phase, which comprises the following steps of preprocessing the actually measured differential phase of the dual-polarization meteorological radar; removing a backward differential phase error existing in the actually measured differential phase value by utilizing the relationship among the differential phase, the horizontal reflectivity factor and the differential reflectivity factor; according to the transformation condition of the differential phase in the same radial direction, LSTM processing is carried out, so that the filtering effect of the differential phase is achieved; traversing the dual-polarized radar rays to omni-directionally scan all radial actually-measured differential phase values, and performing LSTM processing; and performing reflectivity attenuation correction by using a reflectivity attenuation correction algorithm based on the processed precipitation differential phase value. The invention can effectively improve the precision problem existing in the reflectivity attenuation correction and improve the stability of the attenuation correction.)

1. A quality control method for a dual-polarization meteorological radar differential phase is characterized by comprising the following steps:

(1) preprocessing the actually measured differential phase of the dual-polarization radar, including processing pockmark-shaped clutter, mutation clutter and strip-amplitude clutter in a PPI display graph;

(2) removing a backward differential phase error existing in the actually measured differential phase value by utilizing the relationship among the differential phase, the horizontal reflectivity factor and the differential reflectivity factor;

(3) according to the transformation situation of the differential phase in the same radial direction, carrying out LSTM processing and filtering the differential phase;

(4) traversing the dual-polarized radar rays to omni-directionally scan all radial actually-measured differential phase values, and performing LSTM processing;

(5) and performing reflectivity attenuation correction by using a reflectivity attenuation correction algorithm based on the differential phase value after the LSTM processing.

2. The method for controlling the differential phase of the dual-polarization weather radar according to claim 1, wherein in the step (1), the pockmarked clutter is processed by the following steps:

setting the size of the sliding window as 2m multiplied by 2n, i as an azimuth distance library, j as a radial distance, calculating the percentage of the number of distance libraries with effective values in the window to the number of distance libraries in the total window, and when the occupation ratio is less than a set threshold value gamma, considering the central data point of the window as a non-meteorological echo point and carrying out rejection processing, namely:

wherein deg is_i,jGiving differential phase values of distance library points (i, j) for dual-polarization radar data, NaN being actually measured invalid data, P_i,jThe effective differential phase value in the selected window is a percentage of the total distance bin number of the window, when P_i,jIf the noise is less than the set threshold value gamma, the data point (i, j) is judged to be isolated point noise and is eliminated.

3. The method for controlling the differential phase of the dual-polarization weather radar according to claim 1, wherein in the step (1), the abrupt change points are processed by the following steps:

taking each distance library as a center, and scanning data on a full 360-degree full-distance path by using a sliding window; constructing a window with the size of 2m multiplied by 2n, and calculating differential phase values deg on other distance libraries in the window_iDifferential phase deg from center distance bin₀Absolute value of difference Δ deg_i1, 2., (2m × 2n) -1, i.e.:

Δdeg_i＝|deg_i-deg₀|

recording the number N of the obtained data which is larger than a threshold value T1, if N/(2m multiplied by 2N) -1 exceeds a threshold value T2, judging the distance library as a mutation mixed point, and rejecting a data value on the distance library; and then, carrying out interpolation processing by using data in the sliding window to enable the data value on the distance database to meet the change rule of the radar radial data and the distance direction data.

4. The method for controlling the differential phase of the dual-polarization weather radar according to claim 1, wherein in the step (1), the strip-amplitude clutter is processed by the following steps:

confirming the position of the error differential phase, and setting the adjacent radial differential phase difference: deltaΦ_dp＝Φ_dp(j)-Φ_dp(j-1), where j is the distance bin position in the radial direction, Φ_dpIs a differential phase value; setting a parameter N_thThe number of distance bins with the difference value in two adjacent radial directions larger than alpha dB, N_th1Number of range bins for which the differential phase value between row i and row i-1 is greater than α dB, N_th2The number of distance bins for which the differential phase value between the i +1 th row and the i-th row is greater than α dB, the parameter Δ N_th＝|N_th1-N_th2If Δ N_thWhen the value of the interference parameter beta is larger than the set parameter beta, judging that the data in the radial direction has errors, and determining the specific direction of the strip amplitude interference echo; and (3) deducting all the strip-shaped interference echoes, correcting by adopting an interpolation method, selecting a first radial distance library, constructing a window with the distance library as a center on each distance library, and carrying out mean value processing on values in the window.

5. The method for controlling the quality of the differential phase of the dual-polarization meteorological radar according to claim 1, wherein in the step (2), when the rayleigh scattering condition is satisfied, the backward differential phase is ignored, in the dual-polarization meteorological radar, the backward differential phase is set in the same radial radar ray, and if the differential reflectivity factors measured in the two distance bins are the same, the difference value of the differential phase of the two distance bins is the difference value of the precipitation differential phase, that is:

ΔΨ_dp＝ΔΦ_dp

therein, Ψ_dpRepresenting the differential phase, phi, measured directly by a dual-polarization radar_dpIndicating the differential phase of precipitation. Setting the number of the same radial distance library of the radar as M, and setting a matrix P:

let the value of each pixel in the matrix P be represented by P (i, j), where i is the row of the matrix and j is the column of the matrix, then there are:making the ith behavior a of the matrix A be in a binary representation mode to obtain the matrix A;

setting a differential reflectivity factor obtained by actual measurement of radar as Z_drThen, there are:

ΔZ_dr＝P·Z_dr

ΔΨ_dp＝P·Ψ_dp

of interest wherein Δ Z_drLine 0, corresponding to Δ Ψ at that time_dp＝ΔΦ_dpRecording the rows meeting the conditions, and separately listing the corresponding rows of the matrix A to obtain a matrix A';

computingWherein Z is_hhRepresenting the horizontal reflectivity factor obtained by the system actual measurement, and normalizing each line of the parameter w to obtain w by the data sum of the line_normCalculating a differential phase shift rate K_dp：

Wherein, Δ r represents the length of a single distance unit, and K (i, j) and a '(i, j) are the values of each pixel point in K and a', respectively;

the differential phase is reconstructed using the differential phase shift rate by:

wherein r is_jRepresenting the jth distance bin in the radial direction.

6. The method for controlling the differential phase of dual polarized weather radar according to claim 1, wherein in step (3), the length of the input sequence in the LSTM network is selected according to the length of the unit distance library in the radial direction of the dual polarized radar and the length of the radiation distance of the radar.

7. The method for controlling the differential phase of dual-polarized weather radar according to claim 1, wherein in the step (5), the reflectivity attenuation correction algorithm adopts a precipitation profile method:

let the attenuation rate be A_HThe rain zone has a range of (r)₁,r₀) Then, there are:

ΔΦ_dp＝Φ_dp(r₀)-Φ_dp(r₁)

wherein r represents the distance from the reservoir to the radar center in the radial direction, a, b are attenuation correction coefficients, and Z_hhTo attenuate the reflectivity factor, phi, before correction_dpIndicating differential phase of precipitation, Δ Φ_dpThe difference value of the differential phase of the precipitation;

the attenuation correction formula is as follows:

wherein, Z'_hhQ is the reflectance factor after the attenuation correction, and q is the attenuation correction coefficient.

Technical Field

The invention belongs to the technical field of radars, and particularly relates to a quality control method for a dual-polarization meteorological radar differential phase.

Background

In the dual-polarization meteorological radar, the reflectivity factors with high quality and high precision directly determine the capacity of the radar for estimating precipitation and identifying the type of precipitation particles. But there is some attenuation of the reflectivity factor due to the influence of other non-precipitation particles in the air and the presence of other electromagnetic interference. The differential phase shift rate obtained by the echo received by the dual-polarization radar is not influenced by the energy attenuation of radar electromagnetic waves, and the attenuation correction of the reflectivity factor mainly depends on the calculation of the differential phase shift rate. The differential phase value obtained by directly measuring the dual-polarization radar echo consists of a back scattering phase, a precipitation differential phase and an interference clutter phase, so that the correction precision in attenuation correction is improved by controlling the quality of the differential phase.

Disclosure of Invention

In order to solve the technical problems mentioned in the background art, the invention provides a quality control method for a dual-polarization meteorological radar differential phase.

In order to achieve the technical purpose, the technical scheme of the invention is as follows:

a quality control method for a dual-polarization meteorological radar differential phase comprises the following steps:

(1) preprocessing the actually measured differential phase of the dual-polarization radar, including processing pockmark-shaped clutter, mutation clutter and strip-amplitude clutter in a PPI display graph;

(2) removing a backward differential phase error existing in the actually measured differential phase value by utilizing the relationship among the differential phase, the horizontal reflectivity factor and the differential reflectivity factor;

(3) according to the transformation situation of the differential phase in the same radial direction, carrying out LSTM processing and filtering the differential phase;

(4) traversing the dual-polarized radar rays to omni-directionally scan all radial actually-measured differential phase values, and performing LSTM processing;

(5) and performing reflectivity attenuation correction by using a reflectivity attenuation correction algorithm based on the differential phase value after the LSTM processing.

Further, in the step (1), the method for processing the pockmarked clutter includes:

setting the size of the sliding window as 2m multiplied by 2n, i as an azimuth distance library, j as a radial distance, calculating the percentage of the number of distance libraries with effective values in the window to the number of distance libraries in the total window, and when the occupation ratio is less than a set threshold value gamma, considering the central data point of the window as a non-meteorological echo point and carrying out rejection processing, namely:

wherein deg is_i,jGiving differential phase values of distance library points (i, j) for dual-polarization radar data, NaN being actually measured invalid data, P_i,jThe effective differential phase value in the selected window is a percentage of the total distance bin number of the window, when P_i,jIf the noise is less than the set threshold value gamma, the data point (i, j) is judged to be isolated point noise and is eliminated.

Further, in the step (1), the method for treating the mutation site is as follows:

taking each distance library as a center, and scanning data on a full 360-degree full-distance path by using a sliding window; constructing a window with the size of 2m multiplied by 2n, and calculating differential phase values deg on other distance libraries in the window_iDifferential phase deg from center distance bin₀Absolute value of difference Δ deg_i1, 2., (2m × 2n) -1, i.e.:

Δdeg_i＝|deg_i-deg₀|

recording the number N of the obtained data which is larger than a threshold value T1, if N/(2m multiplied by 2N) -1 exceeds a threshold value T2, judging the distance library as a mutation mixed point, and rejecting a data value on the distance library; and then, carrying out interpolation processing by using data in the sliding window to enable the data value on the distance database to meet the change rule of the radar radial data and the distance direction data.

Further, in step (1), the method for processing the strip-amplitude clutter is as follows:

confirming the position of the error differential phase, and setting the adjacent radial differential phase difference: delta phi_dp＝Φ_dp(j)-Φ_dp(j-1), where j is the distance bin position in the radial direction, Φ_dpIs a differential phaseA bit value; setting a parameter N_thThe number of distance bins with the difference value in two adjacent radial directions larger than alpha dB, N_th1Number of range bins for which the differential phase value between row i and row i-1 is greater than α dB, N_th2The number of distance bins for which the differential phase value between the i +1 th row and the i-th row is greater than α dB, the parameter Δ N_th＝|N_th1-N_th2If Δ N_thWhen the value of the interference parameter beta is larger than the set parameter beta, judging that the data in the radial direction has errors, and determining the specific direction of the strip amplitude interference echo; and (3) deducting all the strip-shaped interference echoes, correcting by adopting an interpolation method, selecting a first radial distance library, constructing a window with the distance library as a center on each distance library, and carrying out mean value processing on values in the window.

Further, in step (2), when the rayleigh scattering condition is satisfied, ignoring the backward differential phase, in the dual-polarization meteorological radar, setting in the same radial radar ray, and if the values of the differential reflectivity factors measured in the two distance bins are the same, determining that the difference value of the differential phase of the two distance bins is the difference value of the precipitation differential phase, that is:

ΔΨ_dp＝ΔΦ_dp

therein, Ψ_dpRepresenting the differential phase, phi, measured directly by a dual-polarization radar_dpIndicating the differential phase of precipitation. Setting the number of the same radial distance library of the radar as M, and setting a matrix P:

let the value of each pixel in the matrix P be represented by P (i, j), where i is the row of the matrix and j is the column of the matrix, then there are:making the ith behavior a of the matrix A be in a binary representation mode to obtain the matrix A;

setting a differential reflectivity factor obtained by actual measurement of radar as Z_drThen, there are:

ΔZ_dr＝P·Z_dr

ΔΨ_dp＝P·Ψ_dp

of interest wherein Δ Z_drLine 0, corresponding to Δ Ψ at that time_dp＝ΔΦ_dpRecording the rows meeting the conditions, and separately listing the corresponding rows of the matrix A to obtain a matrix A';

computingWherein Z is_hhRepresenting the horizontal reflectivity factor obtained by the system actual measurement, and normalizing each line of the parameter w to obtain w by the data sum of the line_normCalculating a differential phase shift rate K_dp：

Wherein, Δ r represents the length of a single distance unit, and K (i, j) and a '(i, j) are the values of each pixel point in K and a', respectively;

the differential phase is reconstructed using the differential phase shift rate by:

wherein r is_jRepresenting the jth distance bin in the radial direction.

Further, in step (3), the length of the input sequence in the LSTM network is selected according to the unit distance library length in the radial direction of the dual-polarized radar and the radiation distance length of the radar.

Further, in step (5), the reflectance attenuation correction algorithm adopts a precipitation profile method:

let the attenuation rate be A_HThe rain zone has a range of (r)₁,r₀)，Then there are:

ΔΦ_dp＝Φ_dp(r₀)-Φ_dp(r₁)

wherein r represents the distance from the reservoir to the radar center in the radial direction, a, b are attenuation correction coefficients, and Z_hhTo attenuate the reflectivity factor, phi, before correction_dpIndicating differential phase of precipitation, Δ Φ_dpThe difference value of the differential phase of the precipitation;

the attenuation correction formula is as follows:

wherein, Z'_hhQ is the reflectance factor after the attenuation correction, and q is the attenuation correction coefficient.

Adopt the beneficial effect that above-mentioned technical scheme brought:

the method takes various interference conditions existing in the actually measured data of the dual-polarization radar under the real condition into consideration, and processes pockmark-shaped clutter, mutation clutter and strip-amplitude clutter which interfere in the PPI. The differential phase obtained by direct measurement of the dual-polarization radar comprises the precipitation differential phase and the backward differential phase of the interference. The filtering effect on the precipitation differential phase is achieved by utilizing the training and prediction method of the LSTM network. Compared with the traditional Kalman filtering method, the LSTM does not need to depend on the value of the differential propagation phase shift obtained by calculation and completely depends on the differential phase value obtained by measurement, so that errors possibly introduced by calculating the differential propagation phase shift are reduced. The invention can effectively improve the precision problem existing in the reflectivity attenuation correction and improve the stability of the attenuation correction.

Drawings

FIG. 1 is an overall flow diagram of the present invention;

FIG. 2 is a PPI display graph of differential phase measured by the dual-polarization radar;

FIG. 3 is a PPI graph showing differential phase after elimination of pockmarked clutter;

FIG. 4 is a PPI display of differential phase after eliminating abrupt discontinuities;

FIG. 5 is a PPI display of differential phase after strip-amplitude clutter removal;

FIG. 6 is a diagram of the backward differential phase PPI display;

FIG. 7 is a diagram of the precipitation differential phase PPI;

FIG. 8 is a comparison of the same radial precipitation differential phase before and after filtering;

FIG. 9 is a block diagram of the structure of an LSTM network;

FIG. 10 is a PPI display of the reflectance factor after the relaxation correction.

Detailed Description

The technical scheme of the invention is explained in detail in the following with the accompanying drawings.

The invention designs a quality control method for a dual-polarization meteorological radar differential phase, and the overall flow is shown in figure 1.

The embodiment performs data processing on the basis of the actual measurement data of the existing dual-polarization radar system. The software and hardware configurations used for the experiments are shown in table 1:

TABLE 1 Experimental Environment

The method comprises the following steps:

step 1, preprocessing the actually measured differential phase of the dual-polarization radar, including processing pockmark-shaped clutter, mutation clutter and strip-amplitude clutter in a PPI display graph;

step 2, removing a backward differential phase error existing in an actually measured differential phase value by utilizing the relationship among the differential phase, the horizontal reflectivity factor and the differential reflectivity factor;

step 3, according to the transformation situation of the differential phase in the same radial direction, carrying out LSTM processing and filtering the differential phase;

step 4, traversing the dual-polarized radar rays to omni-directionally scan all radial actually-measured differential phase values, and performing LSTM processing;

and 5, based on the differential phase value processed by the LSTM, performing reflectivity attenuation correction by utilizing a reflectivity attenuation correction algorithm.

In this embodiment, the step 1 is implemented by the following preferred scheme:

the method comprises the steps of preprocessing the actually measured reflectivity factor value of the dual-polarization radar, and processing pockmark-shaped clutter, mutation clutter and strip-amplitude clutter in a plane position display graph. The PPI display of the differential phase measured by the dual-polarized weather radar is shown in FIG. 2.

The method for processing the pockmarked clutter comprises the following main steps:

setting the size of a sliding window to be 2m multiplied by 2n, wherein i is represented as an azimuth distance library, j is represented as a radial distance, calculating the percentage of the number of distance libraries of effective values in the window to the number of distance libraries in the total window, and when the occupation ratio is less than a certain threshold value gamma, considering the central data point of the window as a non-meteorological echo point and carrying out rejection processing, namely:

wherein deg is_i,jGiving differential phase values of distance library points (i, j) for dual-polarization radar data, NaN being actually measured invalid data, P_i,jThe effective differential phase value in the selected window is a percentage of the total distance bin number of the window, when P_i,jIf the value is less than the set threshold value, the data point (i, j) is judged to be isolated point noise and is removed. The result of the spur-like clutter cancellation is shown in fig. 3.

The main steps for removing mutation and miscellaneous points in the data are as follows:

firstly, determining the position of the mutation point, centering on each distance library, and utilizing a sliding window to align data on a full 360 DEG and full distance pathA scan is performed. Still construct a small window of size 2m x 2n and compute the difference between the differential phase values on the other range bins and the center range bin within the window. Let the differential phase of the central distance library be deg₀The differential phase value in the other window is deg_iThe method comprises the following steps:

Δdeg_i＝|deg_i-deg₀|

recording the number N of the obtained data which is larger than a threshold value T1, if N/(2m multiplied by 2N) -1 exceeds a threshold value T2, judging the distance library as a mutation mixed point, and rejecting a data value on the distance library; and then, carrying out interpolation processing by using data in the sliding window to enable the data value on the distance database to meet the change rule of the radar radial data and the distance direction data. The result of eliminating the mutant spots is shown in FIG. 4.

The main steps for eliminating the strip-shaped clutter are as follows:

(1) confirming the position of the error differential phase, and setting the adjacent radial differential phase difference as follows: delta phi_dp＝Φ_dp(j)-Φ_dp(j-1), where j is the distance bin position in the radial direction, Φ_dpIs the value of the differential phase. Setting parameter N_thThe number of the distance library with the difference value larger than alpha dB in the two adjacent radial directions is shown. Note N_th1Number of range bins for which the differential phase value between row i and row i-1 is greater than α dB, N_th2The number of distance bins for which the differential phase value between the i +1 th row and the i-th row is greater than α dB. Setting parameter DeltaN_thComprises the following steps: delta N_th＝|N_th1-N_th2If Δ N_thIf the value of (b) is greater than a predetermined parameter β, it is determined that the data in the radial direction has an error.

(2) After the specific direction of the strip-shaped interference echo is determined, the strip-shaped interference echo is completely deducted, then an interpolation method is adopted for correction, a first radial distance library is selected to start, a window with the distance library as the center is constructed on each distance library, and the value in the window is subjected to mean processing. The resulting differential phase values are shown in fig. 5.

In this embodiment, the step 2 is implemented by the following preferred scheme:

when the Rayleigh scattering condition is met, the backward difference phase can be ignored, in the dual-polarization meteorological radar, the backward difference phase is set in the same radial radar ray, and if the values of the difference reflectivity factors measured in the two distance libraries are the same, the difference value of the difference phases of the two distance libraries is the difference value of the precipitation difference phase. Namely:

ΔΨ_dp＝Ψ_dp(r_b)-Ψ_dp(r_a)＝ΔΦ_dp

therein, Ψ_dpRepresenting the differential phase, phi, measured directly by a dual-polarization radar_dpIndicating the differential phase of precipitation. Assuming that the number of the same radial distance bins of the radar is M, setting a matrix P as follows:

wherein the content of the first and second substances,is the number of combinations in the mathematical permutation combination concept. Obtaining a matrix A according to the value of the matrix P, and assuming that the value of each pixel point in the matrix P is represented by P (i, j), wherein i is a row of the matrix and j is a column of the matrix, then:the ith row of a is represented in binary form (low on the left).

When M is 4, the matrix a can be represented as:

setting a differential reflectivity factor obtained by actual measurement of radar to be Z_drThen, there are:

ΔZ_dr＝P·Z_dr

ΔΨ_dp＝P·Ψ_dp

of interest wherein Δ Z_drLine 0, corresponding to Δ Ψ at that time_dp＝ΔΦ_dpWill satisfyThe number of rows of the above condition is recorded and these corresponding rows of matrix a are listed separately to obtain matrix a'. Then calculateWherein the parameter Z_hhExpressed as the horizontal reflectivity factor measured by the system. Each row of the parameter w is normalized by the sum of the data of the row: w is a_norm. Then the differential dependent rate K is obtained_dpThe values of (A) are:

Δ r represents the length of a single distance element.

The differential phase is reconstructed using the differential phase shift rate by:

wherein r is_jRepresenting the jth distance bin in the radial direction.

The resulting forward differential phase PPI display is shown in fig. 6. The PPI display of the precipitation differential phase is shown in fig. 7.

In this embodiment, the step 3 is implemented by the following preferred scheme:

in an LSTM network, network training is performed based on an input sequence into the network to predict the value of the next data point. In the network training, the input differential phase value sequence needs to be comprehensively learned at multiple points, and the network weakens the characteristics of the points with large variation, so that the filtering effect is achieved on the prediction result of the whole radial data. A comparison of before and after filtering is shown in fig. 8.

LSTM is a special recurrent neural network, which is trained by inputting the state of the previous time and the state data of the previous time into the network, and predicts the state data of the next time by using the training result of each time. The input sequence of the network is set by observing the change condition of the differential phase in the precipitation interval, so that the differential phase value processed by the LSTM is obtained, and the filtering effect on the differential phase is achieved.

In this embodiment, the step 4 is implemented by adopting the following preferred scheme:

and carrying out LSTM processing on all radial data of the full 360-degree azimuth of the dual-polarization radar. The structure of the LSTM is shown in fig. 9.

In this embodiment, the step 5 is implemented by the following preferred scheme:

the obtained differential reflectivity factor is subjected to attenuation correction, the attenuation correction algorithm of the reflectivity of the dual-polarization radar comprises a differential phase correction method, a precipitation profile method, a self-adaptive constraint algorithm and the like, and the precipitation profile method is taken as an example in the invention. Assuming a decay Rate of A_HThe rain zone has a range of (r)₁,r₀) Then, there are:

ΔΦ_dp＝Φ_dp(r₀)-Φ_dp(r₁)

the reflectance factor after attenuation correction is set to Z'_hhThe value of the reflectivity factor before the fall-off correction is Z_hhR represents the distance from the database to the radar center in the radial direction, a and b are attenuation correction coefficients, the coefficient a is related to the temperature and the size of the liquid drop, a priori values need to be set first, and a obtained by fitting according to the Andger relationship is 0.293. When the temperature is constant, b is constant. Therefore, the value of b at a fixed temperature is determined, and the value range of b is (0.76, 0.84).

The formula for obtaining the attenuation correction is:

q is the attenuation correction coefficient and takes the value of 1.224. The PPI display of the reflectance factor after the relaxation correction is shown in fig. 10.

The embodiments are only for illustrating the technical idea of the present invention, and the technical idea of the present invention is not limited thereto, and any modifications made on the basis of the technical scheme according to the technical idea of the present invention fall within the scope of the present invention.