A method for partitioning the sea surface based on the parameter estimation of the sea clutter model

Abstract

The present invention proposes a method for partitioning the sea surface based on the estimation of sea clutter model parameters. For the fast-time - slow-time video data received by the radar, a parametric model including shape parameters is used to model the amplitude data distribution on each range cell, and the sea surface is divided into a clutter area and a noise area according to the robustness of the estimation of the model shape parameters. The present invention utilizes the difference in robustness between the model parameter estimation method in the sea clutter area and the noise area to divide the sea surface into a clutter area and a noise area, providing a basis for the selection of subsequent detection strategies.

Description

Technical Field

[0001] The invention relates to the field of radar technology, and is a method for realizing sea surface area division so as to use a detection strategy in a targeted manner. Background Art

[0002] In radar sea detection, sea clutter is one of the main natural environmental disturbances. Due to the large differences in the formation mechanism, amplitude distribution and other aspects of sea clutter and noise, different detection strategies need to be adopted under the sea clutter background and the noise background. Therefore, it is of obvious practical significance to clarify the boundaries of sea clutter and noise areas in radar echoes. At present, there are few studies on sea surface zoning at home and abroad. Some domestic scholars have divided the sea surface into regions based on the radar grazing angle to implement targeted detection strategies, but the regional division based on the grazing angle is more of an empirical judgment, and it is difficult to distinguish between clutter and noise areas. The sea surface zoning method can be improved from the aspects of clarifying the physical meaning of the zoning standard. Summary of the invention

[0003] Aiming at the problem of determining the boundary between the sea clutter area and the noise area in the radar echo, the present invention proposes a sea surface partitioning method based on sea clutter model parameter estimation.

[0004] The present invention proposes a sea surface partitioning method based on sea clutter model parameter estimation. For the fast-time-slow-time video data received by the radar, a parametric model containing shape parameters is used to model the amplitude data distribution on each distance unit, and the sea surface is divided into a clutter area and a noise area based on the robustness of the model shape parameter estimation. The specific implementation steps of the present invention are as follows:

[0005] Step 1: Use the sea detection radar to collect fast-time-slow-time video data at the target location, where the sea detection radar can adopt any working mode that can obtain fast-time-slow-time video data, such as single-transmit single-receive, single-transmit multiple-receive, multiple-transmit multiple-receive, etc.;

[0006] Step 2: Based on the prior knowledge of the radar system and the relevant sea area, a probability distribution model is selected as the mathematical model of the sea clutter amplitude distribution characteristics;

[0007] Step 3: Using a model parameter estimation method, all slow time data at each distance unit are used as samples to estimate the specific values ​​of the shape parameters in the parametric model;

[0008] Step 4: According to the robustness of the shape parameter estimation results, the area where the shape parameter estimation is robust is divided into the clutter area, and the area where the shape parameter estimation is not robust is divided into the noise area.

[0009] Further, the sea detection radar in step 1 can be single - transmit single - receive, single - transmit multi - receive, or multi - transmit multi - receive.

[0010] Further, the shape parameters of the parametric model in step 2 can be used to describe the sharpness of sea clutter. As the noise component in the sea clutter increases, the shape parameter can decrease or increase, and when the clutter degrades into noise, the shape parameter approaches 0 or infinity.

[0011] Further, the model parameter estimation methods in step 3 include the model fitting method, the moment estimation method, and the quantile method.

[0012] Further, the robustness of the variance evaluation method based on the shape parameter estimation results of consecutive multiple range cells in step 4 is considered. If the variance exceeds a preset threshold, it is considered that the model parameter estimation method fails, and the data of the corresponding range cell is divided into the noise area; otherwise, the data of the corresponding range cell is divided into the clutter area.

[0013] The present invention utilizes the robustness difference of the model parameter estimation method in the sea clutter area and the noise area to divide the sea surface into the clutter area and the noise area, providing a basis for the selection of subsequent detection strategies.

[0014] The following further describes the present invention in detail with reference to the accompanying drawings. Description of the Drawings

[0015] Figure 1 Flowchart of the sea surface partitioning method based on sea clutter model parameter estimation Detailed Embodiment

[0016] The following further illustrates the present invention with reference to the accompanying drawings, but the protection scope of the present invention is not limited by the embodiments.

[0017] The present invention proposes a sea surface partitioning method based on sea clutter model parameter estimation. Considering that the sea clutter model parameter estimation method can only maintain robustness when sea clutter is dominant, and loses robustness when noise is dominant, the sea surface is divided into the clutter area and the noise area according to the robustness of the parameter estimation results, providing a basis for the selection of subsequent detection strategies.

[0018] Combined with the accompanying drawings, the steps of the embodiment of the present invention are as follows:

[0019] Step 1: Use a sea detection radar to collect fast - time - slow - time video data in the target azimuth.

[0020] Step 2: According to the prior knowledge of the radar system and the relevant sea area, select a probability distribution model as the mathematical model for the amplitude distribution characteristics of sea clutter. The shape parameter of this parametric model can be used to describe the sharpness of sea clutter. As the noise component in the sea clutter increases, the shape parameter can decrease or increase, and when the clutter degenerates into noise, the shape parameter approaches 0 or infinity;

[0021] Step 3: Adopt a model parameter estimation method, using all the slow-time data at each range cell as samples to estimate the specific value of the shape parameter in the parametric model. The model parameter estimation methods include the model fitting method, the moment estimation method, and the quantile method;

[0022] Step 4: According to the robustness of the shape parameter estimation results, divide the region with robust shape parameter estimation into the clutter area, and divide the region with non-robust shape parameter estimation into the noise area. Based on the variance evaluation method of the shape parameter estimation results of consecutive multiple range cells for robustness, if the variance exceeds the preset threshold, it is considered that the model parameter estimation method fails, and the data of the corresponding range cell is divided into the noise area, otherwise the data of the corresponding range cell is divided into the clutter area.

Claims

1. A method for partitioning the sea surface based on the estimation of sea clutter model parameters, characterized in that: Step 1: Use a sea detection radar to collect fast time - slow time video data in the target azimuth; Step 2: According to the prior knowledge of the radar system and the relevant sea area, select a probability distribution model as the mathematical model of the amplitude distribution characteristics of sea clutter. The shape parameter of the mathematical model is used to describe the sharpness of sea clutter. As the noise component in the sea clutter increases, the shape parameter decreases or increases, and when the clutter degenerates into noise, the shape parameter approaches 0 or infinity; Step 3: Adopt a model parameter estimation method, and use all slow time data in each range cell as samples to estimate the specific value of the shape parameter in the parametric model; Step 4: According to the robustness of the shape parameter estimation results, divide the area with robust shape parameter estimation into the clutter area, and divide the area with non - robust shape parameter estimation into the noise area; wherein the evaluation criteria for the robustness include: If the variance of the shape parameter estimation results of multiple consecutive range cells exceeds a preset threshold, it is considered that the model parameter estimation method fails, and the data of the corresponding range cells are divided into the noise area, otherwise the data of the corresponding range cells are divided into the clutter area.

2. A method for partitioning the sea surface based on the estimation of sea clutter model parameters according to claim 1, characterized in that: The sea detection radar in Step 1 adopts single - transmit single - receive, single - transmit multi - receive, or multi - transmit multi - receive.

3. A method for partitioning the sea surface based on the estimation of sea clutter model parameters according to claim 1, characterized in that: The model parameter estimation method in Step 2 includes the model fitting method, the moment estimation method, and the quantile method.

Images