A five-component power decomposition method and system for compressed polarization SAR data

Abstract

This invention discloses a five-component power decomposition method and system for compressed polarimetric SAR data, belonging to the field of synthetic aperture radar target decomposition technology. The invention preprocesses the covariance data of each pixel in the acquired compressed polarimetric SAR image to extract two dominant scattering mechanism discriminators for each pixel; determines the dominant scattering mechanism based on the extracted discriminators; selects a scattering model for power decomposition based on the dominant scattering mechanism of each pixel; and performs power decomposition on each scattering model to obtain five-component power. This invention is used for radar target decomposition.

Description

Technical Field

[0001] A five-component power decomposition method and system for compressed polarimetric SAR data is presented for radar target decomposition, belonging to the field of synthetic aperture radar target decomposition technology. Background Technology

[0002] Fully polarimetric (FP) SAR acquires a covariance matrix containing nine independent parameters by alternately transmitting electromagnetic signals of horizontal (H) and vertical (V) linear polarizations. This matrix contains rich backscattering information, aiding in a better understanding of target attributes and enabling applications such as land use and terrain classification, demonstrating its advantages in Earth observation. However, the alternating transmission of two electromagnetic signals results in a pulse repetition frequency twice that of traditional dual-polarimetric (DP) SAR systems, while theoretically limiting its data coverage to only half that of DP systems. The complex system design and maintenance, along with the high cost of data acquisition, restrict the application of FP. Compact polarimetric SAR, as a coherent dual-polarization system, achieves simultaneous transmission of H and V polarizations by transmitting a special polarization (most commonly 45-degree inclined linear polarization and left-right circular polarization). While providing more information (relative phase) than traditional DP systems, it retains the advantages of DP systems in terms of system complexity and data acquisition cost compared to FP systems, thus balancing target information acquisition and observation coverage. Depending on the transmit and receive polarizations, compact polarization SAR mainly has three modes: the pi / 4 mode (transmitting with 45-degree linear polarization and receiving with orthogonal linear polarization), the CTLR mode (transmitting with right-hand circular polarization and receiving with orthogonal linear polarization), and the DCP mode (transmitting and receiving with circular polarization). Among these three modes, the CTLR mode is currently the only mode used in practical systems due to its rotation invariance, simple hardware implementation, and self-calibration capabilities. It is also the most widely used compact polarization SAR for Earth observation applications.

[0003] Currently, data processing methods for compressed polarimetric SAR (SAR) are mainly divided into two categories: pseudo-full polarimetric reconstruction and compressed polarimetric power decomposition. Pseudo-full polarimetric reconstruction methods typically introduce uncertainties based on implicit assumptions (such as reflection symmetry) during the reconstruction of compressed polarimetric SAR data into a full polarimetric covariance matrix, affecting reconstruction accuracy. Therefore, compressed polarimetric power decomposition methods, which directly process compressed polarimetric SAR data, are more favored by researchers. Compressed polarimetric power decomposition methods are mainly divided into two categories: those based on wave decomposition theorem and those based on scattering models. Methods based on wave decomposition theorem decompose the total power into the sum of a completely depolarized component and a completely polarized component using polarimetric parameters, treating the former as a volume scattering component and the latter as the sum of surface scattering and bielastic scattering. Model-based methods are based on the model decomposition theory of full polarimetry, utilizing the projection transformation relationship between full polarimetry and compressed polarimetry to establish a compressed polarimetric SAR scattering model, which has significant physical meaning.

[0004] Compact polarimetric SAR data has only four observations, and even the simplest classic Freeman-Durden three-component model has seven unknowns. Existing compact polarimetric scattering models are all three-component scattering models. By fixing the values ​​of parameters α or β under the condition of surface scattering or bielastic scattering dominance, and solving for reasonable three-component scattering power under the constraint of positive power, their ability to characterize complex targets is limited. Compact polarimetric scattering models with more components are not only difficult to construct, but also challenging in terms of model decomposition methods (more unknowns compared to three-component models). Furthermore, most existing compact polarimetric power decomposition methods equate the completely depolarized component in the total power to the volume scattering power, treating vegetation as the sole source of depolarization effects, ignoring the depolarization effects that may be caused by undulating surfaces and buildings, leading to an overestimation of volume scattering power.

[0005] In summary, the existing compact polarization SAR decomposition method has the following technical problems:

[0006] 1. It can only decompose the three components of the scattering model of each pixel in the compressed polarimetric SAR image, which results in a limited ability to characterize complex targets, and thus limits the application of land cover classification, target detection and recognition based on compressed polarimetric SAR data.

[0007] 2. Most studies equate the complete depolarization component in the total power with volume scattering power, treating vegetation as the sole source of depolarization, neglecting the depolarization effects that may be caused by undulating surfaces and buildings. This leads to an overestimation of volume scattering power, which in turn affects the accuracy of applications such as land cover classification and target recognition. Specifically, in land cover classification, an overestimation of volume scattering may cause some buildings and undulating surfaces to be incorrectly classified as vegetation; in target detection and recognition, it may lead to misidentification of targets.

[0008] 3. Challenging: The fact that the covariance matrix of compressed polarimetric SAR has only 4 independent parameters presents a huge challenge to constructing and solving scattering models with more components (greater than 3). Summary of the Invention

[0009] The purpose of this invention is to provide a five-component power decomposition method and system for compressed polarimetric SAR data, which solves the problem that only three components of the scattering model of each pixel in a compressed polarimetric SAR image can be decomposed, resulting in limited ability to characterize complex targets, and thus limiting applications such as land cover classification, target detection and recognition based on compressed polarimetric SAR data.

[0010] A five-component power decomposition method for compressed polarimetric SAR data includes the following steps:

[0011] S1. Preprocess the covariance data of each pixel in the acquired compressed polarimetric SAR image, and extract the two dominant scattering mechanism discriminators for each pixel, where the dominant scattering mechanism discriminators are discriminators m and m respectively. v and discriminator g3;

[0012] S2. Based on the two dominant scattering mechanisms of each extracted pixel, the dominant scattering mechanism is determined by the discriminator. The dominant scattering mechanisms include volume scattering, surface scattering, and bielastic scattering.

[0013] S3. Select a scattering model based on the dominant scattering mechanism of each pixel to perform power decomposition;

[0014] S4. Perform power decomposition on each scattering model to obtain five-component power.

[0015] Furthermore, the specific steps of step S1 are as follows:

[0016] S10. Convert the covariance matrix C2 of each pixel in the compressed polarimetric SAR image into a compressed polarimetric SARStokes vector SV, where the covariance matrix C2 is:

[0017]

[0018] In the formula, C xy C represents the matrix element in the x-th row and y-th column. 11 and C 22 For real numbers, C 12 and C 21 They are complex numbers and are complex conjugates of each other;

[0019] The specific formula for converting the Stokes vector SV in compact polarization SAR is as follows:

[0020]

[0021] In the formula, g0 represents the total power, g1 represents the power of horizontal or vertical linear polarization, g2 represents the power of 45-degree linear polarization, g3 represents the power of circular polarization, and Re(C 12 ) and Im(C 12 ) represent C respectively 12 The real and imaginary parts;

[0022] S11. Based on the compact polarization SAR Stokes vector, extract the discriminator m for each pixel. v Discriminator m v Specifically:

[0023]

[0024] Where max{a, b} represents taking the larger of a and b;

[0025] S12. Based on the compact polarization SAR Stokes vector, extract the discriminator g3 for each pixel.

[0026] Furthermore, the specific steps of step S2 are as follows:

[0027] Step S20: Determine the discriminator m v If the pixel is less than a given threshold, then the pixel is dominated by volume scattering; otherwise, proceed to step S21.

[0028] Step S21: Determine whether the discriminator g3 is less than zero. If it is less than zero, the pixel is dominated by surface scattering; otherwise, it is dominated by bielastic scattering.

[0029] Furthermore, the specific steps of step S3 are as follows:

[0030] Step S30: If the pixel is dominated by volume scattering, then select the scattering model. V Perform power decomposition;

[0031] scattering model V Specifically:

[0032]

[0033] In the formula, f s f d and f v These are the weighting factors for surface scattering power, bielastic scattering power, and volume scattering power, respectively, and θ0 is the average direction angle;

[0034] Step S31: If the pixel is dominated by surface scattering, then select the scattering model. S Perform power decomposition;

[0035] scattering model S Specifically:

[0036]

[0037] In the formula, f vs β is the weighting factor for the surface volume scattering power, and β∈[-1,1] is the complex observation value of surface scattering;

[0038] Step S32: If the pixel is dominated by bielastic scattering, then select the scattering model. D Perform power decomposition;

[0039] scattering model D Specifically:

[0040]

[0041] In the formula, f vd α is the weighting factor for the biplane scattering power, and α∈[-1,1] is the complex observation of biplane scattering.

[0042] Furthermore, the specific steps of step S4 are as follows:

[0043] Step S40, based on discriminator m v and scattering model V The power P is obtained. s P d and P v :

[0044]

[0045] Default P vs and P vd Zero;

[0046] Therefore, the five-component power is obtained as follows:

[0047]

[0048] Among them, P s Represents surface scattering power, P d Represents the bielastic scattering power, P v Represents volume scattering power, P vs Represents the surface volume scattering power, P vd Indicates the power of the two projectiles scattering;

[0049] Step S41, based on the scattering model S The power P is obtained. s P d and P vs :

[0050]

[0051] Default P v and P vd Zero;

[0052] Therefore, the five-component power is obtained as follows:

[0053]

[0054] in:

[0055]

[0056] Step S42, based on the scattering model D The power P is obtained. s P d and Pvd :

[0057]

[0058] Default P v and P vs Zero;

[0059] Therefore, the five-component power is obtained as follows:

[0060]

[0061] in:

[0062]

[0063] A five-component power decomposition system for compressed polarimetric SAR data includes:

[0064] Extraction module: Preprocesses the covariance data of each pixel in the acquired compressed polarimetric SAR image, and extracts the two dominant scattering mechanism discriminators for each pixel, where the dominant scattering mechanism discriminators are discriminators m and m respectively. v and discriminator g3;

[0065] Judgment module: Based on the two dominant scattering mechanisms of each extracted pixel, the discriminator determines its dominant scattering mechanism, which includes volume scattering, surface scattering, and bielastic scattering.

[0066] Selection module: Select a scattering model for power decomposition based on the dominant scattering mechanism of each pixel;

[0067] Power decomposition module: Performs power decomposition on each scattering model to obtain five-component power.

[0068] Furthermore, the specific implementation steps of the extraction module are as follows:

[0069] S10. Convert the covariance matrix C2 of each pixel in the compressed polarimetric SAR image into a compressed polarimetric SARStokes vector SV, where the covariance matrix C2 is:

[0070]

[0071] In the formula, C xy C represents the matrix element in the x-th row and y-th column. 11 and C 22 For real numbers, C 12 and C 21 They are complex numbers and are complex conjugates of each other;

[0072] The specific formula for converting the Stokes vector SV in compact polarization SAR is as follows:

[0073]

[0074] In the formula, g0 represents the total power, g1 represents the power of horizontal or vertical linear polarization, g2 represents the power of 45-degree linear polarization, g3 represents the power of circular polarization, and Re(C 12 ) and Im(C 12 ) represent C respectively 12 The real and imaginary parts;

[0075] S11. Based on the compact polarization SAR Stokes vector, extract the discriminator m for each pixel. v Discriminator m v Specifically:

[0076]

[0077] Where max{a, b} represents taking the larger of a and b;

[0078] S12. Based on the compact polarization SAR Stokes vector, extract the discriminator g3 for each pixel.

[0079] Furthermore, the specific implementation steps of the judgment module are as follows:

[0080] Step S20: Determine the discriminator m v If the pixel is less than a given threshold, then the pixel is dominated by volume scattering; otherwise, proceed to step S21.

[0081] Step S21: Determine whether the discriminator g3 is less than zero. If it is less than zero, the pixel is dominated by surface scattering; otherwise, it is dominated by bielastic scattering.

[0082] Furthermore, the specific implementation steps of the selection module are as follows:

[0083] Step S30: If the pixel is dominated by volume scattering, then select the scattering model. V Perform power decomposition;

[0084] scattering model V Specifically:

[0085]

[0086] In the formula, f s f d and f v These are the weighting factors for surface scattering power, bielastic scattering power, and volume scattering power, respectively, and θ0 is the average direction angle;

[0087] Step S31: If the pixel is dominated by surface scattering, then select the scattering model. SPerform power decomposition;

[0088] scattering model S Specifically:

[0089]

[0090] In the formula, f vs β is the weighting factor for the surface volume scattering power, and β∈[-1,1] is the complex observation value of surface scattering;

[0091] Step S32: If the pixel is dominated by bielastic scattering, then select the scattering model. D Perform power decomposition;

[0092] scattering model D Specifically:

[0093]

[0094] In the formula, f vd α is the weighting factor for the biplane scattering power, and α∈[-1,1] is the complex observation of biplane scattering.

[0095] Furthermore, the specific implementation steps of the power decomposition module are as follows:

[0096] Step S40, based on discriminator m v and scattering model V The power P is obtained. s P d and P v :

[0097]

[0098] Default P vs and P vd Zero;

[0099] Therefore, the five-component power is obtained as follows:

[0100]

[0101] Among them, P s Represents surface scattering power, P d Represents the bielastic scattering power, P v Represents volume scattering power, P vs Represents the surface volume scattering power, P vd Indicates the power of the two projectiles scattering;

[0102] Step S41, based on the scattering model S The power P is obtained. s P d and Pvs :

[0103]

[0104] Default P v and P vd Zero;

[0105] Therefore, the five-component power is obtained as follows:

[0106]

[0107] in:

[0108]

[0109] Step S42, based on the scattering model D The power P is obtained. s P d and P vd :

[0110]

[0111] Default P v and P vs Zero;

[0112] Therefore, the five-component power is obtained as follows:

[0113]

[0114] in:

[0115]

[0116] Compared with the prior art, the advantages of the present invention are as follows:

[0117] I. The decomposition method in this invention can decompose the scattering model of each pixel in a compressed polarization SAR image into five components (volume scattering power P). v Surface volume scattering power P vs Two-body scattering power P vd Surface scattering power P s and bielastic scattering power P d Using five components allows for a more accurate characterization of complex targets;

[0118] II. This invention decomposes the total power into volume scattering power P. v Surface volume scattering power P vs Two-body scattering power P vd Surface scattering power P s and bielastic scattering power P dThe sum, compared with other compact polarization power decomposition methods, increases the surface volume scattering power P. vs and the two-body scattering power P vd These two components, representing the depolarization effects caused by undulating terrain and buildings respectively, solve the problem of overestimation of volume scattering and achieve a more accurate characterization of complex targets. Attached Figure Description

[0119] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0120] Figure 1 This is the overall flowchart of the present invention;

[0121] Figure 2 This is the dataset used for verification in this invention;

[0122] Figure 3 This is a three-component colorimetric image of the power decomposition result of the present invention;

[0123] Figure 4 This is a five-component pie chart of the power decomposition results for a typical region according to the present invention, wherein (a), (b), (c), and (d) respectively represent Figure 2 Five-component pie charts of power decomposition results for typical regions A, B, C, and D are shown.

[0124] Figure 5 This is a schematic diagram of the specific process of the present invention. In the diagram, R, h and v represent right-hand circular polarization, horizontal polarization and vertical polarization, respectively. Detailed Implementation

[0125] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0126] Based on the classic three-component scattering model, adding two components to each pixel to represent undulating surfaces and buildings can alleviate the problem of volume scattering overestimation. However, no similar method has been proposed in the existing literature. Therefore, it is necessary to improve the existing technology.

[0127] A five-component power decomposition method for compressed polarimetric SAR data includes the following steps:

[0128] S1. Preprocess the covariance data of each pixel in the acquired compressed polarimetric SAR image, and extract the two dominant scattering mechanism discriminators for each pixel, where the dominant scattering mechanism discriminators are discriminators m and m respectively. v and discriminator g3;

[0129] The specific steps are as follows:

[0130] S10. Convert the covariance matrix C2 of each pixel in the compressed polarimetric SAR image into a compressed polarimetric SARStokes vector SV, where the covariance matrix C2 is:

[0131]

[0132] In the formula, C xy C represents the matrix element in the x-th row and y-th column. 11 and C 22 For real numbers, C 12 and C 21 They are complex numbers and are complex conjugates of each other;

[0133] The specific formula for converting the Stokes vector SV in compact polarization SAR is as follows:

[0134]

[0135] In the formula, g0 represents the total power, g1 represents the power of horizontal or vertical linear polarization, g2 represents the power of 45-degree linear polarization, g3 represents the power of circular polarization, and Re(C 12 ) and Im(C 12 ) represent C respectively 12 The real and imaginary parts;

[0136] S11. Based on the compact polarization SAR Stokes vector, extract the discriminator m for each pixel. v Discriminator m v Specifically:

[0137]

[0138] Where max{a, b} represents taking the larger of a and b.

[0139] S12. Extract the discriminator for each pixel based on the compact polarization SAR Stokes vector. g3.

[0140] S2. Based on the two dominant scattering mechanisms of each extracted pixel, the dominant scattering mechanism is determined by the discriminator. The dominant scattering mechanisms include volume scattering, surface scattering, and bielastic scattering.

[0141] The specific steps are as follows:

[0142] Step S20: Determine the discriminator m v Check if it is less than a given threshold of 0.37. If it is less than the threshold, the pixel is dominated by volume scattering; otherwise, proceed to step S21.

[0143] Step S21: Determine whether the discriminator g3 is less than zero. If it is less than zero, the pixel is dominated by surface scattering; otherwise, it is dominated by bielastic scattering.

[0144] S3. Select a scattering model based on the dominant scattering mechanism of each pixel to perform power decomposition;

[0145] The specific steps are as follows:

[0146] Step S30: If the pixel is dominated by volume scattering, then select the scattering model. V Perform power decomposition;

[0147] scattering model V Specifically:

[0148]

[0149] In the formula, f represents the total power. s f d and f v These are the weighting factors for surface scattering power, bielastic scattering power, and volume scattering power, respectively, and θ0 is the average direction angle;

[0150] Step S31: If the pixel is dominated by surface scattering, then select the scattering model. S Perform power decomposition;

[0151] scattering model S Specifically:

[0152]

[0153] In the formula, f vs β is the weighting factor for the surface volume scattering power, and β∈[-1,1] is the complex observation value of surface scattering;

[0154] Step S32: If the pixel is dominated by bielastic scattering, then select the scattering model. D Perform power decomposition;

[0155] scattering model D Specifically:

[0156]

[0157] In the formula, f vd α is the weighting factor for the biplane scattering power, and α∈[-1,1] is the complex observation of biplane scattering.

[0158] S4. Perform power decomposition on each scattering model to obtain five-component power.

[0159] The specific steps are as follows:

[0160] Step S40, based on discriminator m v and scattering model V There are 4 observations g0, g1, g2 and g3, and 5 unknowns f. s f d f v m v and θ0. Let m v Consider the known numbers and solve for the numbers containing m. v Power P s P d and P v :

[0161]

[0162] Based on the constraint that the power is not zero, we can solve for m. v The range of values ​​for:

[0163]

[0164] m v The smaller P v The larger. Therefore This aligns with the assumption that volume scattering is dominant.

[0165] Default P vs and P vd Zero;

[0166] Therefore, the five-component power is obtained as follows:

[0167]

[0168] Among them, P s Represents surface scattering power, P d Represents the bielastic scattering power, P v Represents volume scattering power, P vs Represents the surface volume scattering power, P vd Indicates the power of the two projectiles scattering;

[0169] Step S41, based on the scattering modelS There are 4 observations g0, g1, g2 and g3, and 5 unknowns f. s f d f vs Re(β) and Im(β). Solve for the power P with |β| as a known value. s P d and P vs :

[0170]

[0171] Based on the constraint that the power is not zero, the range of values ​​for |β| can be determined:

[0172]

[0173] When |β| takes the values ​​of the left and right endpoints, P vs and P d The values ​​are all 0, which does not reflect reality. Therefore... It is more in line with the actual situation.

[0174] Default P v and P vd Zero;

[0175] Therefore, the five-component power is obtained as follows:

[0176]

[0177] in:

[0178]

[0179] Step S42, based on the scattering model D There are 4 observations g0, g1, g2 and g3, and 5 unknowns f. s f d f vd Re(α) and Im(α). Treating |α| as a known value, solve for the power P with |α|. s P d and P vd :

[0180]

[0181] Based on the constraint that the power is not zero, the range of values ​​for |α| can be determined:

[0182]

[0183] When |α| takes the values ​​of the left and right endpoints, P vd and P sThe values ​​are all 0, which does not reflect reality. Therefore... It is more in line with the actual situation.

[0184] Default P v and P vs Zero;

[0185] Therefore, the five-component power is obtained as follows:

[0186]

[0187] in:

[0188]

[0189] Figure 2 The experimental image shown is from the C-band RISAT-1 dataset acquired in San Francisco on August 9, 2016. The product ID is 163791211, the resolution is 3.33 × 2.34 m (Azimuth × Range), the image size is 8719 × 13843 pixels, the incident angle is 38.19869 degrees, and the mode is RH, RV. The experimental filter window is 5. Four typical regions are represented: A (water region), B (forest region), C (orthogonal building region), and D (directional building region). All four regions are 400 × 400 pixels in size.

[0190] The five-component power decomposition results of each pixel in the compressed polarimetric SAR image are processed as follows: P s =P s +P vs P d =P d +P vd P was obtained v Updated P s And the updated P d The mixed image of the three-component decomposed power is as follows: Figure 3 As shown, this invention identified high levels of bielastic scattering and volume scattering components in both the built-up and vegetated areas.

[0191] Will as Figure 2 The five-component power decomposition results of each pixel in the compressed polarimetric SAR images of the four typical regions shown are represented by pie charts, as follows: Figure 4 As shown. P vs In region A, the largest proportion was 29%, in regions B and D it was 13% and 10% respectively, and in region C it was the smallest at 5%; P vdIn regions C and D, the proportions were 14% and 10%, respectively, while in regions B and A, the proportions were negligible at 2% and 0%, respectively. This indicates that the present invention effectively separates the surface scattering and bielastic scattering components, which are confused in the fully depolarized component, into surface volume scattering components and bielastic volume scattering components, thereby alleviating the problem of overestimation of volume scattering components and achieving accurate characterization of complex targets.

Claims

1. A five-component power decomposition method for compressed polarimetric SAR data, characterized in that, Includes the following steps: S1. Preprocess the covariance data of each pixel in the acquired compressed polarimetric SAR image, and extract the two dominant scattering mechanism discriminators for each pixel. The dominant scattering mechanism discriminators are respectively the discriminator... and discriminator ; S2. Based on the two dominant scattering mechanisms of each extracted pixel, the dominant scattering mechanism is determined by the discriminator. The dominant scattering mechanisms include volume scattering, surface scattering, and bielastic scattering. S3. Select a scattering model based on the dominant scattering mechanism of each pixel to perform power decomposition; The specific steps are as follows: Step S30: If the pixel is dominated by volume scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, , and These are the weighting factors for surface scattering power, bielastic scattering power, and volume scattering power, respectively. The average direction angle, Indicates total power. Power indicating horizontal or vertical linear polarization This indicates the power of a 45-degree linearly polarized circuit. The power representing circular polarization, i.e., the power of the discriminator. ; Step S31: If the pixel is dominated by surface scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, This is a weighting factor for the surface volume scattering power. For complex observations of surface scattering, and They represent The real and imaginary parts; Step S32: If the pixel is dominated by bielastic scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, This is a weighting factor for the scattering power of the two projectiles. For complex observations of bielastic scattering, express The imaginary part; S4. Perform power decomposition on each scattering model to obtain five-component power.

2. The five-component power decomposition method for compressed polarimetric SAR data according to claim 1, characterized in that, The specific steps of step S1 are as follows: S10. Calculate the covariance matrix of each pixel in the compressed polarimetric SAR image. Converted to compact polarimetric SAR Stokes vectors SV, where the covariance matrix is... for: In the formula, Indicates the first line, number Matrix elements on columns, and For real numbers, and They are complex numbers and are complex conjugates of each other; The specific formula for converting the Stokes vector SV in compact polarization SAR is as follows: In the formula, and They represent The real and imaginary parts; S11. Based on the compact polarization SAR Stokes vector, extract the discriminator for each pixel. Discriminator Specifically: ； in, ; S12. Based on the compact polarization SAR Stokes vector, extract the discriminator for each pixel. .

3. The five-component power decomposition method for compressed polarimetric SAR data according to claim 2, characterized in that, The specific steps of step S2 are as follows: Step S20: Determine the discriminator If the pixel is less than a given threshold, then the pixel is dominated by volume scattering; otherwise, proceed to step S21. Step S21: Determine the discriminator If the value is less than zero, the pixel is dominated by surface scattering; otherwise, it is dominated by bielastic scattering.

4. The five-component power decomposition method for compressed polarimetric SAR data according to claim 3, characterized in that, The specific steps of step S4 are as follows: Step S40, Based on the discriminator and scattering model , obtain power , and : default and Zero; Therefore, the five-component power is obtained as follows: in, Indicates surface scattering power, Indicates the power of bielastic scattering. Indicates volume scattering power, Indicates the surface volume scattering power, Indicates the power of the two projectiles scattering; Step S41, based on the scattering model , obtain power , ,and : default and Zero; Therefore, the five-component power is obtained as follows: in: Step S42, based on the scattering model , obtain power , ,and : default and Zero; Therefore, the five-component power is obtained as follows: in: 。 5. A five-component power decomposition system for compressed polarimetric SAR data, characterized in that, include: Extraction Module: This module preprocesses the covariance data of each pixel in the acquired compressed polarimetric SAR image and extracts the two dominant scattering mechanism discriminators for each pixel. The dominant scattering mechanism discriminators are respectively... and discriminator ; Judgment module: Based on the two dominant scattering mechanisms of each extracted pixel, the discriminator determines its dominant scattering mechanism, which includes volume scattering, surface scattering, and bielastic scattering. Selection module: Select a scattering model for power decomposition based on the dominant scattering mechanism of each pixel; The specific implementation steps are as follows: Step S30: If the pixel is dominated by volume scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, , and These are the weighting factors for surface scattering power, bielastic scattering power, and volume scattering power, respectively. The average direction angle, Indicates total power. Power indicating horizontal or vertical linear polarization This indicates the power of a 45-degree linearly polarized circuit. The power representing circular polarization, i.e., the power of the discriminator. ; Step S31: If the pixel is dominated by surface scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, This is a weighting factor for the surface volume scattering power. For complex observations of surface scattering, and They represent The real and imaginary parts; Step S32: If the pixel is dominated by bielastic scattering, then select the scattering model. Perform power decomposition; Scattering model Specifically: In the formula, This is a weighting factor for the scattering power of the two projectiles. For complex observations of bielastic scattering, express The imaginary part; Power decomposition module: Performs power decomposition on each scattering model to obtain five-component power.

6. A five-component power decomposition system for compressed polarimetric SAR data according to claim 5, characterized in that, The specific implementation steps of the extraction module are as follows: S10. Calculate the covariance matrix of each pixel in the compressed polarimetric SAR image. Converted to compact polarimetric SAR Stokes vectors SV, where the covariance matrix is... for: In the formula, Indicates the first line, number Matrix elements on columns, and For real numbers, and They are complex numbers and are complex conjugates of each other; The specific formula for converting the Stokes vector SV in compact polarization SAR is as follows: In the formula, and They represent The real and imaginary parts; S11. Based on the compact polarization SAR Stokes vector, extract the discriminator for each pixel. Discriminator Specifically: ； in, ; S12. Based on the compact polarization SAR Stokes vector, extract the discriminator for each pixel. .

7. A five-component power decomposition system for compressed polarimetric SAR data according to claim 6, characterized in that, The specific implementation steps of the judgment module are as follows: Step S20: Determine the discriminator If the pixel is less than a given threshold, then the pixel is dominated by volume scattering; otherwise, proceed to step S21. Step S21: Determine the discriminator If the value is less than zero, the pixel is dominated by surface scattering; otherwise, it is dominated by bielastic scattering.

8. A five-component power decomposition system for compressed polarimetric SAR data according to claim 7, characterized in that, The specific implementation steps of the power decomposition module are as follows: Step S40, Based on the discriminator and scattering model , obtain power , and : default and Zero; Therefore, the five-component power is obtained as follows: in, Indicates surface scattering power, Indicates the power of bielastic scattering. Indicates volume scattering power, Indicates the surface volume scattering power, Indicates the power of the two projectiles scattering; Step S41, based on the scattering model , obtain power , ,and : default and Zero; Therefore, the five-component power is obtained as follows: in: Step S42, based on the scattering model , obtain power , ,and : default and Zero; Therefore, the five-component power is obtained as follows: in: 。

Images