An inversion algorithm for insect body length parameters in oblique flight

Abstract

The present invention discloses an algorithm for inverting the body length parameter of insects in a slanting flight posture, aiming to provide a method for estimating the body length of insects in a non-level flight posture, so as to solve the problem of low inversion accuracy caused by the unified use of the level flight assumption for inverting the body size of migrating insects by vertical monitoring insect radars. In the case where the posture angles of insects measured by the radar are known, this method can compensate the body length-inflexion point wavelength fitting curve of level flight insects to obtain a slanting flight fitting curve at the current posture angle of the insects; then obtain the multi-frequency point radar cross section (RCS) data of the insects, and search for the frequency value f0 corresponding to the highest peak inflection point of the multi-frequency RCS curve of the insects based on the least squares idea, and calculate the corresponding wavelength λ0; next, substitute λ0 into the compensated slanting flight fitting curve to obtain the estimated value of the body length of the insects at the current non-level flight posture angle.

Description

Technical Field

[0001] The invention belongs to the technical field of insect radars, and in particular relates to an inversion algorithm for body length parameters of insects in an oblique flying posture. Background Art

[0002] Insect radar has experienced the development from scanning radar system to vertical radar system. The most widely used insect radar is the ZLC system vertical observation radar. The antenna of this type of radar points vertically upward, and the linear polarization beam deviates from the vertical axis at a very small angle to quickly scan the cone, which can measure the RCS of the target in different linear polarization directions, that is, the polarization pattern. Under the assumption of "insects flying horizontally", the polarization pattern characteristics of the insect abdomen when the radar wave is incident in the vertical direction are studied. It is found that the RCS of small and medium-sized insects is the largest when the body axis is parallel to the polarization direction, and the mapping relationship between the average RCS of the insect polarization pattern and the body length. By extracting the direction of the maximum RCS and fitting the relationship between the average RCS and the body length, the inversion capability of the ZLC system vertical observation radar for the insect body length can be realized.

[0003] At present, the inversion of insect biological parameters by insect radar is carried out under the assumption that "insects fly level". In the cruising stage of migration, insects mainly fly level; in the take-off and landing stages, insects mainly fly obliquely. The study of insect swarm migration based on dual-polarization weather radar found that even in the cruising stage, insects fly obliquely at small angles. When insects fly obliquely, the vertically observed radar waves are no longer vertically incident, and the surface of the insect body facing the radar has changed. Since the electromagnetic scattering of insects is sensitive to posture, this will obviously affect the echo characteristics of the insects, and then affect the target biological parameters measured by insect radar. However, to date, there has been no research on the impact of insect oblique flight on the accuracy of the body axis orientation and body length inversion method proposed based on the "insect flying level" assumption, nor has there been any research on the inversion of insect body posture based on insect radar. Summary of the invention

[0004] In view of this, the present invention provides a method for estimating the body length of an insect in a non-level flight posture. The technical solution of the present invention is:

[0005] An inversion algorithm for insect body length parameters in oblique flight posture, the method comprising the following steps:

[0006] Step 1: Use a multi-band radar system to obtain radar scattering data of the insect to be tested at different frequencies, and further analyze the relationship between the frequency and the measured data of the maximum RCS value;

[0007] Step 2: Based on the RCS-frequency measurement data curve obtained in step 1, a search method based on the least squares meaning is used to obtain a parameter estimation value of a peak inflection point in the RCS-frequency measurement data curve;

[0008] Step 3: Use the wavelength corresponding to the peak inflection point in the RCS-frequency measurement data curve obtained in step 2 and the pitch angle of the insect body axis relative to the plane of the radar antenna polarization coordinate system to invert the body length of the oblique flying insect.

[0009] Beneficial effects:

[0010] The present invention proposes a method for estimating the body length of insects in non-level flight postures, which is expected to solve the problem of low inversion accuracy caused by the uniform use of the level flight assumption in vertical monitoring insect radars for body length inversion of migratory insects, and is of great significance for insect radar target recognition and classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 , Experimental scenario of measuring the fully polarized scattering matrix of insects with different body inclination angles;

[0012] Figure 2 , multi-viewing angle antenna support structure;

[0013] Figure 3 , scatter plot of body length error;

[0014] Figure 4 , the average error of body length inversion at different oblique flight angles;

[0015] Figure 5 , algorithm flow chart. DETAILED DESCRIPTION

[0016] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0017] Step 1: Obtain insect multi-frequency RCS data

[0018] Use a multi-band radar system to detect the insects to be tested and obtain the scattered signals of N frequency points. The number of frequencies in each band is not less than 10. For the nth frequency point, obtain the full polarization scattering matrix corresponding to the insect to be tested, expressed as:

[0019]

[0020] in, is the horizontal polarization component, is the vertical polarization component, and is the cross-polarization component.

[0021] Substitute the full polarization scattering matrix into the following equation:

[0022]

[0023] Get the measured data R of the maximum RCS value of the nth frequency point nTraverse N frequency points and obtain the measurement data of the maximum RCS value corresponding to each frequency point. Take the frequency corresponding to each frequency point as the horizontal coordinate and the measurement data of the maximum RCS value as the vertical coordinate to form a measurement data curve. Based on the measurement data curve, the power function relationship between the frequency and the measurement data of the maximum RCS value is obtained:

[0024] R n =α(f-f0) β +σ max (3)

[0025] Among them, f0 represents the horizontal coordinate of the peak inflection point in the curve graph, σ max is the ordinate of the peak inflection point, that is, the measured data of the maximum RCS value, while α and β are parameters that determine the steepness of the power function curve.

[0026] Step 2: Obtain parameter estimates of the peak inflection point in the measured data curve

[0027] A search method based on the least squares meaning is described to obtain parameter estimates of the peak inflection points in the measured data curve graph. The specific steps are as follows:

[0028] 1) Select the data within the peak inflection point window

[0029] From the measurement data curve graph generated in step 1, determine a window containing a peak inflection point, and extract the RCS measurement data within the window.

[0030] 2) Set the search range and interval

[0031] Set the parameters α, β, f0 and σ max The search interval and traversal search interval.

[0032] 3) Parameter search and error calculation

[0033] In their respective search intervals, traverse the parameters α, β, f0 and σ max A set of possible values ​​is obtained by substituting them into the power function relationship to calculate the theoretical value at the corresponding frequency point. Then, the calculated value is subtracted from the measured value of the corresponding frequency point obtained in step 1 to obtain the error value. All frequency points in the window are traversed to obtain the error value corresponding to each frequency point, and the squares of all error values ​​are summed:

[0034]

[0035] 4) Parameter Optimization

[0036] According to the set search interval, the parameters α, β, f0 and σ are continuously changed in their respective search intervals. max, calculate the corresponding sum of squared error values. Select a set of parameters that minimizes the sum of squared error values ​​as the estimation result of the first peak inflection point parameter:

[0037]

[0038] Step 3: Inverse insect body length

[0039] Assuming that the elevation angle of the insect body axis relative to the hv plane of the radar antenna polarization coordinate system is θ (unit: °), the estimation formula of the insect body length can be expressed as follows:

[0040]

[0041] in, is the peak inflection point frequency of the measured data curve obtained by least squares estimation in step 2 The corresponding wavelength, a and b are quantities related to the pitch angle θ. Through a large number of simulation calculations, the calculation formulas of a and b are as follows:

[0042] a=0.0016θ 2 -0.0855θ+5.3313 (7)

[0043] b=-0.0454θ 2 +2.4273θ-137.3155 (8)

[0044] Therefore, the parameter estimate of the peak inflection point in the RCS-frequency measurement data curve in step 2 is The corresponding wavelength Substituting the inversion formula of insect body length at the current oblique flight angle, the estimated value of the body length of the measured oblique flying insect can be obtained.

[0045] Example:

[0046] In order to verify the above-mentioned method of inverting the body length of insects in non-level flight, a full-polarization scattering matrix measurement experiment of insects with different body inclination angles was carried out in a microwave darkroom using a multi-view full-polarization measurement device. Figure 1 The measurement equipment mainly includes full polarization measurement equipment and multi-view antenna support structure. Among them, the full polarization measurement equipment consists of a vector network analyzer and a dual-polarization antenna. The multi-view antenna support structure is shown in Figure 2As shown. The main body of the structure is a quarter-circular arc metal track, and other metal structures are mainly used to support the track. The radius of the large circle where the arc track is located is 2m. On the arc track, a pair of antenna bases (one for transmitting and one for receiving) are set every 15° at the center angle for installing dual-polarized antennas. If the position directly below the center of the large circle where the arc track is located is taken as the reference angle of the center angle (0°), the positions of the antenna bases are 0°, 15°, ..., 75° respectively. In the figure, the dual-polarized antenna and They represent the radar line of sight direction, H polarization direction and V polarization direction respectively.

[0047] For 23 insects (cotton bollworm, fall armyworm, cutworm, etc.), based on the broadband RCS measured data of the three bands of X, Ku and Ka, the body length estimation method of the insect in non-level flight posture described in the present invention is used to complete the inversion of their body length parameters when the pitch angles are 0° / 15° / 30° / 45° respectively. The inflection point information extracted from the 23 insects at different pitch angles is shown in Table 1.

[0048] Table 1 Results of extracting inflection point parameters at various oblique flight angles

[0049]

[0050] The body length inversion formulas for pitch angles of 0° / 15° / 30° / 45° are shown in Table 2. Next, the parameter estimates of the peak inflection point are The corresponding wavelength Substituting the body length inversion formula corresponding to each pitch angle, the corresponding body length estimation value can be obtained. The statistical diagram of the body length inversion error of all insects at different pitch angles is shown in Figure 3 As shown, the horizontal axis is the real body length of the insect, and the vertical axis is the relative error of the body length inversion.

[0051] Table 2 Inversion formula of body length at various oblique flight angles

[0052] angle Body length (unit: mg / dBsm) 0 l=5.4179λ-139.7168 15 l=4.1469λ-103.9118 30 l=4.4584λ-112.5812 45 l=4.6205λ-117.6440

[0053] The statistical results of the average error of body length inversion of 23 insects at different oblique flight angles are shown in the figure below. Figure 4 The statistical results are shown in Table 3.

[0054] Table 3 Average error of body length inversion at different oblique flight angles

[0055] angle Mean relative error 0 7.89％ 15 9.96％ 30 8.63％ 45 9.08％

[0056] Based on the inversion results of the above measured data, the following conclusions can be drawn:

[0057] For insects of different sizes, the method proposed in the present invention can effectively invert the insect body length at different oblique flight angles, and the maximum inversion error does not exceed 1 / 10 of the true value.

[0058] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An algorithm for inverting the body length parameters of insects in oblique flight posture, characterized in that: include: Step 1: Use a multi-band radar system to obtain radar scattering data of the insect to be tested at different frequencies, and further analyze the relationship between the frequency and the measured data of the maximum RCS value; Step 2: Based on the RCS-frequency measurement data curve obtained in step 1, a search method based on the least squares meaning is used to obtain a parameter estimation value of a peak inflection point in the RCS-frequency measurement data curve; Step 3: Use the wavelength corresponding to the peak inflection point in the RCS-frequency measurement data curve obtained in step 2 and the pitch angle of the insect body axis relative to the radar antenna polarization coordinate system plane to invert the body length of the oblique flying insect. The method is: use the pitch angle of the insect body axis relative to the radar antenna polarization coordinate system plane to obtain the insect body length inversion formula at the current oblique flight angle, and use the peak inflection point frequency in the measurement data curve obtained by least squares estimation in step 2 Corresponding wavelength Substitute into the inversion formula to obtain the estimated value of the body length of the measured insect; Assuming that the insect body axis points to an elevation angle of θ° relative to the hv plane of the radar antenna polarization coordinate system, the estimation formula for the insect body length can be expressed as follows: in, is the peak inflection point frequency of the measured data curve obtained by least squares estimation in step 2 The corresponding wavelength, a and b are quantities related to the pitch angle θ, and the calculation formulas for a and b are as follows: a=0.0016θ 2 -0.0855θ+5.3313 (2) b=-0.0454θ 2 +2.4273θ-137.3155 (3).

2. The inversion algorithm for the body length parameters of insects in oblique flight posture according to claim 1 is characterized by: In the step 1, it is assumed that a multi-band radar system is used to detect the insect to be detected, and scattering signals of N frequency points are obtained, and the number of frequencies in each band should be as large as possible; for the nth frequency point, the full polarization scattering matrix corresponding to the insect to be detected is obtained, which is expressed as: in, is the horizontal polarization component, is the vertical polarization component, and is the cross-polarization component; Substitute the full polarization scattering matrix into the following equation: The measured data R of the maximum RCS value at the nth frequency point can be obtained n ; Traverse N frequency points to obtain the measurement data of the RCS maximum value corresponding to each frequency point; take the frequency corresponding to each frequency point as the abscissa and the measurement data of the RCS maximum value as the ordinate to form an RCS-frequency measurement data curve; based on the RCS-frequency measurement data curve, obtain the power function relationship between the frequency and the measurement data of the RCS maximum value, which can be expressed by the following formula: R n =α(f-f0) β +σ max (6) Among them, f0 represents the horizontal coordinate of the peak inflection point in the curve graph, σ max is the ordinate of the peak inflection point, that is, the measured data of the maximum RCS value, while α and β are parameters that determine the steepness of the power function curve.

3. The inversion algorithm for the body length parameters of insects in oblique flight posture according to claim 1 is characterized by: In the step 2, a search method based on the least squares meaning is used to obtain the parameter estimation value of the peak inflection point in the measurement data curve graph; the specific steps are as follows: 1) Select the data within the peak inflection point window From the measurement data curve graph generated in step 1, determine a window containing a peak inflection point, and extract RCS measurement data within the window; 2) Set the search range and interval Set the parameters α, β, f0 and σ max The search interval and traversal search interval; 3) Parameter search and error calculation In their respective search intervals, traverse the parameters α, β, f0 and σ max A set of possible values ​​of is substituted into the power function relationship to calculate the theoretical value at the corresponding frequency point; then, the calculated value is subtracted from the measured value of the corresponding frequency point obtained in step 1 to obtain the error value; all frequency points in the window are traversed to obtain the error value corresponding to each frequency point, and the squares of all error values ​​are summed: Among them, J(α,β,f0,σ max ) is the objective function, which represents the sum of squares of errors under different parameter values; The frequency point response is calculated, which is related to the parameters α, β, f0, σ in this formula. max The relevant frequency response, that is, at the nth frequency point f n The estimated value of RCS under n (f n ) is the actual frequency response value in the observed data, that is, at the nth frequency point f n The measured value of RCS; 4) Parameter Optimization According to the set search interval, the parameters α, β, f0 and σ are continuously changed in their respective search intervals. max , calculate the corresponding sum of squared error values; select a set of parameters that minimizes the sum of squared error values ​​as the estimation result of the first peak inflection point parameter: Finally, according to the relationship between wavelength and frequency, Calculate the wavelength corresponding to the inflection point

Images