CN114879159A - Pretreated sea surface target detection method - Google Patents

Abstract

The present application discloses a preprocessed sea surface target detection method. The preprocessed sea surface target detection method includes: acquiring electromagnetic echo data of the water surface to be detected; removing baseline drift and filtering the electromagnetic echo data of the water surface to be detected, so as to obtain the preprocessed electromagnetic echo data of the water surface to be detected; Extracting the echo features of the preprocessed water surface electromagnetic echo data to be detected; inputting the echo features into the trained neural network detection model to determine whether the water surface electromagnetic echo data to be detected is sea clutter data . The invention is based on a water surface electromagnetic target detection method combining three features and deep learning, adopts a composite denoising method and processes data, further improves the accuracy of target detection, and can effectively reduce the false alarm rate of sea clutter detection. At the same time, the detection speed can be accelerated.

Description

Preprocessed sea surface target detection method

technical field

The invention relates to the field of water surface electromagnetic target detection, in particular to a method of water surface electromagnetic data acquisition, sea clutter feature extraction, sea clutter data processing, sea clutter amplitude distribution modeling and convolutional neural network applied to water surface electromagnetic target detection.

Background technique

The sea detection radar faces a complex detection background, and the echo signals it receives usually include sea clutter in addition to the target echoes of interest. The sea clutter power level is usually high, accompanied by significant non-Gaussian and non-stationary characteristics, and is easily affected by various complex and variable factors, which has become one of the main constraints affecting the performance of radar detection. With the refinement of observation methods, the background clutter and target echoes become so complex that accurate statistical modeling is difficult. In this case, the "double-height" system with high spatial resolution and high Doppler resolution is the main technical approach.

High-resolution small target detection radar needs to face extremely complex high-resolution sea clutter characteristics and small target echo characteristics, so the key to breaking through the critical signal-to-noise ratio detection performance lies in the deep cognition of sea clutter characteristics (deep cognition of sea clutter characteristics). cognition), elaborate perception, and full utilization. In this case, one or more differential features of clutter and target echo are usually used to achieve joint detection. Such methods are called feature-based detection technology, or feature detection technology for short.

The multi-feature-based detection method in the background of sea clutter is to extract the distinctive features from radar and target echoes, and reduce the dimension of the observation space with high overlap between the clutter and the target to the feature space with low overlap. target is detected.

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide a preprocessed sea surface target detection method to overcome or at least alleviate at least one of the above-mentioned defects of the prior art.

One aspect of the present invention provides a preprocessed sea surface target detection method, wherein the preprocessed sea surface target detection method includes:

The preprocessed sea surface target detection method includes:

Obtain the electromagnetic echo data of the water surface to be detected;

Remove baseline drift and filter the electromagnetic echo data of the water surface to be detected, so as to obtain the pre-processed electromagnetic echo data of the water surface to be detected;

extracting the echo features of the preprocessed water surface electromagnetic echo data to be detected;

The echo features are input into the trained neural network detection model to determine whether the electromagnetic echo data on the water surface to be detected is sea clutter data.

Optionally, the filtering processing of the electromagnetic echo data on the water surface to be detected includes:

Median filtering and wavelet threshold denoising are respectively performed on the water surface electromagnetic echo data to be detected.

Optionally, the performing median filtering on the water surface electromagnetic echo data to be detected includes:

The pixel value of the pixel overlapping with the center of the two-dimensional template in the image formed by the acquired electromagnetic echo data of the water surface to be detected is set as the median value of the gray value of each pixel in the coverage area of the two-dimensional template.

Optionally, the wavelet threshold denoising includes:

Perform wavelet transform on the image after median filtering to obtain corresponding scale coefficients and wavelet coefficients;

Filter out noise-dominated wavelet coefficients based on a given threshold;

Based on the remaining wavelet coefficients, wavelet reconstruction is performed to obtain the electromagnetic echo data of the water surface to be detected after denoising.

The present application also provides a preprocessed sea surface target detection device, and the preprocessed sea surface target detection device includes:

a water surface electromagnetic echo data acquisition module to be detected, the water surface electromagnetic echo data acquisition module to be detected is used to acquire the water surface electromagnetic echo data to be detected;

a preprocessing module, which is used for removing baseline drift and filtering the electromagnetic echo data of the water surface to be detected, so as to obtain preprocessed electromagnetic echo data of the water surface to be detected;

an echo feature extraction module, which is used to extract the echo features of the preprocessed water surface electromagnetic echo data to be detected;

An input module, which is used for inputting the echo feature into the trained neural network detection model to determine whether the electromagnetic echo data on the water surface to be detected is sea clutter data.

The present application also provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the above-mentioned preprocessed sea surface target detection method can be implemented.

The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the above-mentioned computer program when the processor executes the computer program Preprocessed sea surface object detection method.

Beneficial effects:

The invention is based on a water surface electromagnetic target detection method combining three features and deep learning, adopts a composite denoising method and processes the data, further improves the accuracy of target detection, and can effectively reduce the false alarm rate of sea clutter detection. At the same time, the detection speed can be accelerated. At the same time, the invention combines the deep learning neural network to further improve the detection speed of the water surface electromagnetic target and improve the detection accuracy.

Description of drawings

FIG. 1 is a flowchart of a preprocessed sea surface target detection method according to an embodiment of the present application.

FIG. 2 is an electronic device for implementing the preprocessed sea surface target detection method shown in FIG. 1 .

Detailed ways

In order to make the implementation purpose, technical solutions and advantages of the present application clearer, the technical solutions in the embodiments of the present application will be described in more detail below with reference to the accompanying drawings in the embodiments of the present application. Throughout the drawings, the same or similar reference numbers refer to the same or similar elements or elements having the same or similar functions. The described embodiments are some, but not all, of the embodiments of the present application. The embodiments described below with reference to the accompanying drawings are exemplary, and are intended to be used to explain the present application, but should not be construed as a limitation to the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the protection scope of the present application. The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

In the description of this application, it should be understood that the terms "center", "portrait", "horizontal", "front", "rear", "left", "right", "vertical", "horizontal", The orientation or positional relationship indicated by "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that The device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore should not be construed as limiting the scope of protection of the present application.

FIG. 1 is a schematic flowchart of a preprocessed sea surface target detection method according to the first embodiment of the present invention.

The preprocessed sea surface target detection method shown in Figure 1 includes:

Obtain the electromagnetic echo data of the water surface to be detected;

Remove baseline drift and filter the electromagnetic echo data of the water surface to be detected, so as to obtain the pre-processed electromagnetic echo data of the water surface to be detected;

extracting the echo features of the preprocessed water surface electromagnetic echo data to be detected;

The echo features are input into the trained neural network detection model to determine whether the electromagnetic echo data on the water surface to be detected is sea clutter data.

Optionally, the filtering processing of the electromagnetic echo data on the water surface to be detected includes:

Median filtering and wavelet threshold denoising are respectively performed on the water surface electromagnetic echo data to be detected.

Optionally, the performing median filtering on the water surface electromagnetic echo data to be detected includes:

The pixel value of the pixel overlapping with the center of the two-dimensional template in the image formed by the acquired electromagnetic echo data of the water surface to be detected is set as the median value of the gray value of each pixel in the coverage area of the two-dimensional template.

Optionally, the wavelet threshold denoising includes:

Perform wavelet transform on the image after median filtering to obtain corresponding scale coefficients and wavelet coefficients;

Filter out noise-dominated wavelet coefficients based on a given threshold;

Based on the remaining wavelet coefficients, wavelet reconstruction is performed to obtain the electromagnetic echo data of the water surface to be detected after denoising.

The present invention is based on a water surface electromagnetic target detection method combining three features and deep learning, and the specific implementation steps are as follows:

First obtain the water surface electromagnetic echo data, including:

The selection of the surface electromagnetic target sea area needs to select the sea area with richer types of surface targets and a wide field of vision;

The installation location of the shore-based scanning radar is determined. Since the radar adopts the solid-state power amplifier combined pulse emission system, the emission time is 40ns to 100μs, and the 360° omnidirectional scanning in the horizontal plane is selected. Installation Location;

Measure and record the echo signal of the measured sea surface, establish the corresponding relationship between the radar scattering coefficient of the measured sea surface and the measured value of the radar video voltage through calibration, measure the receiver output power P _r0 of the calibration body, and re-measure in the same state The receiver output power P _r out of the measured sea surface, then the scattering coefficient σ ⁰ of the measured sea surface is:

Among them, σ0 unit: dBm2/m2; σ represents the radar scattering cross-sectional area of the radar antenna beam illuminating the sea surface, unit: m2; A represents the area of the radar antenna beam illuminating the sea surface, unit: m2; Pr The echo power of the measured sea surface, unit :W; Pr0 means the echo power of the calibration body, unit: W; Rr means the distance from the measured sea surface to the antenna, unit: m; R0 means the distance from the calibration body to the antenna, unit: m; σ0 means the calibration body The radar scattering cross-sectional area, unit: m2; Vr represents the echo voltage of the measured sea surface, unit: V; Vr0 represents the echo voltage of the calibration body, unit: V.

For surface electromagnetic target detection, it can be expressed as the following binary hypothesis test:

Among them, the null hypothesis H0 indicates that there is no target in the test cell, and the alternative hypothesis H1 indicates that there is a target in the test cell, and the noise can be ignored in the case of high clutter-to-noise ratio. x(n), s(n) and c(n) are the vectors received in the test cell, the possible target vectors in the test cell and the sea clutter vector in the test cell, respectively. x _p (n) = c _p (n), p = 1, 2, ..., p is the sea clutter vector at the reference cell.

Features are then extracted from sea clutter data and sequences containing target echo data, including:

The relative mean amplitude of the received vector, the mean amplitude of a time series of length N is expressed as:

和

分别是在测试单元和参考单元处接收到的向量x的平均幅度，则接收向量的相对平均幅度(RRA)表示为：Assume

and

are the average magnitudes of the vectors x received at the test unit and the reference unit, respectively, then the relative average magnitude (RRA) of the received vectors is expressed as:

For non-uniform sea clutter along distance cells, the denominator of RAA fits well the clutter level as a function of distance. RAA is a test statistic for object detection with constant false alarm rate (CFAR) property.

The time-domain Husrt exponent for surface electromagnetic data uses the concepts of fractal theory to define features in the time domain. Let {x(N), n = 1, 2, ..., N} be a time series consisting of echo amplitudes, which can be modeled by the following fractal process:

where F(·) is the fluctuation function, H(2) is the time domain Hurst exponent (Td Hurst), and m is the time interval. In order to describe the Husrt exponent in the time domain more intuitively, take the logarithm operation on both sides of the formula, which is given by the following formula:

where log ₂ F(m) is linearly related to log ₂ (m) in the logarithmic domain. From this realization, the TD Hurst feature can be easily obtained by a first-order least squares polynomial approximation on the logarithmic scale.

由于频谱幅值的平方与功率成正相关，功率谱密度(PSD)由下式给出The frequency-domain Husrt exponent of surface electromagnetic data, in order to enhance the distinguishability of the feature space, introduces fractal theory into the frequency domain, in order to extract a feature that can supplement the feature space with additional spectral information. By Fourier transform of the received signal, we get

Since the square of the spectral magnitude is positively related to the power, the power spectral density (PSD) is given by

If x(N) is a fractal process, its power spectral density (PSD) S(F) also satisfies the fluctuation analysis, and the above formula is used to extract the Hurst exponent in the frequency domain, which is expressed as PSD Hurst.

The relative Doppler peak height of the Doppler amplitude spectrum. The sea surface illuminated by the radar contains a large number of scatterers with different radial velocities, and the radial velocity of the target varies in a small range during the observation time in seconds. Therefore, the energy of the target echo is more concentrated in the Doppler domain than the energy of the sea clutter. Let x(n) be a time series of length N received at the test unit. The Doppler amplitude spectrum is calculated by:

where fd is the Doppler frequency, the location and height of the Doppler peak of the Doppler amplitude spectrum are estimated as:

Let 2δ1 and 2δ2 be the width of the reference Doppler interval and the maximum possible width of the target Doppler peak, the relative peak height (RPH) of the Doppler amplitude spectrum is defined as:

Δ=[-δ ₁ , -δ ₂ ]∪[δ ₂ ,δ ₁ ]

where the symbol #Δ denotes the number of Doppler bins that fall into the set Δ. The denominator of the above equation is the average magnitude of the Doppler magnitude spectrum at the reference Doppler interval near the Doppler peak.

The vector entropy of the relative Doppler amplitude spectrum of the Doppler amplitude spectrum includes:

The Doppler amplitude spectrum of the sea clutter shows that the effective values of the Doppler amplitude spectrum of the echo with the target are concentrated on a small number of Doppler bins. The vector entropy (VE) of the Doppler amplitude spectrum is a useful feature to distinguish target echoes from sea clutter and is expressed by:

是归一化多普勒幅度谱，并被视为概率密度函数。in

is the normalized Doppler magnitude spectrum and is treated as a probability density function.

For the VE of a reference range cell, the reference relative vector entropy (RVE) is defined as:

For returns with targets, RVE takes a small value, and for clutter-only vectors, RVE takes a large value.

Finally, a neural network detection model is built to train the sea clutter data, including:

To build the FasterR-CNN convolutional neural network model, the steps include:

A: Input the labeled training images;

B: Input the image to the feature extraction network for feature extraction;

C: Use the RPN structure to generate candidate frames for target object detection;

D: Map the candidate frame to the feature map of the last layer of the convolutional network;

E: The Rol pooling layer makes each RoI generate a fixed-dimensional feature vector;

F: Use detection classification probability and detection classification regression to correct the position of the predicted frame so that the predicted frame position is closer to the true value.

FasterR-CNN is divided into three modules: the first module uses the public network to extract the features of the image, the second module uses the deep convolutional neural network to generate the candidate window of the target object, and the third module uses the generated The candidate window is used as the FasterR-CNN target detector. The structure of FastR-CNN is designed as follows:

A: Convolutional layer 1: 32 convolution kernels of size 3×3;

B: Convolutional layer 2: 64 convolution kernels of 3×3 size;

C: Pooling layer: the maximum pooling method is adopted, the pooling radius is 2, the pooling range is 2×2, and the activation function is Softmax;

D: Convolutional layer 3: 128 convolution kernels of 3×3 size, and the activation function is the Relu function;

E: RoIPooling layer: used to adjust feature maps of different sizes to the same size;

F: Two fully connected layers in parallel realize size adjustment and classification recognition respectively.

The core RPN (region proposal network) of FasterR-CNN is constructed, which consists of an intermediate convolutional layer and a classification layer and a regression layer after the convolutional layer. The essence of RPN is a classless object detector based on sliding window, its input is the feature map extracted by the convolution layer, and the size of the feature map depends on the structure of the convolution layer. The classification layer and the regression layer are based on the feature vector obtained after convolution and the size and coordinate information of the region extracted by the An-chor mechanism. The structure of RPN is designed as follows:

A: Convolutional layer 1: 32 convolution kernels of size 3×3;

B: Convolutional layer 2: 64 convolution kernels of 3×3 size;

C: Pooling layer: the maximum pooling method is adopted, the pooling radius is 2, the pooling range is 2×2, and the activation function is Softmax;

D: Convolutional layer 3: 128 convolution kernels of 3×3 size, and the activation function is the Relu function;

E: RPN convolution layer: 128 3×3 size convolution kernels;

F: At the end of the structure, the classification and regression functions are implemented by two convolutional layers respectively, and the structure is 8 1×1 convolution kernels and 16 1×1 convolution kernels.

The loss function of training RPN is composed of two parts: the classification decision loss function and the size adjustment loss function:

Among them: L _cls is the loss function of classification judgment, and L _reg is the loss function of size adjustment.

The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the above-mentioned computer program when the processor executes the computer program Preprocessed sea surface object detection method.

The present application also provides a computer-readable storage medium, where a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, the above-mentioned preprocessed sea surface target detection method can be implemented.

FIG. 2 is an exemplary structural diagram of an electronic device capable of implementing the preprocessed sea surface target detection method provided according to an embodiment of the present application.

As shown in FIG. 2 , the electronic device includes an input device 501 , an input interface 502 , a central processing unit 503 , a memory 504 , an output interface 505 and an output device 506 . Among them, the input interface 502, the central processing unit 503, the memory 504 and the output interface 505 are connected to each other through the bus 507, and the input device 501 and the output device 506 are respectively connected to the bus 507 through the input interface 502 and the output interface 505, and then to other electronic devices. Component connection. Specifically, the input device 504 receives input information from the outside, and transmits the input information to the central processing unit 503 through the input interface 502; the central processing unit 503 processes the input information based on the computer-executable instructions stored in the memory 504 to generate output information, temporarily or permanently store the output information in the memory 504, and then transmit the output information to the output device 506 through the output interface 505; the output device 506 outputs the output information to the outside of the electronic device for the user to use.

That is, the electronic device shown in FIG. 2 can also be implemented to include: a memory storing computer-executable instructions; and one or more processors that, when executing the computer-executable instructions, can Implement the preprocessed sea surface object detection method described in conjunction with Figure 1.

In one embodiment, the electronic device shown in FIG. 2 may be implemented to include: a memory 504 configured to store executable program codes; one or more processors 503 configured to execute executable programs stored in the memory 504 The program code is used to execute the preprocessed sea surface target detection method in the above embodiment.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

Memory may include non-persistent memory in computer readable media, random access memory (RAM) and/or non-volatile memory in the form of, for example, read only memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media includes both permanent and non-permanent, removable and non-removable, and the media can be implemented by any method or technology for storage of information. Information may be computer readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Flash Memory or other memory technology, Compact Disc Read Only Memory (CD-ROM), Data Versatile Disc (DVD) or other optical storage, Magnetic tape cartridges, tape-disk storage, or other magnetic storage devices or any other non-transmission medium that can be used to store information that can be accessed by a computing device.

It will be appreciated by those skilled in the art that the embodiments of the present application may be provided as a method, a system or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.

Furthermore, it is clear that the word "comprising" does not exclude other elements or steps. Several units, modules or means recited in the device claims can also be realized by one unit or total means by means of software or hardware. The terms first, second, etc. are used to identify the names, not any particular order.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which includes one or more functions for implementing the specified logical function(s) executable instructions. It should also be noted that, in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks identified in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowchart illustrations, and combinations of blocks in the block diagrams and/or general flowchart illustrations, can be implemented by dedicated hardware-based systems that perform the specified functions or operations. implementation, or may be implemented in a combination of special purpose hardware and computer instructions.

The processor in this embodiment may be a central processing unit (Central Processing Unit, CPU), and may also be other general-purpose processors, digital signal processors (Digital Signal Processors, DSP), application specific integrated circuits (Application Specific Integrated Circuits) , ASIC), off-the-shelf programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The memory can be used to store computer programs and/or modules, and the processor implements various functions of the apparatus/terminal device by running or executing the computer programs and/or modules stored in the memory and calling data stored in the memory. The memory can mainly include a stored program area and a stored data area, wherein the stored program area can store an operating system, an application program (such as a sound playback function, an image playback function, etc.) required for at least one function; data (such as audio data, phone book, etc.) In addition, the memory may include high-speed random access memory, and may also include non-volatile memory such as hard disk, internal memory, plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card , a flash memory card (Flash Card), at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device.

In this embodiment, if the modules/units integrated in the device/terminal device are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the processes in the methods of the above embodiments, and can also be completed by instructing relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program is in When executed by the processor, the steps of the foregoing method embodiments can be implemented. Wherein, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file or some intermediate forms, and the like. The computer readable medium may include: any entity or device capable of carrying computer program code, recording medium, U disk, removable hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access Memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction.

Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the present invention, but not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements to some of the technical features; and these Modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (7)

1. a pre-processed sea surface target detection method is characterized in that, the described pre-processed sea surface target detection method comprises: Obtain the electromagnetic echo data of the water surface to be detected; Remove baseline drift and filter the electromagnetic echo data of the water surface to be detected, so as to obtain the pre-processed electromagnetic echo data of the water surface to be detected; extracting the echo features of the preprocessed water surface electromagnetic echo data to be detected; The echo features are input into the trained neural network detection model to determine whether the electromagnetic echo data on the water surface to be detected is sea clutter data. 2. The preprocessed sea surface target detection method according to claim 1, wherein the filtering processing of the water surface electromagnetic echo data to be detected comprises: Median filtering and wavelet threshold denoising are respectively performed on the water surface electromagnetic echo data to be detected. 3. The preprocessed sea surface target detection method according to claim 2, wherein the performing median filtering on the water surface electromagnetic echo data to be detected comprises: The pixel value of the pixel overlapping with the center of the two-dimensional template in the image formed by the acquired electromagnetic echo data of the water surface to be detected is set as the median value of the gray value of each pixel in the coverage area of the two-dimensional template. 4. The preprocessed sea surface target detection method according to claim 3, wherein the wavelet threshold denoising comprises: Perform wavelet transform on the image after median filtering to obtain corresponding scale coefficients and wavelet coefficients; Filter out noise-dominated wavelet coefficients based on a given threshold; Based on the remaining wavelet coefficients, wavelet reconstruction is performed to obtain the electromagnetic echo data of the water surface to be detected after denoising. 5. A preprocessed sea surface target detection device, characterized in that the preprocessed sea surface target detection device comprises: a water surface electromagnetic echo data acquisition module to be detected, the water surface electromagnetic echo data acquisition module to be detected is used to acquire the water surface electromagnetic echo data to be detected; a preprocessing module, which is used for removing baseline drift and filtering the electromagnetic echo data of the water surface to be detected, so as to obtain preprocessed electromagnetic echo data of the water surface to be detected; an echo feature extraction module, which is used to extract the echo features of the preprocessed water surface electromagnetic echo data to be detected; An input module, which is used for inputting the echo feature into the trained neural network detection model to determine whether the electromagnetic echo data on the water surface to be detected is sea clutter data. 6. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, wherein, when the computer program is executed by a processor, the computer program according to any one of claims 1 to 4 can be implemented Preprocessed sea surface object detection method. 7. An electronic device, characterized in that it comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, the processor implementing the computer program as claimed in the claims The preprocessed sea surface target detection method described in any one of 1 to 4.

Images