Improved high orbit sar adaptive error estimation and compensation method

Abstract

This invention discloses an improved adaptive error estimation and compensation method for high-orbit SAR, which can effectively estimate and compensate for amplitude and phase errors caused by non-ideal factors, and solve the spatial variation error caused by large imaging stripes in high-orbit SAR, achieving focused imaging. The method includes the following steps: A GEO SAR satellite images the imaging area, and the echo data of the imaging area is processed into two-dimensional blocks to obtain m two-dimensional sub-blocks. The total synthetic aperture is divided into n sub-apertures. The SPECAN algorithm is used to image each sub-aperture and two-dimensional sub-block, resulting in n×m images. Discrete windowing is applied to the n×m image data, followed by PGA correction to obtain the amplitude and phase errors corresponding to each image. Error fusion is performed on the amplitude and phase errors corresponding to each sub-aperture image. After full-aperture error compensation, full-aperture imaging is performed on each two-dimensional sub-block, and then the images are stitched together to obtain the final imaging result of the GEO SAR for the entire imaging area.

Description

An Improved Adaptive Error Estimation and Compensation Method for High-Orbit SAR Technical Field

[0001] This invention relates to the field of synthetic aperture radar technology, and more specifically to an improved adaptive error estimation and compensation method for high-orbit SAR. Background Technology

[0002] Geosynchronous Synthetic Aperture Radar (GEO SAR) is a high-orbit SAR operating in geosynchronous orbit. It boasts advantages such as wide image acquisition swath, short revisit time, and strong weather adaptability, and has received widespread attention and application in recent years. However, due to the extremely high orbital altitude and extremely long synthetic aperture time of GEO SAR, the imaging process is inevitably affected by non-ideal factors. For practical GEO SAR systems, these non-ideal factors specifically include:

[0003] (1) Non-ideal atmosphere. Due to the extremely high altitude of GEO SAR orbits, the signal travels through the entire atmosphere in its round-trip path, so the impact of non-ideal atmosphere on signal transmission should be considered. The effects of non-ideal atmosphere generally include the effects of the ionosphere and troposphere. Considering that most GEO SARs operate in the L-band, the effect of the troposphere can be ignored. Therefore, only the effect of the ionosphere on GEO SAR signals needs to be considered. The errors generated by the ionosphere come from two aspects: first, ionospheric scintillation causes random amplitude and phase fluctuations, affecting the sidelobes of the SAR image; second, the background ionosphere causes polynomial phase errors, affecting image quality and even causing the image to defocus.

[0004] (2) Orbital perturbation. In actual operation, GEO SAR satellites are subject to orbital perturbation errors caused by factors such as non-spherical gravity, lunar and solar perturbations, and solar radiation pressure, which in turn affect imaging quality. The main orbital parameters include semi-major axis, eccentricity, orbital inclination, and longitude of the ascending node. Orbital perturbation can be described as fluctuations in the above orbital parameters, which will produce Doppler parameter errors and affect imaging quality.

[0005] (3) Antenna vibration. GEO SAR antennas can be several meters or even longer, and antenna vibration during long-term focusing cannot be eliminated. Antenna vibration mainly includes two parts: rotational vibration and translational vibration, which can cause slant range error and antenna gain error, thus affecting imaging quality.

[0006] In summary, non-ideal factors such as atmospheric conditions, orbital perturbations, and antenna vibrations have a significant impact on GEO SAR imaging. Furthermore, due to the very large imaging bands in GEO SAR, these non-ideal factors also introduce severe spatially varying errors. To estimate and compensate for GEO SAR errors, methods such as joint amplitude and phase compensation algorithms based on SPECAN and minimum entropy, and autofocusing based on interferometry, have been proposed. However, existing compensation algorithms cannot efficiently and accurately compensate for the amplitude and phase errors caused by these non-ideal factors, thus failing to achieve high-quality focused imaging. Summary of the Invention

[0007] In view of this, the present invention provides an improved adaptive error estimation and compensation method for high-orbit SAR, which can effectively estimate and compensate for amplitude and phase errors caused by non-ideal factors, and solve the spatial variation error caused by the large imaging stripe in high-orbit SAR, thereby achieving focused imaging.

[0008] To achieve the above objectives, the technical solution of the present invention includes the following steps:

[0009] S1: The GEO SAR satellite images the imaging area and acquires echo data; the echo data of the imaging area is processed into two-dimensional blocks to obtain m two-dimensional sub-blocks.

[0010] S2: Divide the total synthetic aperture into n sub-apertures, with the sub-aperture length set to a predetermined value.

[0011] S3: Use the SPECAN algorithm to image each sub-aperture and two-dimensional sub-block to obtain n×m images.

[0012] S4: Discrete windowing is applied to n×m image data. The amplitude and phase errors of each image after discrete windowing are corrected by PGA to obtain the amplitude error and phase error corresponding to each image.

[0013] S5: Perform error fusion on the amplitude and phase errors corresponding to each sub-aperture image to obtain the amplitude and phase errors of the full aperture. After full aperture error compensation, perform full aperture imaging on each two-dimensional sub-block, and then stitch the images together to obtain the final imaging result of GEO SAR for the entire imaging area.

[0014] Preferably, the echo data of the imaging area is subjected to two-dimensional block processing to obtain m two-dimensional sub-blocks. Specifically, the echo data of the imaging area is divided equally on a two-dimensional plane to obtain m equal two-dimensional sub-blocks.

[0015] Preferably, the fully synthesized aperture is divided into n sub-apertures, and the length of each sub-aperture is a set value. Specifically, the sub-aperture length is initially set, and the fully synthesized aperture is divided into n sub-apertures on an average basis. Imaging processing is performed on the divided sub-apertures. If the imaging processing fails, the length of each sub-aperture is reduced to 80% of its original length until the imaging processing is successful.

[0016] Preferably, the sub-aperture length is a set value, specifically: the sub-aperture length is set to 10% of the total aperture length.

[0017] Preferably, there is an overlap between adjacent sub-apertures, with the overlapping portion accounting for 30% to 60% of the sub-aperture.

[0018] Preferably, discrete windowing is applied to n×m image data, specifically by multiplying the image data by a series of discrete windows. The width of the discrete windows is set to the superposition of multiple rectangular windows, with each rectangular window having a width 10 times the resolution of the image. The center of each window is located at the main lobe and the first three pairs of paired echoes.

[0019] Beneficial effects:

[0020] This invention provides an improved adaptive error estimation and compensation method for high-orbit SAR. It effectively addresses the combined errors caused by non-ideal atmospheric conditions, orbital perturbations, and antenna vibrations during high-orbit SAR imaging by employing two-dimensional block partitioning, aperture division, discrete windowing, error compensation, and fusion, ultimately achieving focused imaging. Real-world imaging results demonstrate that the proposed method significantly improves imaging quality compared to the uncompensated method, proving its feasibility and superior performance. Attached Figure Description

[0021] Figure 1 is a schematic diagram of the GEO SAR imaging geometry of the present invention;

[0022] Figure 2 is a schematic diagram of the GEO SAR imaging geometry considering non-ideal factors in this invention;

[0023] Figure 3 is a flowchart of the technical solution proposed in this invention;

[0024] Figure 4 is a schematic diagram of the sub-aperture division steps of the present invention;

[0025] Figure 5 is a schematic diagram of the discrete windowing steps of the present invention;

[0026] Figure 6 is the original reference image in a specific embodiment of the present invention;

[0027] Figure 7 shows the imaging results without error compensation in a specific embodiment of the present invention;

[0028] Figure 8 shows the imaging results after error compensation using this method in a specific embodiment of the present invention. Detailed Implementation

[0029] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0030] The geometry of GEO SAR is shown in Figure 1. The origin of the coordinate system is located at the center of the illuminated area. The y-axis is parallel to the velocity vector of the geostationary satellite, the z-axis is perpendicular to the Earth's surface, and the x-axis is determined by the right-hand rule. In this coordinate system, the motion of the geostationary satellite can be modeled as a vector function of slow time. Simultaneously, for a target T in a given scene, its historical slant range can be calculated and expanded using a Taylor series. Without considering non-ideal factors, the GEO SAR echo can be modeled in the two-dimensional time domain as s ideal (t r ,t a ), where t r and t a These represent fast time and slow time, respectively.

[0031] However, in practical applications, non-ideal factors such as imperfect atmosphere, orbital perturbations, and antenna vibrations are unavoidable. The effects of these non-ideal factors are shown in Figure 2. We will now discuss the errors caused by these non-ideal factors to GEO SAR signals.

[0032] Non-ideal atmosphere: Non-ideal atmospheric factors generally include the effects of the ionosphere and troposphere on the signal. For most GEO SARs operating in the L-band, the tropospheric effect can be ignored. Ionospheric effects mainly include background ionospheric delay and ionospheric scintillation. Regarding background ionospheric delay, when the GEO SAR signal passes through the ionosphere, the additional delay related to the background ionospheric electron content will accumulate in the slant range, thus introducing phase error. For GEO SAR with extremely long aperture times, the total electron content (TEC) will not remain constant throughout the imaging process and should be modeled as a function of slow time. Furthermore, due to the wide swath of GEO SAR, the spatial variation of TEC cannot be ignored and should be modeled as a function of target position. In summary, the background ionosphere produces a Δφ effect on the GEO SAR echo signal. BI The additional phase. For ionospheric scintillation, ionospheric scintillation causes fluctuations in signal amplitude and phase, and the error can be modeled as slowly time-varying random amplitude and random phase. Typically, for targets at different locations in a scene, if the corresponding distance between targets is greater than the decorrelation distance (usually several kilometers), then spatial variation within a region can be ignored. Based on the above analysis, the additional phase and amplitude of the ionospheric scintillation background ionosphere on the GEOSAR echo signal can be expressed as δ IS (t a ) and φ IS (t a ).

[0033] Orbital perturbation: The impact of orbital perturbation on echo signals is similar to that of the background ionosphere. The orbit can be described by orbital parameters such as semi-major axis, eccentricity, dip angle, and the longitude of the rising node. Orbital perturbation introduces parameter fluctuations, which in turn introduce Doppler parameter errors. Based on the above analysis, the additional phase introduced by orbital perturbation into the GEO SAR echo signal can be expressed as Δφ. OP

[0034] Antenna vibration: The effects of antenna vibration mainly include rotational vibration and translational vibration. For antenna translation, at extremely high GEO SAR orbital altitudes, we only need to consider single-frequency periodic antenna translation. Antenna translation introduces a slant range error, which can be expressed as ΔR. AT It also introduces additional phase error into the GEO SAR echo signal, which can be expressed as: Where A T f T and These represent the amplitude, magnitude, and phase of the periodic error introduced by antenna vibration. For antenna rotation, its main effect is on the direction of the illumination beam, thus leading to gain error. Generally, the antenna rotation angle affecting antenna gain can be modeled as a slow-time function. Therefore, the error term introduced by antenna rotation to the GEO SAR echo signal can be expressed as g. AR (t a ).

[0035] Based on the discussion of the various non-ideal factors that cause errors in GEO SAR signal echoes, the actual GEO SAR echo signal can be expressed as follows: Where g E (t a )and These represent the total amplitude and phase error of the signal caused by each non-ideal factor, which is the superposition of the various error terms mentioned above.

[0036] To address the impact of various non-ideal factors on GEO SAR imaging, this invention proposes an improved adaptive error estimation and compensation method for high-orbit SAR. The flowchart of the method can be summarized in Figure 3.

[0037] The following section will detail the specific process of an improved adaptive error estimation and compensation method for high-orbit SAR proposed in this invention.

[0038] Step 1: The GEO SAR satellite images the imaging area and acquires echo data; the echo data of the imaging area is processed into two-dimensional blocks to obtain m two-dimensional sub-blocks.

[0039] GEO SAR imaging faces a severe spatial variation problem because the TEC (Effective Chromaticity) varies across the entire scene. The resulting phase error variance degrades the performance of autofocus methods. Therefore, the control and resolution problem must be addressed before autofocus processing. Two-dimensional segmentation is an effective method for solving this problem. This method divides a very wide scene into several smaller blocks, eliminating spatial variation within each block. The size of the two-dimensional blocks can be adjusted and limited based on the overall higher-order phase error. In this embodiment, the echo data of the imaging area is evenly divided on a two-dimensional plane to obtain m equal two-dimensional sub-blocks.

[0040] Step 2: Divide the total synthetic aperture into n sub-apertures, with the sub-aperture length set to a predetermined value.

[0041] The GEO SAR satellite operates in an orbit at an altitude of 36,000 kilometers, which results in a suboptimal noise equivalent scattering coefficient (NESZ). With synthetic aperture times reaching hundreds of seconds, the low signal-to-noise ratio and large errors can lead to complete defocusing of strong points, causing phase error estimation failures in strong-point-related autofocusing methods such as the PGA algorithm.

[0042] This method addresses the problem of effective focusing by dividing the entire synthetic aperture into several short sub-apertures. As shown in Figure 4, within each sub-aperture, the phase error is limited to an appropriate range, ensuring that the defocusing of strong points is not too severe. In this way, the SPECAN algorithm can effectively complete the original imaging and estimate the phase error.

[0043] In this embodiment of the invention, the length of each sub-aperture is initially set to be the same, and can be set to 10-30 seconds, or about 10% of the entire aperture. Then, imaging processing is performed on the divided sub-apertures. If the imaging processing fails, the length of the sub-aperture can be appropriately changed, reducing the sub-aperture length to 80% of the original length, until a sufficiently well-focused strong point is obtained, that is, until imaging is successful.

[0044] In this embodiment of the invention, an overlap is set between adjacent sub-apertures. The aperture overlap can be set to 30% to 60%. After dividing the sub-apertures in this way, the imaging error of each sub-aperture segment can be estimated, and then the overlapping part can be fused.

[0045] Step 3: Use the SPECAN algorithm to image each sub-aperture n and sub-block m, resulting in a total of n×m images.

[0046] In embodiments of the present invention, for GEO SAR imaging considering non-ideal factors, the SPECAN algorithm can easily perform range compression and range migration correction within each sub-aperture and each two-dimensional block. For the range-compressed signal, azimuth compression can be achieved through higher-order phase multiplication and azimuth-to-Fourier transform. Using the fixed-phase principle, series inversion method, and Jacobi-Anger expansion method, the non-ideal phase in the azimuth frequency domain can be obtained. As for the amplitude error, considering that it is estimated in the azimuth time domain by extracting the envelope of a strong target, its analytical form is not required.

[0047] According to the SPECAN imaging results, the final image of a point target (each point in the scene) can be regarded as the convolution of an ideal two-dimensional sinc function with the error spectrum. This will widen the main lobe of the image, causing paired echoes, thereby reducing the image quality.

[0048] Step 4: Discretely window the n×m image data. Perform PGA correction on the amplitude and phase errors of each image after discrete windowing to obtain the corresponding amplitude and phase errors for each image.

[0049] The periodic phase error caused by antenna vibration produces additional paired echo peaks in the image. These paired echo peaks appear at positions that are integer multiples of the vibration frequency from the actual main lobe location. This causes the power of the strong points to be discretely distributed in the neighborhood around the main lobe and the paired echo peaks, thus diffusing the target's energy.

[0050] Discrete windowing can be used to solve the problem of paired echo peaks. As shown in Figure 5, the basic idea of ​​discrete windowing is to multiply the image data by a series of discrete windows. These windows have non-zero values ​​only near the main lobe and paired echo peaks, thus concentrating the acquisition of target energy as much as possible. The width of the discrete windows is set as the superposition of multiple rectangular windows, each with a width of 10 times the image resolution. The center of each window is located at the main lobe and the first three pairs of paired echoes.

[0051] Each image after discrete windowing is corrected for amplitude and phase errors.

[0052] In the subsequent autofocusing process, this method estimates and compensates for amplitude and phase errors in the azimuth time domain. Specifically, the method of extracting the envelope of strong points is used to correct the amplitude error, and the PGA method is used to correct the phase error.

[0053] For amplitude errors, GEO SAR amplitude errors introduce fluctuations in the data domain, which can be expressed as multiplicative real noise in the azimuth time domain. These fluctuations transform the constant envelope of strong point data into a fluctuating form, degrading image quality. Therefore, to address the problem of amplitude non-uniformity, this method employs an amplitude compensation procedure by extracting the fluctuation amplitude of each strong point to eliminate the impact of amplitude errors.

[0054] Phase errors can be corrected using the PGA (Programmatical Generator) method. In PGA processing, a cyclic shift is first applied to several selected strong points to remove linear phase. Then, the phase gradient is estimated. Since the phase gradient of isolated, well-focused strong points in the data domain should be zero, the PGA algorithm can compensate for additional nonlinear phase errors caused by non-ideal factors, thus achieving accurate focusing. In this step, we can iteratively compensate for the estimated phase error until the phase error is below a preset value.

[0055] This step allows us to obtain the amplitude error and phase error corresponding to each sub-aperture.

[0056] Step 6: Perform error fusion on the amplitude and phase errors corresponding to each sub-aperture image to obtain the amplitude and phase errors of the full aperture. After full aperture error compensation, perform full aperture imaging on each two-dimensional sub-block, and then stitch the images together to obtain the final imaging result of GEO SAR for the entire imaging area.

[0057] Example 1

[0058] The improved high-orbit SAR adaptive error estimation and compensation method proposed in this invention was implemented using computer simulation to perform error compensation and accurate imaging of real targets. The simulation parameters are shown in Table 1, and the error parameters are shown in Table 2.

[0059] Table 1. Simulation parameters of GEO SAR system

[0060] Parameters / Unit Values: Orbital Altitude / km 36000; Orbital Inclination / Deg 55; Orbital Eccentricity / -0; Total Synthetic Aperture Time / s 100; Sub-Aperture Time / s 20; Sub-Aperture Overlap / % 50; Carrier Band / -L; Signal Bandwidth / MHz 18; Sampling Frequency / MHz 30; Downward Viewing Angle / Deg 4.5; Ground Distance Resolution / m 14.375; PRF / Hz 150; Azimuth Resolution / m 13.5; Pulse Width / us 20 surface

[0061] Table 2 Error Simulation Parameters

[0062]

[0063]

[0064] In the simulation, distributed targets are set using real SAR images. Each pixel in the image is designated as a point target, and the echoes of the point targets are accumulated to generate the echoes of the distributed targets. In the parameter settings, the distance between each pixel is less than the two-dimensional resolution. The original reference image is shown in Figure 6. After adding an error term to the original reference image, the imaging result without error compensation is shown in Figure 7. The imaging result using the method proposed in this invention is shown in Figure 8. By comparison, it can be seen that the imaging result quality of the method proposed in this invention is significantly better than the imaging result without error compensation. This indicates that the method proposed in this invention can effectively compensate for errors caused by various non-ideal factors and can automatically focus to obtain high-resolution GEO SAR imaging results.

[0065] In summary, the above are merely preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An improved adaptive error estimation and compensation method for high-orbit SAR, characterized in that, The method includes the following steps: S1: The GEO SAR satellite images the imaging area and acquires echo data; the echo data of the imaging area is divided into m equal two-dimensional sub-blocks on a two-dimensional plane. S2: Divide the fully synthesized aperture into n sub-apertures, with a set sub-aperture length. Specifically: initially set the sub-aperture length, and divide the fully synthesized aperture into n sub-apertures on average; perform imaging processing on the divided sub-apertures. If imaging processing fails, reduce the sub-aperture length to 80% of its original length until imaging processing is successful; the sub-aperture length is set to 10% of the total aperture length; adjacent sub-apertures overlap, with the overlap portion accounting for 30%~60% of the sub-aperture length; S3: Use the SPECAN algorithm to image each sub-aperture and the two-dimensional sub-block, obtaining n×m images; S4: Discretely window the n×m image data. Specifically, the image data is multiplied by a series of discrete windows, the width of which is set to the superposition of multiple rectangular windows, each rectangular window being 10 times the resolution of the image, with the center of the window located at the main lobe and the first three pairs of paired echo positions, respectively; the amplitude and phase errors of each image after discrete windowing are corrected by PGA to obtain the amplitude and phase errors corresponding to each image; S5: the amplitude and phase errors corresponding to each sub-aperture image are fused to obtain the amplitude and phase errors of the full aperture, and after full aperture error compensation, each two-dimensional sub-block is imaged with full aperture, and then the images are stitched together to obtain the final imaging result of GEO SAR for the entire imaging area.

Images