Satellite-borne SAR observation topographic characteristic estimation method based on generalized function beam

Abstract

The invention discloses a generalized function beam-based satellite-borne SAR (synthetic aperture radar) observed topographic characteristic estimation method, which comprises the following steps of: dividing a surveying and mapping area into a plurality of sub-surveying and mapping bands Si along a ground distance direction, marking any point in each sub-surveying and mapping band Si as a strong scattering target which is used for generating an echo signal; carrying out terrain contour modeling through a plurality of strong scattering targets marked in the surveying and mapping area; transmitting a pulse to the surveying and mapping area based on a satellite-borne SAR, segmenting the transmitted pulse into a plurality of sub-pulses corresponding to the sub-surveying and mapping band Si by adopting an imaging waveform segmentation method, and obtaining a transmitted waveform for surveying and mapping scene terrain estimation based on the sub-surveying and mapping band Si; extracting pitching direction signal vectors of echo signals at a plurality of sampling moments on the basis of the echo signals of the transmitted pulses, and calculating the position of each strong scattering target and sub-satellite point visual angle information on the basis of a preset algorithm; and based on the sub-satellite point view angle of each strong scattering target, calculating slope distance and height information of each strong scattering target, and reconstructing the terrain contour of the surveying and mapping scene along the ground distance direction.

Description

Technical Field

[0001] The present invention relates to the technical field of spaceborne SAR terrain estimation, and particularly to a method for estimating the terrain characteristics of spaceborne SAR observations based on a generalized function beam. Background Art

[0002] When a spaceborne synthetic aperture radar (SAR) realizes high-resolution wide-swath imaging, it usually uses a small-aperture antenna to form a wide beam to transmit pulses, so as to ensure the effective coverage of the transmitted beam over the entire wide-swath scene; when receiving the echo, the digital beamforming technology (DBF) of a phased array antenna is used to form a high-gain digital receiving beam to track the echo, and the arrival angle of the echo is always pointed to by the center of the receiving beam to improve the receiving gain and compensate for the gain loss when transmitting pulses with a small aperture. This process is as Figure 1 shown.

[0003] Based on the working process of the above high-resolution wide-swath SAR, it can be seen that in order to ensure that the receiving beam formed by DBF can effectively receive the echo, it is necessary to determine the real-time corresponding relationship between the echo and the arrival angle in advance, otherwise a large beam pointing deviation will be caused, resulting in the loss of echo gain. This phenomenon is as Figure 2 shown, which shows two different terrains, where the first terrain (left figure) is relatively flat, while the second terrain (right figure) has relatively large undulations. Under these two terrains, the same echo time will correspond to different arrival angle directions. Therefore, before using DBF to receive the echo, it is necessary to use relevant technical means to obtain the contour characteristics of the observed terrain, and design the receiving beam pointing based on this contour to ensure the matching with the echo direction.

[0004] Currently, the common algorithm for estimating the terrain contour of the observed scene is the spatial spectrum estimation algorithm. The steps of this algorithm include the following two steps:

[0005] 1) Pulse compress the original echo signals received by all elevation receiving channels;

[0006] 2) For each range bin after range compression, use the elevation multi-channel to perform spatial spectrum estimation to estimate the arrival angle direction corresponding to each range bin;

[0007] When performing spatial spectrum estimation for a certain range bin, first establish a covariance matrix using the signals of the elevation channels in this range bin, and then use this covariance matrix to obtain the spatial power spectrum within the observed viewing angle range, and search for the arrival angle corresponding to the peak position in this spatial power spectrum. This arrival angle is the estimated value of the arrival angle corresponding to this range bin.

[0008] The algorithm using spatial power spectrum is used to estimate the terrain profile of the observed scene of spaceborne SAR. Its biggest drawback is that the on-board processing calculation amount is too large. On the one hand, pulse compression processing needs to be performed on the echo signals of all pitch channels. On the other hand, spatial spectrum estimation needs to be carried out for each range gate. Spatial spectrum estimation involves covariance matrix operation and power value calculation of different arrival angles, all of which will greatly increase the workload of on-board processing operations.

[0009] Therefore, it is necessary to provide a method for estimating the terrain characteristics of spaceborne SAR observation based on the generalized function bundle to solve the problems mentioned in the above background technology. Summary of the Invention

[0010] To achieve the above object, the present invention provides the following technical solution: A method for estimating the terrain characteristics of spaceborne SAR observation based on the generalized function bundle, including:

[0011] Step 1: Divide the mapping area into several sub-mapping bands S along the ground range direction i and mark any point in each of the sub-mapping bands S i as a strong scattering target. The strong scattering target is used to generate echo signals, and terrain profile modeling is carried out through several strong scattering targets marked in the mapping area.

[0012] Step 2: Based on the spaceborne SAR transmitting pulses to the mapping area, divide the transmitted pulses into several sub-pulses corresponding to the sub-mapping bands S i by using the imaging waveform slicing method, and obtain the transmitted waveform for terrain estimation of the mapping scene based on each of the sub-mapping bands S i

[0013] Step 3: Based on the echo signals of the transmitted pulses, extract the pitch-direction signal vectors of the echo signals at several sampling moments, and calculate the positions of each of the strong scattering targets and the sub-satellite point viewing angle information based on a preset algorithm.

[0014] Step 4: Calculate the slant range and height information of each of the strong scattering targets based on the sub-satellite point viewing angles of each of the strong scattering targets, and then reconstruct the terrain profile of the mapping scene along the ground range direction.

[0015] Preferably, in step 2, when designing the transmitted waveform, any moment in the echo receiving window is set to: at most, it can simultaneously receive the echo signals of two adjacent strong scattering targets along the ground range direction.

[0016] Preferably, in step 3, for different terrain estimation pulses, extract the echo signal vectors at different interval distance sampling moments for processing to extract the position information of each of the strong scattering targets.

[0017] ​Preferably, in step 3, for the signal vector extracted at a certain moment, its processing algorithm flow includes:

[0018] S3-1. Extract the echo signal vector S r (τ) in the pitch direction at time τ;

[0019] S3-2. Construct signal vector matrices Y1 and Y2;

[0020] S3-3. Based on the signal vector matrices Y1 and Y2, evaluate whether there is echo signal data. If not, repeat S3-1 - S3-3 until there is echo signal data;

[0021] S3-4. If there is echo signal data, determine the pseudo-inverse matrix (Y1) H ;

[0022] S3-5. Extract the eigenvalues z H of the pseudo-inverse matrix (Y1) m and Y2;

[0023] S3-6. Evaluate whether these eigenvalues z m are located on the unit circle. If not, repeat S3-1 - S3-6 until the eigenvalues z m are located on the unit circle;

[0024] S3-7. If the eigenvalues z m are located on the unit circle, extract the beam center offset θ m of the detected strong scattering target;

[0025] S3-8. Output the distance of the position where the strong scattering target is located and the sub-satellite point viewing angle α m .

[0026] Preferably, in S3-1, set Y = S r (τ). The calculation method of the k-th element of the echo signal vector Y in the pitch direction is:

[0027]

[0028] where N tar represents the number of strong scattering targets contained in the echo signal at this moment, A m represents the echo signal amplitude generated by the backscattering coefficient of the m-th target, represents the eigenvalue.

[0029] Preferably, based on the sub-satellite point viewing angles α m of the extracted strong scattering targets, determine the slant range R m of the strong scattering targets. The expression is:

[0030]

[0031] Among them, the τ min is the minimum distance sampling moment, and the τ max is the maximum distance sampling moment. The c is the propagation speed of the on-orbit SAR transmitted pulse in the air, and it can be approximately considered that c≈3×10 8 m / s.

[0032] Preferably, according to the sub-satellite point view angle α m and the slant range R m， of each of the strong scattering targets, the height h m of each of the strong scattering targets is determined. The expression is:

[0033]

[0034] Among them, R E is the radius of the earth, and H is the satellite orbit height.

[0035] Preferably, based on the sub-satellite point view angle α m , the slant range R m and the height h m of each of the strong scattering targets, the terrain profile features in the ground distance direction are obtained by using the methods of linear fitting and data smoothing.

[0036] Compared with the prior art, the present invention provides a method for estimating the terrain characteristics of an on-orbit SAR based on a generalized function beam, and has the following beneficial effects:

[0037] In the present invention, by dividing the observation scene into several sub-swaths in the ground distance direction and designing the transmitted waveform based on the sub-swaths, there is no need to transmit a wide-spectrum and wide-pulse-width pulse waveform, and only a narrow-spectrum and narrow-pulse-width waveform is required, which greatly reduces the amount of data for on-orbit real-time processing;

[0038] Compared with the traditional terrain estimation method, there is no need to perform pulse compression processing, spatial spreading analysis, and searching for the peak position on the echo signal to locate the position of the characteristic scattering target. Instead, the collected original echo signal is directly processed, omitting the on-orbit pulse compression processing step, and performing calculations in the data echo, resulting in a significant reduction in the processing volume. On the one hand, it alleviates the demand for on-orbit hardware processing resources, and on the other hand, it improves the efficiency of terrain profile estimation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of a high-resolution wide-swath on-orbit SAR receiving echoes using digital beamforming technology;

[0040] Figure 2The angle-of-arrival deviation corresponding to the same echo moment under different observed-scene terrains;

[0041] Figure 3 It is a schematic structural diagram of the process for an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0042] Figure 4 It is a schematic structural diagram of the terrain modeling of the observed scene for an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0043] Figure 5 It is a schematic diagram of the estimated transmitted waveform for the terrain segmented based on the imaging transmitted waveform in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0044] Figure 6 It is the pulse waveform transmission sequence for the terrain profile estimation of the observed scene in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0045] Figure 7 It is a schematic diagram of the three-dimensional echo data block of the terrain estimation pulse train in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0046] Figure 8 It is a flow chart for extracting the position information of strong scattering targets based on the elevation signal vector in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0047] Figure 9 It is a schematic diagram of the terrain profile modeling result of the observed scene in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0048] Figure 10 It is the change of the target sub-satellite point view angle with the slant range when considering the terrain profile (blue line) and not considering the terrain profile (red line) in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0049] Figure 11 It is a schematic diagram of the deviation between the true sub-satellite point view angle and the theoretical sub-satellite point view angle of the target in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0050] Figure 12 It is a schematic diagram of the slant range deviation between the centers of adjacent sub-swaths in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0051] Figure 13 It is a schematic diagram of the comparison between the terrain profile estimation based on GPOF and the terrain modeling result in an on-board SAR observed-terrain characteristic estimation method based on the generalized function bundle;

[0052] Figure 14 It is a schematic diagram of the deviation between the terrain profile estimation based on the Generalized Function Bundle (GPOF) and the terrain modeling results for a method of estimating the terrain characteristics of spaceborne SAR observations based on the Generalized Function Bundle;

[0053] Figure 15 It is a schematic diagram of the estimated results of the sub-satellite point view of the characteristic scattering target based on the Generalized Function Bundle (GPOF) for a method of estimating the terrain characteristics of spaceborne SAR observations based on the Generalized Function Bundle;

[0054] Figure 16 It is a schematic diagram of the deviation between the estimated results of the sub-satellite point view of the characteristic scattering target based on the Generalized Function Bundle (GPOF) and the modeling values for a method of estimating the terrain characteristics of spaceborne SAR observations based on the Generalized Function Bundle;

[0055] In the figure:. Specific implementation manner

[0056] Please refer to Figure 1-16 , the present invention provides a method for estimating the terrain characteristics of spaceborne SAR observations based on the Generalized Function Bundle, including:

[0057] Step 1: Terrain modeling of the observation scene

[0058] Model the observation scene using the geometric parameter method. It can be seen from Figure 2 that for a specific slant range, the arrival angle of the echo signal is closely related to the height of the target in the observation scene. Therefore, divide the mapping area along the ground range direction into several sub-mapping strips S i (i = 1, 2,..., N), and mark any point in each of the sub-mapping strips S i as a strong scattering target, and the height of the strong scattering target is h i (i = 1, 2,..., N). The strong scattering target is used to generate echo signals, and terrain profile modeling is carried out through several strong scattering targets marked in the mapping area.

[0059] Specifically, a strong scattering target refers to an object that has significant reflection or scattering characteristics for incident electromagnetic waves or sound waves.

[0060] As Figure 4 shown, a schematic diagram of the terrain modeling of the present invention is given. The entire wide-swath observation scene is divided into S1 to S in the ground range direction N , and in each sub-mapping strip S i , the altitude of the target representing its scattering characteristics is h i .

[0061] Step 2: Design of the transmitted waveform for terrain estimation

[0062] A pulse is transmitted to the mapping area based on the spaceborne SAR, and the transmitted pulse is segmented into a number of sub-pulses corresponding to the sub-mapping band S by using the imaging waveform segmentation method. i Based on each sub-mapping band S i An emission waveform for terrain estimation of the mapping scene is obtained, and the terrain estimation of the mapping scene is obtained by using the original echoes of a number of sub-pulses. Therefore, compared with the imaging pulse waveform, the pulse and bandwidth of the selected emission waveform will be greatly reduced, thereby extremely reducing the sampling frequency of the analog-to-digital converter (ADC) and the amount of data processed in real time on the satellite, and effectively improving the efficiency of terrain assessment.

[0063] Among them, when designing the emission waveform, any moment within the echo reception window (ERW) is set to: be able to receive at most the echo signals of two adjacent strong scattering targets along the ground range direction simultaneously. That is to say, each sampling point along the ground range direction contains at most the information of two strong scattering target points, ensuring that no singular value appears when calculating the angle of arrival of the target.

[0064] As Figure 5 shown, an emission waveform for terrain estimation of the observation scene is obtained by the method of imaging waveform segmentation, which is represented by a series of rectangular frames as Sp1~Sp N . It can be seen from this waveform that both the spectral bandwidth and the time-domain pulse width of this waveform are only 1 / N of the imaging pulse, so that the sampling frequency of the analog-to-digital converter (ADC) and the satellite data rate can be reduced to 1 / N of the imaging process. When estimating the terrain profile, the emission sequence of these pulses is as Figure 6 shown.

[0065] Step 3: Echo processing and angle-of-arrival extraction

[0066] Based on the echo signal of the transmitted pulse, the elevation signal vectors of the echo signals at a number of sampling moments are extracted, and the positions of each strong scattering target and the sub-satellite point view angle information are calculated based on a preset algorithm.

[0067] When the resources for signal processing on the satellite are limited and the analog-to-digital converter (ADC) uses a relatively high rate, the echoes of these pulses can be processed alternately. As Figure 7 shown, for different terrain estimation pulses, the echo signal vectors at sampling moments with different interval distances are extracted for processing to extract the positions of each strong scattering target and the sub-satellite point view angle information. On the one hand, the redundancy between data can be removed, and on the other hand, the complementarity between their processing results can be ensured, so that the terrain profile of the entire observation scene can be obtained efficiently. Figure 7 The signal vectors used for on-satellite processing in

[0068] are marked with different colors. Figure 8As shown, it is the flow chart of the processing algorithm for the signal vector extracted at a certain moment.

[0069] Specifically, it includes:

[0070] S3-1. Extract the echo signal vector S r (τ) in the elevation direction at time τ;

[0071] S3-2. Construct the signal vector matrices Y1 and Y2;

[0072] S3-3. Evaluate whether there is echo signal data based on the signal vector matrices Y1 and Y2. If not, repeat S3-1 - S3-3 until there is echo signal data;

[0073] S3-4. If there is echo signal data, determine the pseudo-inverse matrix (Y1) of the signal vector matrix Y1 H ;

[0074] S3-5. Extract the eigenvalues z of the pseudo-inverse matrix (Y1) H and Y2; m ;

[0075] S3-6. Evaluate whether these eigenvalues z m are located on the unit circle. If not, repeat S3-1 - S3-6 until the eigenvalues z m are located on the unit circle;

[0076] S3-7. If the eigenvalues z m are located on the unit circle, extract the beam center offset θ of the detected strong scattering target m ;

[0077] S3-8. Output the distance of the position where the strong scattering target is located and the sub-satellite point view angle α m .

[0078] In practical applications, at time τ within the echo reception window (ERW), the echo signal vector S r (τ) received by the SAR payload can be expressed as:

[0079]

[0080] where N tar represents the number of target information contained in the echo signal at this moment. Since in the waveform design of step two, it can at most simultaneously acquire the echoes of two targets. Therefore, N tar satisfies the constraint condition 0 ≤ N tar ≤ 2. S m (τ) is the echo signal vector of the m-th strong scattering target, which can be expressed as

[0081] Sm = A m ·s rm (τ)·V m (2)

[0082] where A m is the echo signal amplitude generated by the backscattering coefficient of the m-th target, s rm (τ) and V m respectively represent the echo signal received by the elevation upward reference channel from the m-th target and the steering vector of the m-th target. Their analytical expressions are respectively

[0083]

[0084]

[0085] where rect[.] in equation (3) is a rectangular window function, R m is the slant range of the m-th target, T p is the transmitted pulse width, f c is the carrier frequency of the transmitted pulse, K r is the pulse frequency modulation slope. d in equation (4) e is the subarray interval in the elevation direction, θ m is the normal deviation angle corresponding to the m-th target, N e is the number of receiving channels in the elevation upward direction, and c is the propagation speed of the spaceborne SAR transmitted pulse in the air, which can be approximately considered as c ≈ 3×10^8 m / s.

[0086] Set Y = S r (τ), Therefore, the k-th element Y(k) of vector Y can be expressed as

[0087]

[0088] According to equation (5), use the Generalized Polynomial Orthogonal Function (GPOF) algorithm to obtain the relevant information of strong scattering targets, specifically as follows:

[0089] Construct two matrices Y1 and Y2 with dimensions (N e - L)×L from the signal vector Y. These two matrices are respectively

[0090]

[0091]

[0092] where L is the beam parameter in the Generalized Polynomial Orthogonal Function (GPOF) algorithm. The beam parameter L is used to eliminate the noise effect in the data, and its value range is [L / 2, L / 3].

[0093] The expression of Y(k) is given by formula (5), and Y1 and Y2 can be decomposed into analytical expressions of the product of multiple matrices, which are respectively

[0094] Y1 = Z1ΑZ2 (8)

[0095] Y2 = Z1ΑZ0Z2 (9)

[0096] Specifically,

[0097]

[0098]

[0099]

[0100]

[0101] where diag[...] represents a diagonal matrix.

[0102] Based on the above formula, a generalized matrix pencil is established

[0103] Y2 - λY1 = Z1Α{Z0 - λΙ}Z2 (14)

[0104] where I is an N tar ×N tar identity matrix. When the value range interval of the pencil parameter L is N tar ≤L≤N e -N tar the rank of the matrix Y2 - λY1 is N tar , so the parameter z m is the generalized eigenvalue of the matrix pair {Y2, Y1}. Based on this, obtaining the value of z m can be mapped to solve the traditional eigenvalue equation given by formula (16).

[0105]

[0106] where Y1 + is the pseudo-inverse of the matrix Y1 and can be expressed as

[0107]

[0108] where (.) H represents the conjugate transpose of the matrix. Since has an amplitude of 1, the effective eigenvalues z m calculated must be located on the unit circle, while the other eigenvalues are all singular values.

[0109] According to the number of eigenvalues on the unit circle, the number of target information N contained in the echo signal at this moment can be further determinedtar If all the eigenvalues are not on the unit circle, it indicates that the echo at this moment does not contain the information of any characteristic scattering point; if only one eigenvalue is on the unit circle, it means that the echo signal at this moment only contains the information of any one characteristic scattering point; since it is stipulated during waveform design that it can cover the echoes of at most two characteristic targets simultaneously, there will be at most two eigenvalues on the unit circle, and at this time N tar = 2.

[0110] After determining z m and N tar based on the above steps, the echo signal amplitude A m of each characteristic target is further obtained using the following equation.

[0111]

[0112] Based on the obtained z m , the beam center offset angle where the target is located - that is, the angle of arrival θ can be further calculated using m . Then, according to the sub-satellite point viewing angle α c of the antenna normal, the sub-satellite point viewing angle α m where the target is located can be obtained as

[0113] α m = α c + θ m (18)

[0114] Step Four: Scene Terrain Profile Reconstruction

[0115] For each extracted sub-satellite point viewing angle α m of the target, after processing the echo signals of all sub-pulses, the minimum and maximum range sampling times when it appears are found to be τ min and τ max respectively. At this time, the slant range R m of this target can be obtained as

[0116]

[0117] Finally, using the sub-satellite point viewing angle α m obtained from Equation (19) and the target slant range R m , the height h m of this target is finally obtained as

[0118]

[0119] where R E is the radius of the earth and H is the satellite orbit height.

[0120] After obtaining the sub-satellite point view angle α m , slant range R m , and altitude h m of the target using Equations (18), (19), and (20), the terrain profile of the observation scene in the ground range direction can be fully restored, thereby providing prior information for the spaceborne SAR to use the digital beamforming (DBF) technology for high-gain receiving beams.

[0121] Furthermore, based on the sub-satellite point view angle α m , slant range R m , and altitude h m of each of the strong scattering targets, the terrain profile features in the ground range direction are obtained using the methods of linear fitting and data smoothing, as shown by the Figure 4 dashed line part.

[0122] To verify the effectiveness of the present invention for the terrain profile of the observation scene, taking a spaceborne SAR system as an example, the spaceborne SAR adopts a phased array system and is equipped with two-dimensional multi-channels in both the elevation and azimuth directions, and can use the digital beamforming (DBF) technology to form digital high-gain beams to receive echoes. The parameters of the selected spaceborne SAR system are shown in the following table:

[0123]

[0124]

[0125] Through the orbital altitude of the satellite of 700 km and the view angle range of 24° to 26° of the observation scene, we can calculate that the near and far ground ranges of the observation scene are 315.26 km to 346.17 km respectively. Therefore, the entire mapping swath width is 30.9 km. In terrain modeling, we evenly divide the entire mapping swath width into a series of sub-swaths with a width of 300 m, and the total number of sub-swaths divided is N s = 104. For the representative strong scattering targets in each sub-swath, we set their positions in the ground range direction to be evenly spaced, and the altitude of these targets gradually increases from the near end to the far end of the scene to avoid the occlusion effect in the spaceborne SAR. The terrain profile modeling results are shown in Figure 9 , where the red solid dot targets are the representative scattering targets of the sub-swath.

[0126] Using Figure 9 the sub-satellite ground range position G0 and altitude h of the scattering target given in, based on the satellite-ground geometric relationship, the slant range R0 and sub-satellite point view angle α0 of the scattering target can be calculated, as shown in Equations (21) to (23).

[0127]

[0128]

[0129]

[0130] Among them, β is the geocentric angle of the target. Based on the above results, the variation curve of the target viewing angle α0 with the slant range R0 can be obtained, as shown by Figure 10 the blue line in. This correspondence will be used to generate digital receiving beams by subsequent digital beamforming (DBF) technology.

[0131] To illustrate the necessity of terrain profile estimation, we also give the correspondence between the target slant range and the sub-satellite point viewing angle when terrain undulation is not considered, as shown by Figure 10 the red line in. The difference between them is as shown by Figure 11 . It can be seen that the maximum angular deviation has reached 0.45°, accounting for 3 / 4 of the receiving antenna beam width. This angular mismatch will cause the complete failure of the digital beamforming (DBF) technology. Therefore, it also fully illustrates the necessity of terrain profile estimation before DBF in spaceborne SAR.

[0132] Through Figure 9 the sub-swath division results given, the slant range deviation between adjacent sub-swaths can be calculated, and the corresponding calculation results are as shown by Figure 12 . It can be seen that the minimum slant range deviation is 135.5 m at the proximal end. Therefore, according to the transmitting waveform design principle, the transmitting pulse width tp applied to scene terrain profile estimation can be designed as

[0133]

[0134] to meet the constraint of obtaining at most two point target echoes simultaneously at any time within the ERW.

[0135] Based on the pulse width given by Equation (24), the number of pulses N sub for terrain profile estimation that can be cut out from the imaging pulse is

[0136] N sub = T p / t p = 68 (25)

[0137] Among them, T p is the pulse width of the imaging pulse, 60 μs. Therefore, the bandwidth b w of the terrain estimation pulse is

[0138]

[0139] Among them, B wThe total bandwidth of the imaging pulse is 600 MHz. Based on the result given by Equation (26), the sampling frequency of the analog-to-digital converter (ADC) during terrain profile estimation is set to F s = 10 MHz to avoid aliasing effects.

[0140] Finally, Equations (27) and (28) give the center frequency and time offset of this series of transmitted pulses applied to terrain profile estimation relative to the imaging pulse.

[0141]

[0142]

[0143] After completing terrain profile modeling and transmit waveform design, the terrain profile estimation can be performed using the GPOF-based algorithm given by the present invention. Figure 13 The terrain profile estimation results for terrain modeling (red solid circles) are given, Figure 14 and the deviation of the estimation results is given. It can be seen from the figure that the maximum deviation of the altitude estimation for characteristic scattering targets does not exceed 3 m; Figure 15 The nadir view angle estimation results for these characteristic scattering points in terrain modeling are given. Due to the nadir view angle

[0144] of the modeling target (red) and the deviation between the "estimated target nadir view angle (blue)" is very small, less than 8×10 -4 deg, while Figure 15 the minimum measurement unit of the Y-axis in [figure] is 0.5 degrees, and such a small deviation cannot be effectively shown in this figure; therefore, in Figure 16 the deviation of the estimation results relative to the nadir view angle of the target modeling position is given. It can be seen that the target view angle deviation is within 0.001°, fully demonstrating the effectiveness of the method of the present invention.

[0145] It should be explained that the traditional estimation methods for the terrain profile of the observation scene all use spatial spectrum estimation algorithms, such as Capon and MUSIC algorithms. Compared with this algorithm, these algorithms require a very large amount of on-board resources, and the reasons are as follows:

[0146] 1. The transmitted pulses of the spatial spectrum estimation algorithm are wide-spectrum + wide-pulse-width pulses. Therefore, the amount of scene echo data to be processed is much larger than the echo data volume obtained by the present invention using narrow-band spectrum + narrow-pulse-width pulses;

[0147] 2. The spatial spectrum estimation algorithm needs to first perform pulse compression processing on the echo signals received by all channels, and use the signals after pulse compression processing to perform spatial spectrum estimation for each range gate; while the present invention directly processes the original echo sampling data without first performing on-board pulse compression processing;

[0148] 3. The spatial spectrum estimation algorithm needs to perform spatial spectrum operations for each range gate and search for the peak position of the spatial spectrum within the entire observation angle range. Therefore, it is necessary to traverse all the angle values to obtain the peak position of the spatial spectrum. However, the present invention directly calculates the angle position of the target by using the eigenvalue calculation method, so the on-board computational load is much less.

[0149] Considering the above three aspects, the present invention can greatly reduce the on-board real-time processing computational load on the basis of ensuring the terrain profile estimation accuracy. Thus, on the one hand, it can alleviate the demand for on-board hardware processing resources, and on the other hand, it can improve the efficiency of terrain profile estimation.

[0150] The above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent replacements or changes should be covered within the protection scope of the present invention.

Claims

1. A method for estimating the terrain characteristics of spaceborne SAR observations based on a generalized function bundle, characterized in that Including: Step 1: Divide the surveying and mapping area into several sub-surveying and mapping bands S along the ground distance direction i , and mark any point in each of the sub-surveying and mapping bands S i as a strong scattering target. The strong scattering target is used to generate an echo signal, and terrain profile modeling is performed through several strong scattering targets marked in the surveying and mapping area; Step 2: Transmit pulses to the mapping area based on the spaceborne SAR, and divide the transmitted pulses into a number of sub-pulses corresponding to the sub-mapping bands S i using the imaging waveform segmentation method, and obtain the transmitted waveform for terrain estimation of the mapping scene based on each sub-mapping band S i ​ Step 3: Based on the echo signals of the transmitted pulses, extract the elevation signal vectors of the echo signals at several sampling moments, and calculate the positions of each of the strong scattering targets and the sub-satellite point viewing angle information based on a preset algorithm; Step 4: Based on the sub-satellite point viewing angles of each of the strong scattering targets, calculate the slant range and altitude information of each of the strong scattering targets, and then reconstruct the terrain profile of the mapping scene along the ground distance direction.

2. The method for estimating the terrain characteristics of spaceborne SAR observation based on the generalized function bundle according to claim 1, wherein In Step 2, when designing the transmitted waveform, any moment within the echo receiving window is set to: be able to receive at most the echo signals of two adjacent strong scattering targets along the ground distance direction simultaneously.

3. The method for estimating the terrain characteristics of spaceborne SAR observation based on a generalized function bundle according to claim 1, characterized in that In Step 3, for different terrain estimation pulses, extract the echo signal vectors at sampling moments with different interval distances for processing to extract the position information of each of the strong scattering targets.

4. A method for estimating the terrain characteristics of spaceborne SAR observation based on a generalized function bundle according to claim 1, characterized in that In Step 3, for the signal vector extracted at a certain moment, the processing algorithm flow includes: S3-1. Extract the echo signal vector S in the pitch direction at time τ r (τ); S3-2: Construct signal vector matrices Y1 and Y2; S3-3: Based on the signal vector matrices Y1 and Y2, evaluate whether there is echo signal data. If not, repeat S3-1 - S3-3 until there is echo signal data; S3-4. If there is echo signal data, determine the pseudo-inverse matrix (Y1) of the signal vector matrix Y1 H ; S3-5. Extract the pseudo-inverse matrix (Y1) H and the eigenvalue z of Y2 m ; S3-6. Evaluate these eigenvalue z m Whether it is on the unit circle. If not, repeat S3-1 - S3-6 until the eigenvalue z m is on the unit circle; S3-7. If the eigenvalue z m is located on the unit circle, extract the beam center offset θ m of the detected strong scattering target; S3-8. Output the distance to the location of the strong scattering target and the sub-satellite point viewing angle α m .

5. The on-orbit SAR observation terrain feature estimation method based on the generalized function bundle according to claim 4, characterized in that In S3-1, set Y = S r (τ). The calculation method of the echo signal vector in the pitch direction is as follows: Among them, Y(k) represents k elements of the echo signal vector, and N tar represents the number of strong scattering targets contained in the echo signal at this moment, and A m represents the amplitude of the echo signal generated by the backscattering coefficient of the m-th target, represents the eigenvalue.

6. The method for estimating the terrain characteristics of spaceborne SAR observation based on the generalized function bundle according to claim 4, wherein Based on the sub-satellite point view angle α of each of the extracted strong scattering targets m , determine the slant range R of each of the strong scattering targets m The expression is: wherein, the τ min is the minimum distance sampling time, the τ max is the maximum distance sampling time, and c is the propagation speed of the spaceborne SAR transmitted pulse in the air.

7. A method for estimating the terrain characteristics of spaceborne SAR observation based on a generalized function bundle according to claim 6, characterized in that According to the sub-satellite point view angle α of each said strong scattering target m and the slant range R m， to determine the height h of each said strong scattering target m The expression is: Among them, R E is the radius of the earth, and H is the satellite orbit altitude.

8. A method for estimating the terrain characteristics of spaceborne SAR observations based on a generalized function bundle according to claim 7, characterized in that, Based on the sub-satellite point view angle α of each of the strong scattering targets m , the slant range R m and the altitude h m , a method of linear fitting and data smoothing is used to obtain the topographic contour features in the ground distance direction.

Images