Terahertz SAR compensation method based on LM algorithm under complex platform vibration
By establishing a multi-component platform vibration model with time-varying amplitude in the terahertz SAR system and using the LM algorithm to compensate, the problem of poor focusability of imaging results caused by complex platform vibration is solved, and high-quality SAR imaging is achieved.
Patent Information
- Application Number
- CN202411015692.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-26
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-07-26
AI Technical Summary
Existing vibration models cannot accurately describe complex platform vibrations, resulting in poor focusability of terahertz SAR imaging results and difficulty in effective compensation.
Based on the LM algorithm, a multi-component platform vibration model with time-varying amplitude was established. The image focusing effect was evaluated through Tsallis entropy, and the Tsallis entropy minimization problem was transformed into a system of nonlinear equations. The Levenberg-Marquardt algorithm was used to solve it to build a compensation function to realize the platform. Accurate compensation for vibration.
Accurate description and compensation of complex platform vibrations is achieved, the focusability and image quality of terahertz SAR imaging results are improved, and the clarity of SAR images is ensured by minimizing Tsallis entropy.
Smart Images

Figure CN118938221A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of radar, and in particular relates to a terahertz SAR compensation method. Background Art
[0002] Terahertz SAR has a small wavelength and a large signal bandwidth, which enables it to obtain higher imaging resolution and better imaging details. Therefore, terahertz SAR imaging is gradually becoming a research hotspot. However, due to the small wavelength of terahertz SAR, a small platform vibration error can also cause severe defocusing along the azimuth direction, so it needs to be compensated to obtain a clearly focused imaging result. The premise of platform vibration error compensation is to establish a suitable and accurate model to describe it. Usually, the platform vibration is described as a superposition of multiple simple harmonic vibrations. However, the actual platform vibration is more complex, and there may be a situation where the amplitude changes with time. The complex platform vibration cannot be accurately described by the form of simple harmonic vibration alone. Complex platform vibration is a multi-component platform vibration with time-varying amplitude. In addition, complex platform vibration is difficult to compensate and will cause serious defocusing of terahertz SAR imaging results. Summary of the invention
[0003] The purpose of the present invention is to solve the problem that the existing vibration model cannot accurately describe the complex platform vibration, and the complex platform vibration is difficult to be compensated, resulting in poor focusing of the obtained terahertz SAR imaging results, and propose a terahertz SAR compensation method based on the LM algorithm under complex platform vibration.
[0004] The specific process of the terahertz SAR compensation method based on the LM algorithm under complex platform vibration is as follows:
[0005] Step 1: Establish a multi-component platform vibration model with time-varying amplitude;
[0006] It is expressed as;
[0007]
[0008] Among them, ΔR(m) is the platform vibration, a j (m) is the time-varying amplitude of the platform vibration, m is the discrete slow time, m = 0, 1, ..., M-1, M is the number of discrete slow times, f j is the vibration frequency of the platform, is the phase of platform vibration, j is the jth platform vibration component, and J is the total number of platform vibration components;
[0009] Step 2: Based on the multi-component platform vibration model with time-varying amplitude in step 1, a terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude is obtained;
[0010] Step 3: Process the terahertz SAR echo signal s(m,n) of step 2 to obtain the SAR imaging result, and calculate the Tsallis entropy of the SAR imaging result;
[0011] Step 4: Construct a compensation function to compensate for platform vibration and obtain a compensated SAR image;
[0012] Calculate the Tsallis entropy of the compensated SAR image, and use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image;
[0013] Step 5: Compensate for the platform vibration based on a compensation function that minimizes the Tsallis entropy of the compensated SAR image to obtain the final focused SAR imaging result.
[0014] The beneficial effects of the present invention are:
[0015] Based on the analysis of complex platform vibration, the present invention establishes a multi-component platform vibration model with time-varying amplitude, and studies the corresponding platform vibration compensation algorithm. The model can more accurately describe the complex and changeable platform vibration in the actual scene. The purpose of platform vibration compensation is to obtain imaging results with good image quality. Tsallis entropy is introduced to evaluate the image focusing effect. With the minimization of Tsallis entropy, the SAR image quality reaches the best, and the Tsallis entropy minimization problem is converted into a nonlinear equation group solution problem. Tsallis entropy is an extended form of Shannon entropy and is also a commonly used indicator to describe the quality of SAR images. When Tsallis entropy is the smallest, it can be considered that a well-focused imaging result is obtained. In addition, in the above-mentioned nonlinear equation group solution problem based on Tsallis entropy, how to solve the complex high-order derivatives of Tsallis entropy of SAR images and obtain the actual platform vibration error through iteration, which is also a research focus. In order to solve the above problems, the present invention introduces the LM (Levenberg-Marquardt) algorithm to solve the nonlinear equations. The LM algorithm can be regarded as an optimization method that combines the ideas of the gradient descent method and the Gauss-Newton method. By dynamically adjusting the step size to balance the advantages of the gradient descent and the Newton method, the LM algorithm can quickly and accurately construct a platform vibration compensation function. Finally, the platform vibration is compensated using the compensation function finally constructed to obtain a clearly focused SAR imaging result. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] Figure 1 It is a flow chart of the present invention;
[0017] Figure 2 This is the SAR image before platform vibration compensation in the 10dB scenario;
[0018] Figure 3 This is a SAR image after platform vibration compensation using the method proposed in the present invention in a 10 dB scenario;
[0019] Figure 4 This is the SAR image before platform vibration compensation in the -10dB scenario;
[0020] Figure 5 This is a SAR image after platform vibration compensation using the method proposed in the present invention in a -10dB scenario. DETAILED DESCRIPTION
[0021] Specific implementation method 1: Combination Figure 1 This embodiment is described. The specific process of the terahertz SAR compensation method based on the LM algorithm under complex platform vibration in this embodiment is as follows:
[0022] The purpose of the present invention is to compensate for the complex platform vibration of terahertz SAR based on the LM algorithm and obtain a focused imaging result. In the terahertz SAR system, the complex platform vibration will seriously affect the final imaging quality. Therefore, firstly, the complex characteristics of the platform vibration in the actual situation are considered, and a platform vibration model with time-varying amplitude is established. Secondly, the Tsallis entropy is introduced to describe the SAR image quality, and the Tsallis entropy minimization problem is converted into an equivalent nonlinear equation group solution problem. Finally, the complex high-order derivatives of the Tsallis entropy are solved, and the LM algorithm is introduced to iteratively obtain the focused SAR imaging result by dynamically adjusting the step size.
[0023] Step 1: Establish a multi-component platform vibration model with time-varying amplitude;
[0024] It is expressed as;
[0025]
[0026] Among them, ΔR(m) is the platform vibration, a j (m) is the time-varying amplitude of the platform vibration, m is the discrete slow time, m = 0, 1, ..., M-1, M is the number of discrete slow times, f j is the vibration frequency of the platform, is the phase of platform vibration, j is the jth platform vibration component, and J is the total number of platform vibration components;
[0027] Multi-component means there are J vibration components superimposed; a j (m) represents the change of amplitude over time m;
[0028] Step 2: Based on the multi-component platform vibration model with time-varying amplitude in step 1, a terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude is obtained;
[0029] Step 3: Process the terahertz SAR echo signal s(m,n) of step 2 to obtain the SAR imaging result that is defocused due to the platform vibration, and calculate the Tsallis entropy of the SAR imaging result;
[0030] Step 4: Construct a compensation function to compensate for platform vibration and obtain a compensated SAR image;
[0031] Calculate the Tsallis entropy of the compensated SAR image, and use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image;
[0032] Step 5: Compensate for the platform vibration based on the compensation function that minimizes the Tsallis entropy of the compensated SAR image to obtain the final focused SAR imaging result (substitute the compensation function that minimizes the Tsallis entropy of the compensated SAR image into formula (7) to obtain the final focused SAR imaging result).
[0033] Specific implementation method 2: This implementation method is different from specific implementation method 1 in that, in step 2, based on the multi-component platform vibration model with time-varying amplitude in step 1, a terahertz SAR echo signal s(m,n) of multi-component platform vibration with time-varying amplitude is obtained;
[0034] The specific process is:
[0035] Assume there is a target point in the scene, and the coordinates of the target point are (x, y, 0);
[0036] According to the multi-component platform vibration model with time-varying amplitude established in step 1, the terahertz SAR echo signal is range compressed, and the range migration correction is performed on the terahertz SAR echo signal after the range compression to obtain the terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude;
[0037] The expression of the terahertz SAR echo signal s(n,m) of multi-component platform vibration with time-varying amplitude is:
[0038]
[0039] Where s(m,n) is the terahertz SAR echo signal of multi-component platform vibration with time-varying amplitude, n is the discrete fast time, n = 0, 1, ..., N-1, N is the number of discrete fast times; ξ is the complex amplitude; B is the signal bandwidth; R0 is the shortest slant distance from the radar platform to the target; c is the speed of light; j is the imaginary unit, j 2 =-1; k a is the Doppler modulation frequency; v is the platform velocity; λ is the wavelength; ΔR is the platform vibration.
[0040] The other steps and parameters are the same as those in the first embodiment.
[0041] Specific implementation method three: This implementation method is different from specific implementation method one or two in that in step three, the terahertz SAR echo signal s(m,n) of step two is processed to obtain a SAR imaging result that is defocused due to the presence of platform vibration, and the Tsallis entropy of the SAR imaging result is calculated;
[0042] The specific process is:
[0043] Step 31: Compensate the terahertz SAR echo signal s(m,n) in step 2 in azimuth to obtain a signal s0(m,n) after azimuth compensation;
[0044] The expression is:
[0045] s0(m,n)=s(m,n)×exp(-jπk a m 2 ) (3)
[0046] Step 32: Perform Fourier transform on the azimuth-compensated signal s0(n,m) along the azimuth direction to obtain the SAR imaging result g0(h,n) defocused due to the platform vibration;
[0047] The expression is:
[0048]
[0049] Where, h is the Doppler frequency index;
[0050] Step 33: Calculate the Tsallis entropy T0 of the defocused SAR imaging result g0(h,n);
[0051] The expression is:
[0052]
[0053] Among them, α is a real number; S0 is an intermediate variable,
[0054] The other steps and parameters are the same as those in the first or second embodiment.
[0055] Specific implementation method 4: This implementation method is different from any one of specific implementation methods 1 to 3 in that in step 4, a compensation function is constructed to compensate for platform vibration to obtain a compensated SAR image;
[0056] Calculate the Tsallis entropy of the compensated SAR image, and use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image;
[0057] The specific process is:
[0058] Step 41: Construct compensation function;
[0059] The compensation function expression is:
[0060] H=exp(jΔΦ(m)) (6)
[0061] Wherein, ΔΦ(m) is the phase error caused by platform vibration. The phase error is the error introduced by platform vibration ΔR(m) in the phase of the terahertz SAR echo signal s(m,n), which is the main factor causing SAR image defocusing; H is the compensation function;
[0062] Step 42: Compensate the platform vibration based on the compensation function H to obtain the compensated SAR image g(h,n);
[0063] The expression is:
[0064]
[0065] Step 43: Calculate the Tsallis entropy of the compensated SAR image g(h,n);
[0066] The expression is:
[0067]
[0068] Where T is the Tsallis entropy of the compensated SAR image g(h,n); S is an intermediate variable,
[0069] Step 44: Use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image g(h,n).
[0070] The other steps and parameters are the same as those in Specific Embodiments 1 to 3.
[0071] Specific implementation method five: This implementation method is different from specific implementation methods one to four in that the LM algorithm is used in step four to find a compensation function that minimizes the Tsallis entropy of the compensated SAR image g(h,n);
[0072] The specific process is:
[0073] Step 441: Convert the Tsallis entropy minimization problem into an equivalent nonlinear equations solution problem;
[0074] The expression is:
[0075]
[0076] Step 442: Calculate the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n);
[0077] Step 443: Initialize phase error and maximum number of cycles K;
[0078] Let the number of loops k = 1;
[0079] Step 444: Based on the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n), the LM algorithm is used to solve the nonlinear equations (Formula (9)) to obtain the updated phase error ΔΦ k (m);
[0080] Step 445: Use the updated phase error ΔΦ k (m) Construct the compensation function H for the kth iteration k The signal s0(m,n) is compensated for platform vibration to obtain the compensated SAR imaging result g(h,n); based on the compensated SAR imaging result g(h,n), the Tsallis entropy T of the kth iterative compensated SAR image is calculated. k ;
[0081] Step 446: According to the Tsallis entropy T of the compensated SAR image k Adjust the step size μ of the LM algorithm k ;
[0082] Step 447: If the loop termination condition is met or the maximum number of loops is reached, the algorithm ends and the compensation function H is found to minimize the Tsallis entropy of the compensated SAR image g(h,n) k ; Otherwise, let k=k+1 and repeat steps 444 to 447.
[0083] The other steps and parameters are the same as those in Specific Embodiments 1 to 4.
[0084] Specific implementation method 6: This implementation method is different from specific implementation methods 1 to 5 in that the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n) are solved in step 442;
[0085] The specific process is:
[0086] Step 4421: Calculate the first-order partial derivative F(ΔΦ(m)) of the Tsallis entropy of the compensated SAR image g(h,n); the expression is:
[0087]
[0088] Where, F(ΔΦ(m)) represents the first-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n); Im[] represents the operation of taking the imaginary part; conj(·) represents the conjugate operation; IFT(·) represents the inverse Fourier transform operation;
[0089] Step 4422: Solve the second-order partial derivative J(ΔΦ(m)) of the Tsallis entropy of the compensated SAR image g(h,n); the expression is:
[0090]
[0091] Among them, T_s1, T_s2, and T_s3 represent intermediate variables;
[0092] The expressions of intermediate variables T_s1, T_s2, and T_s3 are:
[0093]
[0094] Among them, Re{} represents the real part operation; g * (h,n) represents the conjugate of g(h,n); [s0(m,n)] * represents the conjugate of s0(m,n).
[0095] The other steps and parameters are the same as those in Specific Implementation Methods 1 to 5.
[0096] Specific implementation method 7: This implementation method is different from any one of the specific implementation methods 1 to 6 in that in step 444, the first-order and second-order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n) are used to solve the nonlinear equations using the LM algorithm to obtain the updated phase error ΔΦ k (m);
[0097] The specific process is:
[0098] Step 4441: Calculate the search direction s in the kth iteration k-1 (m); the expression is:
[0099] s k-1 (m) = -(J T (ΔΦ k-1 (m))J(ΔΦ k-1 (m))+μ k-1 I) -1 J T (ΔΦ k-1 (m))F(ΔΦ k-1 (m))(15)
[0100] Among them, J(ΔΦ k-1(m)) represents the second-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n) obtained at the k-1th iteration; the superscript T represents the matrix transpose, ΔΦ k-1 (m) represents the phase error caused by platform vibration obtained in the k-1th iteration, μ k-1 represents the step size corresponding to the k-1th iteration, μ k-1 >0; I is the unit matrix, F(ΔΦ k-1 (m)) represents the first-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n) obtained at the k-1th iteration;
[0101] Step 4442: Use search direction s k-1 (m) Update the phase error to obtain the updated phase error ΔΦ k (m); the expression is:
[0102] ΔΦ k (m) = ΔΦ k-1 (m)+s k-1 (m) (16)
[0103] Among them, ΔΦ k (m) represents the phase error caused by platform vibration obtained at the kth iteration.
[0104] The other steps and parameters are the same as those in Specific Embodiments 1 to 6.
[0105] Specific implementation eight: This implementation differs from any one of the specific implementations one to seven in that the updated phase error ΔΦ is used in step four or four. k (m) Construct the compensation function H for the kth iteration k The signal s0(m,n) is compensated for platform vibration to obtain the compensated SAR imaging result g(h,n); based on the compensated SAR imaging result g(h,n), the Tsallis entropy T of the kth iterative compensated SAR image is calculated. k ;
[0106] The specific process is:
[0107] The updated phase error ΔΦ k Substitute (m) into formula (6) to obtain the compensation function H of the kth iteration k ;
[0108] The compensation function H of the kth iteration k Substitute into formula (7) to obtain the k-th iteratively compensated SAR image g(h,n);
[0109] Substitute the k-th iteratively compensated SAR image g(h,n) into formula (8) to obtain the Tsallis entropy T of the k-th iteratively compensated SAR image g(h,n): k .
[0110] The other steps and parameters are the same as those in Specific Embodiments 1 to 7.
[0111] Specific implementation method 9: This implementation method is different from any one of the specific implementation methods 1 to 8 in that in step 446, the Tsallis entropy T of the compensated SAR image is calculated based on the Tsallis entropy T of the compensated SAR image. k Adjust the step size μ of the LM algorithm k ;
[0112] The specific process is:
[0113] If T k ≥T k-1 , then μ k =μ k-1 ×δ, and ΔΦ k (m) = ΔΦ k-1 (m), T k =T k-1 ;
[0114] If T k <T k-1 , then μ k =μ k-1 / β;
[0115] Among them, δ and β are constants, δ>1, β>1.
[0116] The other steps and parameters are the same as those in Specific Embodiments 1 to 8.
[0117] Specific implementation method 10: This implementation method is different from the specific implementation methods 1 to 9 in that the loop termination condition in step 447 is:
[0118] |T k -T k-1 |≤threshold (17)
[0119] Here, threshold indicates the threshold value.
[0120] The other steps and parameters are the same as those in Specific Embodiments 1 to 9.
[0121] Embodiment 1:
[0122] In this embodiment, a terahertz SAR simulation system with a carrier frequency of 220 GHz and a bandwidth of 3 GHz is first constructed under the condition of a signal-to-noise ratio of 10 dB. At the same time, a multi-component platform vibration with time-varying amplitude is established, and a terahertz SAR echo signal is constructed on this basis. The echo signal is range compressed and range migration correction is performed, and then Fourier transform is performed on it along the azimuth to obtain a defocused SAR echo image, such as Figure 2 As shown. It can be seen that Figure 2 The SAR image in is severely defocused in azimuth, which is caused by the complex time-varying amplitude multi-component platform vibration. If you want to suppress the azimuth defocus and obtain a focused terahertz SAR image, you need to compensate for the platform vibration. Therefore, we use the LM-based terahertz SAR compensation algorithm proposed in this invention to compensate for the terahertz platform vibration. Through iteration, the final compensation function can be obtained to compensate for the platform vibration and obtain a focused SAR image, such as Figure 3 As shown. It can be seen that after the terahertz SAR compensation algorithm based on the LM algorithm proposed in the present invention is processed, the platform vibration is well compensated and the image is focused. In addition, the Tsallis entropy of the SAR image obtained before and after the platform vibration compensation under the condition of 10 dB is shown in Table 1. It can be seen that the reduction of Tsallis entropy further proves the effectiveness of the algorithm proposed in the present invention.
[0123] Table 1 Tsallis entropy of SAR image before and after platform vibration compensation under the condition of 10dB signal-to-noise ratio
[0124]
[0125] Embodiment 2
[0126] In this embodiment, a terahertz SAR simulation system with a carrier frequency of 220 GHz and a bandwidth of 3 GHz is constructed under the condition of a signal-to-noise ratio of -10 dB. At the same time, a multi-component platform vibration with time-varying amplitude is established, and a terahertz SAR echo signal is constructed on this basis. The terahertz SAR echo signal containing complex platform vibration is processed and Fourier transformed along the azimuth direction to obtain a defocused SAR image, such as Figure 4 As shown. Due to the influence of multi-component platform vibration with time-varying amplitude, a clearly focused terahertz SAR image cannot be obtained. Therefore, we use the LM-based platform vibration compensation algorithm proposed in the present invention to compensate. Through iteration, the phase error introduced by the platform vibration can be estimated, and a compensation function can be constructed to compensate for the platform vibration. The terahertz SAR image after compensation by the algorithm proposed in the present invention is shown as follows: Figure 5As shown in the figure, it can be seen that the image defocus is well suppressed after compensation by the platform vibration compensation algorithm based on LM, and the point target in the image is clearly visible. In addition, Table 2 shows the Tsallis entropy results of the image before and after compensation by the algorithm proposed by the present invention under the condition of -10dB. It can be seen that the reduction of entropy has proved the effectiveness of the algorithm proposed by the present invention.
[0127] Table 2 Tsallis entropy of SAR images before and after platform vibration compensation under the condition of -10dB signal-to-noise ratio
[0128]
[0129] The present invention may also have many other embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art may make various corresponding changes and modifications based on the present invention, but these corresponding changes and modifications should all fall within the scope of protection of the claims attached to the present invention.
Claims
1. Terahertz SAR compensation method based on LM algorithm under complex platform vibration, Features: The specific process of the method is: Step 1: Establish a multi-component platform vibration model with time-varying amplitude; It is expressed as: Among them, ΔR(m) is the platform vibration, a j (m) is the time-varying amplitude of the platform vibration, m is the discrete slow time, m = 0, 1, ..., M-1, M is the number of discrete slow times, f j is the vibration frequency of the platform, is the phase of platform vibration, j is the jth platform vibration component, and J is the total number of platform vibration components; Step 2: Based on the multi-component platform vibration model with time-varying amplitude in step 1, a terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude is obtained; Step 3: Process the terahertz SAR echo signal s(m,n) of step 2 to obtain the SAR imaging result, and calculate the Tsallis entropy of the SAR imaging result; Step 4: Construct a compensation function to compensate for platform vibration and obtain a compensated SAR image; Calculate the Tsallis entropy of the compensated SAR image, and use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image; Step 5: Compensate for the platform vibration based on a compensation function that minimizes the Tsallis entropy of the compensated SAR image to obtain a final focused SAR imaging result.
2. According to claim 1, the terahertz SAR compensation method based on LM algorithm under complex platform vibration, Features: In the step 2, based on the multi-component platform vibration model with time-varying amplitude in step 1, a terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude is obtained; The specific process is: Assume there is a target point in the scene, and the coordinates of the target point are (x, y, 0); According to the multi-component platform vibration model with time-varying amplitude established in step 1, the terahertz SAR echo signal is range compressed, and the range migration correction is performed on the terahertz SAR echo signal after the range compression to obtain the terahertz SAR echo signal s(m,n) of the multi-component platform vibration with time-varying amplitude; The expression of the terahertz SAR echo signal s(n,m) of multi-component platform vibration with time-varying amplitude is: Where s(m,n) is the terahertz SAR echo signal of multi-component platform vibration with time-varying amplitude, n is the discrete fast time, n = 0, 1, ..., N-1, N is the number of discrete fast times; ξ is the complex amplitude; B is the signal bandwidth; R 0 is the shortest slant distance from the radar platform to the target; c is the speed of light; j is the imaginary unit, j 2 =-1; k a is the Doppler modulation frequency; v is the platform velocity; λ is the wavelength; ΔR is the platform vibration.
3. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 2, Features: In the step 3, the terahertz SAR echo signal s(m,n) of the step 2 is processed to obtain a SAR imaging result, and the Tsallis entropy of the SAR imaging result is calculated; The specific process is: Step 31: Compensate the terahertz SAR echo signal s(m,n) in step 2 along the azimuth direction to obtain the azimuth compensated signal s 0 (m,n); The expression is: s 0 (m,n)=s(m,n)×exp(-jπk a m 2 ) (3) Step 32: Compensate the signal s for the azimuth 0 (n,m) performs Fourier transform along the azimuth direction to obtain the SAR imaging result g 0 (h,n); The expression is: Where, h is the Doppler frequency index; Step 33: Calculate the SAR imaging result g 0 Tsallis entropy T of (h,n) 0 ; The expression is: Among them, α is a real number; S 0 is the intermediate variable, 4. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 3, Features: In the step 4, a compensation function is constructed to compensate for the platform vibration to obtain a compensated SAR image; Calculate the Tsallis entropy of the compensated SAR image, and use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image; The specific process is: Step 41: Construct compensation function; The compensation function expression is: H=exp(jΔΦ(m)) (6) Where ΔΦ(m) is the phase error caused by platform vibration; H is the compensation function; Step 42: Compensate the platform vibration based on the compensation function H to obtain the compensated SAR image g(h,n); The expression is: Step 43: Calculate the Tsallis entropy of the compensated SAR image g(h,n); The expression is: Where T is the Tsallis entropy of the compensated SAR image g(h,n); S is an intermediate variable, Step 44: Use the LM algorithm to find the compensation function that minimizes the Tsallis entropy of the compensated SAR image g(h,n).
5. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 4, Features: In the step 44, the LM algorithm is used to find a compensation function that minimizes the Tsallis entropy of the compensated SAR image g(h,n); The specific process is: Step 441: Convert the Tsallis entropy minimization problem into an equivalent nonlinear equations solution problem; The expression is: Step 442: Calculate the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n); Step 443: Initialize phase error and maximum number of cycles K; Let the number of loops k = 1; Step 444: Based on the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n), the LM algorithm is used to solve the nonlinear equations to obtain the updated phase error ΔΦ k (m); Step 445: Use the updated phase error ΔΦ k (m) Construct the compensation function H for the kth iteration k For signal s 0 (m,n) compensates for the platform vibration and obtains the compensated SAR imaging result g(h,n); based on the compensated SAR imaging result g(h,n), calculates the Tsallis entropy T of the kth iteration compensated SAR image k ; Step 446: According to the Tsallis entropy T of the compensated SAR image k Adjust the step size μ of the LM algorithm k ; Step 447: If the loop termination condition is met or the maximum number of loops is reached, the algorithm ends and the compensation function H is found to minimize the Tsallis entropy of the compensated SAR image g(h,n) k ; Otherwise, let k=k+1 and repeat steps 444 to 447.
6. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 5, Features: In the step 442, the first and second order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n) are solved; The specific process is: Step 4421: Calculate the first-order partial derivative F(ΔΦ(m)) of the Tsallis entropy of the compensated SAR image g(h,n); the expression is: Where, F(ΔΦ(m)) represents the first-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n); Im[] represents the operation of taking the imaginary part; conj(·) represents the conjugate operation; IFT(·) represents the inverse Fourier transform operation; Step 4422: Solve the second-order partial derivative J(ΔΦ(m)) of the Tsallis entropy of the compensated SAR image g(h,n); the expression is: Among them, T_s 1 、T_s 2 、T_s 3 represents an intermediate variable; Intermediate variable T_s 1 、T_s 2 、T_s 3 The expression is: Among them, Re{} represents the real part operation; g * (h,n) represents the conjugate of g(h,n); [s 0 (m,n)] * Indicates 0 The conjugate of (m,n).
7. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 6, Features: In step 444, based on the first-order and second-order partial derivatives of the Tsallis entropy of the compensated SAR image g(h,n), the LM algorithm is used to solve the nonlinear equations to obtain the updated phase error ΔΦ k (m); The specific process is: Step 4441: Calculate the search direction s in the kth iteration k-1 (m); the expression is: s k-1 (m)=-(J T (DF k-1 (m))J(ΔΦ k-1 (m))+μ k-1 I) -1 J T (DF k-1 (m))F(ΔΦ k-1 (m)) (15) Among them, J(ΔΦ k-1 (m)) represents the second-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n) obtained at the k-1th iteration; the superscript T represents the matrix transpose, ΔΦ k-1 (m) represents the phase error caused by platform vibration obtained in the k-1th iteration, μ k-1 represents the step size corresponding to the k-1th iteration, μ k-1 >0; I is the unit matrix, F(ΔΦ k-1 (m)) represents the first-order partial derivative of the Tsallis entropy of the compensated SAR image g(h,n) obtained at the k-1th iteration; Step 4442: Use search direction s k-1 (m) Update the phase error to obtain the updated phase error ΔΦ k (m); the expression is: ΔΦ k (m)=ΔΦ k-1 (m)+s k-1 (m) (16) Among them, ΔΦ k (m) represents the phase error caused by platform vibration obtained at the kth iteration.
8. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 7, Features: In step 445, the updated phase error ΔΦ is used k (m) Construct the compensation function H for the kth iteration k For signal s 0 (m,n) compensates for the platform vibration and obtains the compensated SAR imaging result g(h,n); based on the compensated SAR imaging result g(h,n), calculates the Tsallis entropy T of the kth iteration compensated SAR image k ; The specific process is: The updated phase error ΔΦ k Substitute (m) into formula (6) to obtain the compensation function H of the kth iteration k ; The compensation function H of the kth iteration k Substitute into formula (7) to obtain the k-th iteratively compensated SAR image g(h,n); Substitute the k-th iteratively compensated SAR image g(h,n) into formula (8) to obtain the Tsallis entropy T of the k-th iteratively compensated SAR image g(h,n): k .
9. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 8, Features: In the step 446, the Tsallis entropy T of the compensated SAR image is calculated based on the k Adjust the step size μ of the LM algorithm k ; The specific process is: If T k ≥T k-1 , then μ k = μ k-1 ×δ, and ΔΦ k (m) = ΔΦ k-1 (m), T k = T k-1 ; If T k <T k-1 , then μ k = μ k-1 / β; Among them, δ and β are constants, δ>1, β>1.
10. The terahertz SAR compensation method based on LM algorithm under complex platform vibration according to claim 9, Features: The loop termination condition in step 447 is: |T k -T k-1 |≤threshold (17) Here, threshold indicates the threshold value.
Citation Information
Patent Citations
Small target MRI image segmentation method of minimum cross Tsallis entropy
CN105261009A
Imaging method and device for moving target of synthetic aperture radar (SAR) based on iterative minimum entropy
CN109298420A
Threshold segmentation method based on quick iteration of two-dimensional Tsallis gray entropy
CN111553926A
Maneuvering weak target detection method based on dynamic planning and minimum image entropy
CN111965614A
SAR (Synthetic Aperture Radar) platform high-frequency vibration compensation method under terahertz wave band
CN115792839A
Cited By
ISARadar Image Autofocus Correction System
KR102981375B1
