CN111856421A - A method and device for feature extraction and radar target enhancement in polarization rotation domain - Google Patents

Abstract

The invention discloses a method and device for feature extraction in polarization rotation domain and radar target enhancement. The method includes: performing rotation processing on an acquired polarization scattering matrix around the line of sight of the polarization radar to obtain a rotating polarization scattering matrix; Rotate the polarization scattering matrix to extract the scattering vector; extract the polarization channel combination based on the scattering vector, and visualize its polarization correlation value to obtain the polarization correlation pattern; extract the following polarization correlation pattern features from the polarization correlation pattern to carry out Parametric characterization: original polarization-related eigenvalue, polarization-related characteristic maximum value, polarization-related characteristic minimum value, polarization-related degree, polarization-related fluctuation degree, polarization-related contrast, polarization-related anti-entropy, maximum rotation angle, minimized rotation angle, polarization correlation width. Solve the problems in the prior art that the target is not prominent and difficult to identify, and achieve target enhancement.

Description

A method and device for feature extraction and radar target enhancement in polarization rotation domain

technical field

The invention relates to the technical field of radar polarization information processing and application, in particular to a method and device for feature extraction in polarization rotation domain and radar target enhancement.

Background technique

Polarization radar can obtain the polarization information of the target by sending and receiving a set of electromagnetic waves whose polarization states are orthogonal. Mainstream sensors in many important fields such as observation, air defense and missile defense, weather detection, and sea surface monitoring have been widely used.

The polarization information of the target obtained by the polarization radar can be characterized by the polarization scattering matrix. The polarization correlation characteristics between different polarization channels are sensitive to the relative geometric relationship between the target attitude and the radar line of sight. For the same target, when its attitude relative to the polarization radar line of sight is different, its polarization scattering characteristics may be significantly different. This phenomenon causes a lot of inconvenience to the processing and application of radar polarization information, and is one of the technical bottlenecks facing the current polarization radar target polarization scattering mechanism fine interpretation and quantitative application.

The polarization data obtained under specific imaging geometry is rotated around the direction of the radar line of sight and extended to the polarization rotation domain for analysis. It is expected to mine the hidden polarization information of the target and achieve radar target enhancement, becoming an improved target scatterer The key to understanding translation and application performance. Therefore, it is of great value to develop a method and device for feature extraction and radar target enhancement in polarization rotation domain.

SUMMARY OF THE INVENTION

The present invention provides a method and device for feature extraction in polarization rotation domain and radar target enhancement, which are used to overcome the defects of the prior art, such as low contrast of radar targets, and target scattering characteristics that are sensitive to the relative geometric relationship between target attitude and radar line of sight, etc. The polarization correlation value between the two polarization channels is visualized and parameterized in the polarization rotation domain, which can extract the characteristics of the polarization rotation domain and effectively enhance the identifiability of the radar target. It is a polarized radar system for air surveillance, earth observation, meteorological detection, sea surface surveillance, etc.), and has application value in the fields of air target classification and recognition, ground object classification, and damage assessment.

In order to achieve the above objects, the present invention provides a method for feature extraction and radar target enhancement in polarization rotation domain, including:

Step 1: Rotate the obtained polarization scattering matrix around the line of sight of the polarization radar to obtain a rotating polarization scattering matrix;

Step 2, extract the Pauli vector and the Lexicographic vector based on the rotational polarization scattering matrix;

Step 3, extract the polarization channel combination from the Pauli vector and the Lexicographic vector respectively, and visualize the polarization correlation value of the polarization channel combination to obtain the polarization correlation pattern;

Step 4: Extract the following polarization-related pattern features from the polarization-related pattern for parametric characterization: original polarization-related feature value, polarization-related feature maximum value, polarization-related feature minimum value, polarization correlation degree, polar Polarization-dependent fluctuation, polarization-dependent contrast, polarization-dependent inverse entropy, maximize rotation angle, minimize rotation angle, polarization-dependent width.

In order to achieve the above object, the present invention also provides a device for feature extraction and radar target enhancement in polarization rotation domain, including a memory and a processor, wherein the memory stores a program for feature extraction in polarization rotation domain and radar target enhancement. The processor executes the steps of the above method when running the program for feature extraction and radar target enhancement in the polarization rotation domain.

The method and device for feature extraction in the polarization rotation domain and radar target enhancement provided by the present invention can visualize and parameterize the characteristics of the polarization correlation value of the target in the rotation domain around the radar line of sight by extracting the features in the polarization rotation domain. The feature parameters of the polarization correlation pattern are extracted for subsequent applications such as radar target enhancement. The invention is simple and convenient to realize, and can be directly applied to the target polarization scattering matrix data obtained by polarization radar systems with different purposes. The invention has important reference value for application fields such as earth observation, sea surface monitoring, disaster mitigation and prevention.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention, and for those of ordinary skill in the art, other drawings can also be obtained according to the structures shown in these drawings without creative efforts.

FIG. 1 is a flowchart of a method for feature extraction in polarization rotation domain and radar target enhancement proposed in Embodiment 1 of the invention;

Figure 2(a) is the RGB pseudo-color map under the Pauli basis decomposition of the measured polarimetric synthetic aperture radar data;

Figure 2(b) The true value map of the ship target in the measured polarimetric synthetic aperture radar data;

Figure 3 is a schematic diagram of the relevant characteristics of the rotation domain polarization under the horizontal and vertical polarization base; wherein:

Figure 3(a) shows the polarization correlation characteristics of the rotation domain derived from the polarization correlation values between the polarization channels _SHH and S _HV ;

Figure 3(b) shows the polarization correlation characteristics of the rotation domain derived from the polarization correlation values between the polarization channels S _HH and S _VV ;

Figure 3(c) shows the polarization correlation characteristics of the rotation domain derived from the polarization correlation values between the polarization channels S _HH + S _VV and S _HH -S _VV ;

Figure 3(d) shows the polarization correlation characteristics of the spin domain derived from the polarization correlation values between the polarization channels _SHH - _SVV and _SHV ;

Figure 4. Comparison of radar target enhancement performance based on contrast analysis.

The realization, functional characteristics and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

It should be noted that all directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present invention are only used to explain the relationship between various components under a certain posture (as shown in the accompanying drawings). The relative positional relationship, the movement situation, etc., if the specific posture changes, the directional indication also changes accordingly.

In addition, descriptions such as "first", "second", etc. in the present invention are only for descriptive purposes, and should not be construed as indicating or implying their relative importance or implicitly indicating the number of indicated technical features. Thus, a feature delimited with "first", "second" may expressly or implicitly include at least one of that feature. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise expressly and specifically defined.

In the present invention, unless otherwise expressly specified and limited, the terms "connected", "fixed" and the like should be understood in a broad sense, for example, "fixed" may be a fixed connection, a detachable connection, or an integrated; It can be a mechanical connection, an electrical connection, a physical connection or a wireless communication connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction between the two elements. unless otherwise expressly qualified. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but must be based on the realization by those of ordinary skill in the art. When the combination of technical solutions is contradictory or cannot be realized, it should be considered that the combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

Example 1

As shown in FIG. 1 , an embodiment of the present invention provides a method for feature extraction in polarization rotation domain and radar target enhancement, which mainly consists of four steps:

为例。其中S_HH为在水平极化发射和水平极化接收条件下获取的复后向散射系数；S_HV为在水平极化发射和垂直极化接收条件下获取的复后向散射系数；S_VH为在垂直极化发射和水平极化接收条件下获取的复后向散射系数；S_VV为在垂直极化发射和垂直极化接收条件下获取的复后向散射系数。The polarization scattering matrix obtained by the polarization radar serves as the input of the present invention. The polarization scattering matrix of the target in the horizontal and vertical polarization base

For example. where S _HH is the complex backscattering coefficient obtained under the conditions of horizontal polarization transmission and horizontal polarization receiving; S _HV is the complex backscattering coefficient obtained under the conditions of horizontal polarization emission and vertical polarization receiving; S _VH is The complex backscattering coefficient obtained under the condition of vertical polarization transmission and horizontal polarization receiving; S _VV is the complex backscattering coefficient obtained under the condition of vertical polarization emission and vertical polarization receiving.

In the first step, the polarization scattering matrix is rotated around the radar line of sight to obtain the polarization scattering matrix in the polarization rotation domain; before the first step, the polarization scattering matrix of the target is obtained based on the polarization radar, and then the first step is carried out. , the obtained polarization scattering matrix is rotated around the line-of-sight direction of the polarization radar to obtain the polarization scattering matrix in the polarization rotation domain, that is, the rotating polarization scattering matrix;

Specifically, in the direction of the line of sight of the polarized radar, the polarization scattering matrix S is rotated, and the rotated polarization scattering matrix S after the rotation processing is calculated for the rotation angle θ, θ∈[-π,π] in the rotation domain. The expression of (θ) is:

上标^T为转置处理。Among them, the rotation matrix

The superscript ^T is the transposition process.

Lexicographic矢量定义为

The second step is to extract the Pauli vector and Lexicographic vector based on the rotational polarization scattering matrix; where the Pauli vector is defined as

A Lexicographic vector is defined as

和

对于极化通道X(S_X)和Y(S_Y)，其极化相关值拓展到旋转域的极化相关方向图定义为：In the third step, a total of six polarization channel combinations are extracted from the Pauli vector and the Lexicographic vector, and the polarization correlation values between the polarization channels are visualized to obtain six polarization correlation patterns, which are

and

For polarization channels X(S _X ) and Y(S _Y ), the polarization correlation pattern whose polarization correlation value extends to the rotation domain is defined as:

且

因此剩余的四个极化相关方向图

留作进一步分析。Verifiable

and

So the remaining four polarization-dependent patterns

Reserved for further analysis.

极化相关特征最大值

极化相关特征最小值

极化相关度

极化相关起伏度

极化相关对比度

极化相关反熵

最大化旋转角

最小化旋转角

和极化相关宽度BW_α等刻画参数。其中下标X和Y分别表示两个不同的极化通道。The fourth step is the parametric characterization of the polarization correlation pattern, that is, the following characteristic parameters are extracted from each polarization correlation pattern for parametric characterization: Extract the original polarization correlation eigenvalues

Polarization dependent characteristic maximum

Polarization Dependent Feature Minimum

Polarization correlation

Polarization Dependent Fluctuation

Polarization Dependent Contrast

Polarization Dependent Anti-Entropy

maximize the rotation angle

Minimize the rotation angle

and the polarization-dependent width BW _α and other characterization parameters. The subscripts X and Y represent two different polarization channels, respectively.

The polarization-dependent pattern characteristic of the polarization-dependent pattern in the rotating domain characterizes the scattering characteristics of a polarized radar target in the rotating domain around the radar line of sight. In order to facilitate the use of the polarization correlation pattern, the following characteristic parameters are extracted to describe it parametrically:

为：1. Raw polarization-related eigenvalues

for:

为：2. Polarization-dependent characteristic maximum

for:

为：3. Polarization Dependent Feature Minimum

for:

为：4. Polarization correlation

for:

为：5. Polarization Dependent Fluctuations

for:

为：6. Polarization Dependent Contrast

for:

7. Polarization Dependent Anti-Entropy

为：8. Maximize the rotation angle

for:

为：9. Minimize the rotation angle

for:

10. The polarization correlation width BW _α is:

且

其中，max{·}为求序列的最大值；min{·}为求序列的最小值；mean{·}为求序列的均值；std{·}为求序列的标准差；α为调节因子，通常取α＝0.95。BW _α = θ″-θ′, where

and

Among them, max{·} is the maximum value of the sequence; min{·} is the minimum value of the sequence; mean{·} is the mean value of the sequence; std{·} is the standard deviation of the sequence; α is the adjustment factor, Usually take α=0.95.

Combining the polarization correlation pattern and the above characterization parameters, a visualization method of the polarization correlation characteristics of the radar target is obtained. Using the above parameters to characterize the change characteristics of the polarization correlation value in the polarization rotation domain, a series of polarization features can be extracted for radar target enhancement, which provides support for target detection and classification and identification applications.

Correlation features between polarized channels in polarimetric data are sensitive to the relative geometry of target attitude and radar line-of-sight. The value of polarization-related features of the same target may change significantly under different attitude conditions. The polarization scattering matrix obtained under the specific imaging geometry is rotated around the radar line of sight, which can change the relative geometric relationship between the target attitude and the radar line of sight. Visual processing and parametric characterization of the polarization correlation pattern derived from the polarization correlation value in the rotation domain can completely describe the change characteristics of the target polarization correlation value in the rotation domain, and accurately interpret the target in the polarization rotation domain. scattering mechanism in . In this way, feature extraction and radar target enhancement can be realized in the polarization rotation domain, which can be used in the fields of physical parameter inversion and target detection and classification.

FIG. 2 takes the measured data of polarimetric synthetic aperture radar (Synthetic Aperture Radar, SAR) as an example, which is used for intuitive analysis of the method and device proposed in the present invention. Figure 2(a) is an RGB pseudocolor map under Pauli basis decomposition. Figure 2(b) is the true value map of the measured data, in which the black pixels are the sea clutter background, the white pixels are the ship targets, and the gray is the land area. It should be pointed out that the present invention is not only applicable to polarimetric SAR, but also applicable to polarimetric radars of various other systems.

极化相关特征最大值

极化相关特征最小值

极化相关度

极化相关起伏度

极化相关对比度

等特征中，舰船之间的取值相近，且与海杂波的取值差异较大。相较于原始极化SAR图像，雷达图像中的舰船目标从目视的角度得到显著增强。Figure 3. Feature extraction results of polarization rotation domain based on measured data. Figure 3(a)-(d) represent the four sets of polarization correlation pattern characteristics derived from the combination of four sets of polarization correlation channels, respectively. From the visual analysis, it can be seen that there are obvious differences in the values of the ship target and the sea clutter background in the polarization correlation pattern characteristics. In particular, in the original polarization-dependent eigenvalues

Polarization dependent characteristic maximum

Polarization Dependent Feature Minimum

Polarization correlation

Polarization Dependent Fluctuation

Polarization Dependent Contrast

Among other characteristics, the values between ships are similar, and the values of sea clutter are quite different. Compared with the original polarimetric SAR image, the ship target in the radar image is significantly enhanced from the visual point of view.

Figure 4 is a comparison chart of radar target enhancement performance based on contrast analysis. Contrast (Target to Clutter Ratio, TCR) is defined as the ratio of the polarization correlation pattern feature values of the target area and the clutter area in the polarization correlation pattern feature. In a polarimetric SAR image containing a ship target, it is the ratio of the feature size of the polarization correlation pattern in the ship area and the sea clutter area:

In the above formula, the numerator represents the mean value of the ship area, and the denominator represents the mean value of the sea clutter area. The TCR reflects the magnitude of the difference between the target and the background in the polarization feature map. The larger the TCR, the more prominent the target is in the background, and the more obvious the radar target enhancement effect is.

In this example, TCR is selected as the index to analyze the performance of different polarization characteristics in radar target enhancement. The total power SPAN and polarization correlation pattern characteristics of polarimetric SAR data are selected for comparison. Among them, the characteristic of the polarization correlation pattern is to extend the polarization correlation value between polarization channels to the polarization rotation domain. For polarization channels X and Y, the polarization correlation pattern whose polarization correlation value extends to the rotation domain is defined as:

The original polarization-related eigenvalues |γ _XY | _org , the maximum polarization-related characteristics |γ _XY | _max , the minimum polarization-related characteristics |γ _XY | _min , the polarization correlation degree |γ _XY | _mean , the polarization Correlation fluctuation |γ _XY | _std , polarization-dependent contrast |γ _XY | _contrast , polarization-dependent inverse entropy |γ _XY | _Ani , maximum rotation angle θ _γ-max , minimum rotation angle θ _γ-min and polarization Correlation width BW _α and other characterization parameters, the definition of polarization-related characteristic quantity is similar to that of polarization-related characteristic quantity, the difference is that the definition of polarization-related pattern is different. In Figure 4, dark gray is the TCR of SPAN; light gray represents the TCR of polarization-dependent pattern features; black represents the TCR of polarization-dependent pattern features. Except for SPAN, other polarized spin domain features are divided into four groups according to the different combinations of polarized channels. In the polarization-related features, each group of features from left to right represents polarization-related inverse entropy, polarization-related contrast, polarization-related maximum value, polarization-related correlation degree, polarization-related minimum value, polarization-related fluctuation degree, and polarization-related correlation characteristics. Correlation raw value. In the polarization-related features, each group of features from left to right represents polarization-related inverse entropy, polarization-related contrast, polarization-related maximum value, polarization-related correlation degree, polarization-related minimum value, polarization-related fluctuation degree, and polarization-related correlation characteristics. Correlation raw value.

和

其TCR分别为449，209，119。TCR最高的极化相关方向图特征

其性能相较于SPAN提升了15dB以上，相较于TCR最高的极化相关特征提升了22dB以上，该极化相关方向图特征用于极化雷达目标增强的性能最好。由此，对比实验验证了提取的极化旋转域特征可有效应用于雷达目标增强。Polarimetric SAR data total power SPAN with a TCR of 13. Compared with SPAN, the TCR of the polarization-dependent pattern features is smaller and neither higher than 3, as shown in the enlarged view in the rectangular box at the bottom of Fig. 4. Among them, the polarization correlation pattern with the highest TCR is |γ _(HH-VV)-(HV) (θ)| _org , and its TCR is 3. Compared with SPAN and polarization-related pattern features, the TCR of polarization-related pattern features is significantly improved, that is, the radar target is significantly enhanced. In particular, the three polarization correlation pattern features with the highest TCR are:

and

Its TCRs are 449, 209, 119, respectively. Polarization-dependent pattern features with the highest TCR

Compared with SPAN, its performance is improved by more than 15dB, and the polarization correlation feature with the highest TCR is improved by more than 22dB. The polarization correlation pattern feature has the best performance for polarization radar target enhancement. Therefore, the comparative experiment verifies that the extracted polarization rotation domain features can be effectively applied to radar target enhancement.

Embodiment 2

Based on the first embodiment, the present invention provides an apparatus for feature extraction and radar target enhancement in polarization rotation domain, including a memory and a processor, wherein the memory stores a program for feature extraction in polarization rotation domain and radar target enhancement. The processor executes the steps of any of the foregoing method embodiments when running the program for feature extraction in the polarization rotation domain and radar target enhancement.

The above descriptions are only the preferred embodiments of the present invention, and are not intended to limit the scope of the present invention. Under the inventive concept of the present invention, the equivalent structural transformations made by the contents of the description and drawings of the present invention, or the direct/indirect application Other related technical fields are included in the scope of patent protection of the present invention.

Claims (10)

1. a method for polarization rotation domain feature extraction and radar target enhancement, is characterized in that, comprises: Step 1: Rotate the obtained polarization scattering matrix around the line of sight of the polarization radar to obtain a rotating polarization scattering matrix; Step 2, extract the Pauli vector and the Lexicographic vector based on the rotational polarization scattering matrix; Step 3, extract the polarization channel combination from the Pauli vector and the Lexicographic vector respectively, and visualize the polarization correlation value of the polarization channel combination to obtain the polarization correlation pattern; Step 4: Extract the following polarization-related pattern features from the polarization-related pattern for parametric characterization: original polarization-related feature value, polarization-related feature maximum value, polarization-related feature minimum value, polarization correlation degree, polar Polarization-dependent fluctuation, polarization-dependent contrast, polarization-dependent inverse entropy, maximize rotation angle, minimize rotation angle, polarization-dependent width. 2. The method for feature extraction of polarization rotation domain and radar target enhancement as claimed in claim 1, wherein in the step 1, the target polarization scattering matrix under the horizontal and vertical polarization base obtained by the polarization radar is:

Among them, S _HH is the complex backscattering coefficient obtained under the conditions of horizontal polarization transmission and horizontal polarization receiving; S _HV is the complex backscattering coefficient obtained under the conditions of horizontal polarization emission and vertical polarization receiving; S _VH is the complex backscattering coefficient obtained under the conditions of vertical polarization transmitting and horizontal polarization receiving; S _VV is the complex backscattering coefficient obtained under the conditions of vertical polarization transmitting and vertical polarization receiving; The step 1 includes: In the direction of the line of sight of the polarized radar, the polarization scattering matrix S is rotated. For the rotation angle θ in the rotation domain, the expression of the rotated polarization scattering matrix S(θ) after the rotation processing is calculated as:

Among them, the rotation matrix R ₂ (θ) is:

The superscript ^T is the transposition process, and the value range of the rotation angle θ is θ∈[-π,π]. 3. The method for polarization rotation domain feature extraction and radar target enhancement as claimed in claim 2, wherein step 2 comprises: For any two polarization channels S _X and S _Y , a polarization correlation pattern can be obtained by visualizing the polarization correlation values of the polarization channels S _X and S _Y in the rotation domain, which is defined as:

Among them, <·> is the set average processing, |·| is the absolute value processing, and the superscript ^* is the conjugate processing. 4. The method for feature extraction of polarization rotation domain and radar target enhancement as claimed in claim 3, characterized in that, in step 4, the following ten polarization correlation pattern features are extracted from each polarization correlation pattern. Characterize and characterize polarization-dependent patterns: Raw polarization-dependent eigenvalues

for:

Polarization dependent characteristic maximum

for:

Polarization Dependent Feature Minimum

for:

Polarization correlation

for:

Polarization Dependent Fluctuation

for:

Polarization Dependent Contrast

for:

Polarization Dependent Anti-Entropy

for:

maximize the rotation angle

for:

Minimize the rotation angle

for:

The polarization correlation width BW _α is: BW _α = θ″-θ′, where

and

Among them, max{·} is the maximum value of the sequence; min{·} is the minimum value of the sequence; mean{·} is the mean value of the sequence; std{·} is the standard deviation of the sequence; α is the adjustment factor, Take α=0.95. 5. The method for feature extraction and radar target enhancement in the polarization rotation domain according to any one of claims 1 to 4, characterized in that, further comprising based on the feature contrast based on the feature of the polarization correlation pattern extracted in step 4 This metric prefers polarization correlation pattern features for target enhancement. 6 . The method for feature extraction and radar target enhancement in polarization rotation domain according to claim 5 , wherein the polarization-related feature contrast is defined as the value of the target area and the value of the clutter area in the polarization direction. 7 . The ratio under the graph feature. 7. The method for polarization rotation domain feature extraction and radar target enhancement as claimed in claim 6, wherein the polarization correlation contrast TCR expression is:

In the above formula, the numerator Ship _mean represents the mean value of the feature values of the polarization-related characteristic patterns in the ship region, and the denominator Clutter _mean represents the mean value of the polarization-related pattern feature values in the sea clutter region. 8. The method for feature extraction and radar target enhancement in polarization rotation domain as claimed in claim 7, wherein the three polarization correlation pattern features with higher TCR are respectively

and

in,

represents the original polarization-dependent eigenvalues between polarization channels _SHH -S _VV and S _HV ,

represents the raw polarization-dependent eigenvalues between polarization channels _SHH and _SHV ,

represents the polarization-dependent characteristic minimum between the polarization channels S _HH -S _VV and S _HV , with the rotation angle θ∈[-π,π]. 9. The method for feature extraction in polarization rotation domain and radar target enhancement as claimed in claim 8, wherein the polarization correlation pattern feature with the highest TCR is:

10. An apparatus for feature extraction and radar target enhancement in polarization rotation domain, comprising a memory and a processor, wherein the memory stores a program for feature extraction in polarization rotation domain and radar target enhancement, and the processor is running Steps of the method according to any one of claims 1 to 9 are executed when the program of feature extraction in the polarization rotation domain and radar target enhancement is performed.

Images