Clutter sequence simulation method

Abstract

The invention discloses a clutter sequence simulation method. The method comprises the following steps: firstly, acquiring clutters, and dividing clutter categories; then, according to the clutter category, a corresponding composite Gaussian model is adopted to simulate and sample clutters, and a random sequence is obtained; and finally, simulating the random sequence to obtain a clutter simulation sequence. The method is simple in algorithm design, high in operation efficiency and capable of generating large-scale data.

Description

Technical Field

[0001] The present invention relates to the field of signal processing, and more specifically, to a noise sequence simulation method. Background Art

[0002] Random sequences simulating specific distributions are often used in practical engineering to verify and evaluate the performance of various systems. In the field of radar, random sequences simulating specific distributions are mainly used to replace actual radar background clutter data to reduce the number of radar field test experiments and reduce the cost of manpower and material resources.

[0003] The traditional method of simulating random sequences with specific distributions takes into account the power spectrum characteristics of clutter, so the algorithm design is complex and the computational efficiency is low, making it difficult to generate large-scale data.

[0004] The prior art discloses a method, system, device and storage medium for estimating sea clutter parameters. The method first obtains measured sea clutter data, estimates the shape parameters of the measured sea clutter data using a relationship model between shape parameters and multiple feature points in a multi-point fitting method, and finally establishes a final K-distributed sea clutter model based on the estimated shape parameters using a spherical invariant random process method. The algorithm design of this method is complex and the computational efficiency is low. Summary of the invention

[0005] The present invention provides a noise sequence simulation method to solve the defects of the prior art, such as complex algorithm design, low computational efficiency and difficulty in realizing large-scale data generation.

[0006] The primary purpose of the present invention is to solve the above technical problems. The technical solution of the present invention is as follows:

[0007] A noise sequence simulation method, comprising:

[0008] S1: Obtain clutter and classify it;

[0009] S2: According to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence;

[0010] S3: Simulate the random sequence to obtain a clutter simulation sequence.

[0011] Furthermore, the clutter categories include K distribution, generalized Pareto distribution, and IG-CG distribution.

[0012] Furthermore, the probability density function of the K distribution is as follows:

[0013]

[0014] Among them, K v-1represents the v-1th order second kind modified Bessel function, x represents fast-varying clutter data, c represents the scale parameter, v represents the shape parameter, Γ represents the gamma function, and τ represents the texture component.

[0015] Furthermore, the probability density function of the generalized Pareto distribution is as follows:

[0016]

[0017] Among them, a represents the shape parameter, b represents the scale parameter, Γ represents the gamma function, τ represents the texture component, x represents the fast-changing clutter data, and y represents the slowly changing clutter data.

[0018] Furthermore, the probability density function of the IG-CG distribution is as follows:

[0019]

[0020] Among them, λ represents the shape parameter, μ represents the scale parameter, τ represents the texture component, and x represents the fast-changing clutter data.

[0021] Furthermore, the sampling method is divided into direct sampling and selection sampling;

[0022] Among them, K distribution and generalized Pareto distribution use direct sampling, and IG-CG distribution uses cut-off sampling.

[0023] Furthermore, the generalized Pareto distribution sampling formula is as follows:

[0024]

[0025] Among them, u i represents the i-th uniform distribution value on the interval [0, 1], a represents the shape parameter, b represents the scale parameter, and Z i Represents the i-th sampling result.

[0026] Furthermore, the sampling formula of the K distribution is as follows:

[0027]

[0028]

[0029]

[0030] Where k represents the degrees of freedom of the gamma distribution, ω represents the exponential distribution parameter, and u i represents the i-th uniformly distributed value on the interval [0, 1], p i represents the sampling result of the i-th exponential distribution, q i Represents the sampling result of the i-th gamma distribution.

[0031] Furthermore, the sampling results of the IG-CG distribution need to meet the following sampling conditions:

[0032]

[0033] Among them, h 1i represents a random number in the interval [0, 1], h 2i represents a random number in the interval [d, g], where d and g are the two endpoints of the interval [d, g], and f max is the maximum value of the probability density function of the IG-CG distribution, f(d+(gd)h 1i ) represents the sampling results of IG-CG distribution.

[0034] A noise sequence simulation device, comprising:

[0035] Category classification module: obtain clutter and classify it into categories;

[0036] Simulation sampling module: according to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence;

[0037] Sequence simulation module: simulates the random sequence to obtain a clutter simulation sequence.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] The present invention obtains clutter and classifies the clutter; then, according to the clutter class, a corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence, thereby making the algorithm design simple and the operation efficiency high, and realizing the generation of large-scale data. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A flow chart of a noise sequence simulation method provided in an embodiment. DETAILED DESCRIPTION

[0041] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0042] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0043] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0044] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0045] Example

[0046] like Figure 1 As shown, a noise sequence simulation method comprises:

[0047] S1: Obtain clutter and classify it;

[0048] S2: According to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence;

[0049] S3: Simulate the random sequence to obtain a clutter simulation sequence.

[0050] In a specific embodiment, the case where the noise has an explicit statistical expression and is decomposable is classified as K distribution; the case where the noise has an explicit statistical expression is classified as generalized Pareto distribution; and the case where the noise does not have an explicit statistical expression is classified as IG-CG distribution.

[0051] Furthermore, the clutter categories include K distribution, generalized Pareto distribution, and IG-CG distribution.

[0052] It should be noted that the K distribution, generalized Pareto distribution, and IG-CG distribution correspond to the K distribution composite Gaussian model, the generalized Pareto distribution composite Gaussian model, and the IG-CG distribution composite Gaussian model, respectively. The K distribution composite Gaussian model, the generalized Pareto distribution composite Gaussian model, and the IG-CG distribution composite Gaussian model are all composite Gaussian models. The abbreviation of the composite Gaussian model is CG, which is composed of the product of a positive texture component and a speckle component that obeys a complex Gaussian distribution. In the field of sea clutter, the texture component is generated by scatterers of large-scale structures on the sea surface, while the texture component depends on a large number of small scatterers.

[0053] It should be noted that the K-distribution composite Gaussian model has high simulation accuracy and fast running speed, but it cannot simulate random sequences with shape parameters of arbitrary real numbers. The generalized Pareto distribution composite Gaussian model has a simple sampling form, the overall scheme is easy to implement, and can simulate and generate random sequences with arbitrary distribution parameters. The simulation of random sequences of the IG-CG distribution composite Gaussian model is not restricted by the values ​​of the distribution parameters, but the operation speed is slow. This method combines the advantages of the above three schemes to provide a noise sequence simulation method with simple algorithm design, high operation efficiency, and the ability to generate large-scale data.

[0054] In a specific embodiment, the product of the composite Gaussian model is as follows:

[0055]

[0056] Among them, U represents the speckle component, which obeys the complex Gaussian distribution with zero mean and unit covariance; τ represents the texture component. The texture component τ corresponding to the K-distribution compound Gaussian model obeys the gamma distribution; the texture component τ corresponding to the generalized Pareto distribution compound Gaussian model obeys the inverse gamma distribution, that is, It conforms to the gamma distribution; the texture component τ corresponding to the IG-CG distribution composite Gaussian model obeys the inverse Gaussian distribution.

[0057] In a specific embodiment, the probability density function of the composite Gaussian model is as follows:

[0058]

[0059] Among them, f τ (τ; θ) represents the clutter K distribution amplitude model, f U (x|τ) represents the probability density function of the speckle component. U (x|τ) is also called the probability density function of the clutter amplitude condition, which is usually in the form of Rayleigh distribution. The probability density function of the compound Gaussian model can be used to derive the probability density function of the K-distribution compound Gaussian model, the probability density function of the generalized Pareto distribution compound Gaussian model, and the probability density function of the IG-CG distribution compound Gaussian model.

[0060] The probability density function of the speckle component is as follows:

[0061]

[0062] When the probability density function of the texture component τ obeys the gamma distribution, the K distribution amplitude model is as follows:

[0063]

[0064] Furthermore, the probability density function of the K distribution is as follows:

[0065]

[0066] Among them, K v-1 represents the v-1th order second kind modified Bessel function, x represents fast-varying clutter data, c represents the scale parameter, v represents the shape parameter, Γ represents the gamma function, and τ represents the texture component.

[0067] It should be noted that x≥0, v≥0, c≥0.

[0068] In a specific embodiment, v=2, c=1.

[0069] Furthermore, the probability density function of the generalized Pareto distribution is as follows:

[0070]

[0071] Among them, a represents the shape parameter, b represents the scale parameter, Γ represents the gamma function, τ represents the texture component, x represents the fast-changing clutter data, and y represents the slowly changing clutter data.

[0072] It should be noted that x≥0, a≥0, and b≥0.

[0073] In a specific embodiment, a=0.5, b=1.5.

[0074] Furthermore, the probability density function of the IG-CG distribution is as follows:

[0075]

[0076] Among them, λ represents the shape parameter, μ represents the scale parameter, τ represents the texture component, and x represents the fast-changing clutter data.

[0077] It should be noted that x≥0, λ≥0, and μ≥0.

[0078] In a specific embodiment, λ=1.5, μ=3.5.

[0079] It should be noted that the probability density function of the inverse Gaussian distribution is as follows:

[0080]

[0081] Furthermore, the sampling method is divided into direct sampling and selection sampling;

[0082] Among them, K distribution and generalized Pareto distribution use direct sampling, and IG-CG distribution uses cut-off sampling.

[0083] Furthermore, the generalized Pareto distribution sampling formula is as follows:

[0084]

[0085] Among them, u i represents the i-th uniform distribution value on the interval [0, 1], a represents the shape parameter, b represents the scale parameter, z i Represents the i-th sampling result.

[0086] It should be noted that the sampling formula of the generalized Pareto distribution can be derived from the cumulative distribution function of the generalized Pareto distribution.

[0087] The cumulative distribution function of the generalized Pareto distribution is as follows:

[0088]

[0089]

[0090] x i Represents the i-th fast-changing clutter data.

[0091] Furthermore, the sampling formula of the K distribution is as follows:

[0092]

[0093]

[0094]

[0095] Where k represents the degrees of freedom of the gamma distribution, ω represents the exponential distribution parameter, and u i represents the i-th uniformly distributed value on the interval [0, 1], p i represents the sampling result of the i-th exponential distribution, q i Represents the sampling result of the i-th gamma distribution.

[0096] Furthermore, the sampling results of the IG-CG distribution need to meet the following sampling conditions:

[0097]

[0098] Among them, h 1i represents a random number in the interval [0, 1], h 2i represents a random number in the interval [d, g], where d and g are the two endpoints of the interval [d, g], and f max is the maximum value of the probability density function of the IG-CG distribution, f(d+(gd)h 1i ) represents the sampling results of IG-CG distribution.

[0099] A noise sequence simulation device, comprising:

[0100] Category classification module: obtain clutter and classify it into categories;

[0101] Simulation sampling module: according to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence;

[0102] Sequence simulation module: simulates the random sequence to obtain a clutter simulation sequence.

[0103] The same or similar reference numerals correspond to the same or similar components;

[0104] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;

[0105] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the embodiments of the present invention. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the embodiments here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for simulating a clutter sequence, characterized in that: include: S1: Obtain clutter and classify it; S2: According to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence; S3: Simulate the random sequence to obtain a clutter simulation sequence.

2. The method for simulating a clutter sequence according to claim 1, characterized in that: The clutter categories include K distribution, generalized Pareto distribution, and IG-CG distribution.

3. A method for simulating a clutter sequence according to claim 2, characterized in that: The probability density function of the K distribution is as follows: Among them, K v-1 represents the v-1th order second kind modified Bessel function, x represents fast-varying clutter data, c represents the scale parameter, v represents the shape parameter, Γ represents the gamma function, and τ represents the texture component.

4. The method for simulating a clutter sequence according to claim 2, characterized in that: The probability density function of the generalized Pareto distribution is as follows: Among them, a represents the shape parameter, b represents the scale parameter, Γ represents the gamma function, τ represents the texture component, x represents the fast-changing clutter data, and y represents the slowly changing clutter data.

5. The method for simulating a clutter sequence according to claim 2, characterized in that: The probability density function of the IG-CG distribution is as follows: Among them, λ represents the shape parameter, μ represents the scale parameter, τ represents the texture component, and x represents the fast-changing clutter data.

6. A noise sequence simulation method according to claim 2, characterized in that: The sampling methods are divided into direct sampling and selection sampling; Among them, K distribution and generalized Pareto distribution use direct sampling, and IG-CG distribution uses cut-off sampling.

7. A noise sequence simulation method according to claim 6, characterized in that: The generalized Pareto distribution sampling formula is as follows: Among them, u i represents the i-th uniform distribution value on the interval [0,1], a represents the shape parameter, b represents the scale parameter, z i Represents the i-th sampling result.

8. A noise sequence simulation method according to claim 6, characterized in that: The sampling formula of the K distribution is as follows: Where k represents the degrees of freedom of the gamma distribution, ω represents the exponential distribution parameter, and u i represents the i-th uniform distribution value on the interval [0,1], p i represents the sampling result of the i-th exponential distribution, q i Represents the sampling result of the i-th gamma distribution.

9. A noise sequence simulation method according to claim 6, characterized in that: The sampling results of the IG-CG distribution must meet the following sampling conditions: Among them, h 1i represents a random number in the interval [0,1], h 2i represents a random number in the interval [d,g], where d and g are the two endpoints of the interval [d,g], and f max is the maximum value of the probability density function of the IG-CG distribution, f(d+(gd)h 1i ) represents the sampling results of IG-CG distribution.

10. A noise sequence simulation device, applied to the simulation method according to any one of claims 1 to 9, characterized in that: include: Category classification module: obtain clutter and classify it into categories; Simulation sampling module: according to the clutter category, the corresponding composite Gaussian model is used to simulate and sample the clutter to obtain a random sequence; Sequence simulation module: simulates the random sequence to obtain a clutter simulation sequence.

Images