Method for detecting offshore wind driven generator by maritime radar

Abstract

The invention discloses a method for detecting an offshore wind driven generator by a maritime radar, which comprises the following steps of: 1, starting the maritime radar and preheating the maritime radar, and starting scanning by an antenna; 2, during antenna scanning, recording N pulse echo powers of a given distance gate in a certain direction in a coherence interval; 3, periodically detecting and analyzing non-stationary signals of the time sequence; 4, obtaining a periodic index through the periodic detection analysis in the step 3; 5, the periodic indexes are judged; 6, when the periodic index is larger than or equal to a preset threshold value, the target in the specific distance door is judged to be the offshore wind turbine; according to the method, the power time sequence data is acquired through the single and multiple scanning speed modes of the antenna, the characteristic indexes are comprehensively extracted in combination with multiple periodic detection algorithms, the wind driven generator target can be accurately identified, and the method has the characteristics of improving the detection accuracy, reducing the radar false alarm rate, guaranteeing the marine navigation safety and being wide in applicability.

Description

Technical Field

[0001] The present invention relates to the field of offshore wind turbine detection, and in particular to a method for detecting offshore wind turbines using a maritime radar. Background Art

[0002] Currently, maritime radars primarily consist of shipborne navigation radars and shore-based surveillance radars. Due to cost and practical considerations, the former often utilizes non-coherent radars with magnetron transmission mechanisms, which rely solely on the power signal of the target's echo to detect the target. The latter, based on practical needs, includes both non-coherent radars with magnetron transmission mechanisms and coherent radars based on solid-state transmission technology. Because coherent radars can provide additional Doppler information about the target, they offer superior performance in target detection, identification, and tracking. With the rapid development of the global offshore wind power industry, the number of offshore wind turbines is increasing. The complex electromagnetic scattering characteristics caused by the large size and rotating blades of these offshore wind turbines can interfere with the normal operation of maritime radars, potentially leading to misjudgments or missed detection of targets such as ships, impacting maritime traffic safety. Furthermore, radar-based detection of offshore wind turbines is crucial for ensuring the safe operation and maintenance of offshore wind farms. Traditional maritime radar target detection methods often struggle to accurately identify offshore wind turbines amidst the complex background noise of the marine environment. Summary of the Invention

[0003] The purpose of the present invention is to solve the problem that the complex electromagnetic characteristics caused by the large geometric size and blade rotation of the wind turbine mentioned above will interfere with the normal operation of the maritime radar and cause misjudgment. Therefore, a method for detecting offshore wind turbines by maritime radar is proposed.

[0004] The object of the present invention can be achieved by the following technical solution: A method for detecting offshore wind turbines using a maritime radar, comprising:

[0005] S01: Turn on the maritime radar and preheat it. The maritime radar operates in the X-band and performs a 360-degree scan of the surrounding environment through the mechanical scanning method of the antenna;

[0006] S02: The speed of the maritime radar antenna is adjustable, and K (K≥1) speeds are set; when the antenna is at the i (1≤i≤K) At a certain scanning speed, for a given range gate at a certain azimuth, the N pulse echo powers within a coherent interval are recorded to form a time series S with a length of N. iN ;

[0007] S03: For the above i-th time series S iNPerform periodic detection and analysis of non-stationary signals, the detection and analysis adopts a time-frequency method based on wavelet analysis, an adaptive method based on empirical mode decomposition, and an entropy value analysis method based on samples; or, for the K power sequences S under the above K antenna speeds, iN (i=1, 2, ..., K) respectively perform periodic analysis of non-stationary signals, wherein the periodic detection analysis adopts a time-frequency analysis method based on wavelet analysis or an adaptive decomposition method based on empirical mode decomposition or a sample entropy analysis method;

[0008] S04: Through the periodic detection analysis of step S03, from the i-th time series S iN Obtain three periodic indicators including peak signal-to-noise ratio PSNR, periodic strength PS, and sample entropy SE, or obtain K power sequences S under the above K antenna speeds iN At least one periodicity indicator including peak signal-to-noise ratio (PSNR), periodicity strength (PS), and sample entropy (SE) in (i=1, 2, ..., K);

[0009] S05: For the i-th time series S iN The three periodic indicators are jointly determined; or the K power time series S obtained under the K types of antenna speeds are iN (i=1,2,...,K) and the same periodicity index is obtained for judgment;

[0010] S06: When the periodicity index is greater than or equal to a preset threshold, the target in the specific distance gate is determined to be an offshore wind turbine; otherwise, it is not determined to be an offshore wind turbine.

[0011] Furthermore, the operating frequency of the maritime radar is approximately 9.4 GHz.

[0012] Furthermore, the mechanical scanning speed of the antenna is in the range of 20-40 rpm.

[0013] Furthermore, in the step S02, the larger the value of N is, the greater the time series S is. iN The more obvious the periodic characteristics are.

[0014] Furthermore, in step S02, the antenna preferentially selects the lowest scanning speed to obtain multiple pulse echoes within the coherent interval, and constructs a longer time series to facilitate the extraction of periodic features.

[0015] Furthermore, in the time-frequency analysis method based on wavelet analysis in step S03, discrete wavelet transform is used to perform time-frequency analysis on the i-th or all sequences of the K power echo time series obtained under the K types of antenna speeds to obtain the periodic characteristics of the signal peak signal-to-noise ratio PSNR.

[0016] Furthermore, the adaptive decomposition method based on empirical mode decomposition in step S03 is used to perform modal decomposition and IMF analysis on the i-th or all sequences of the K power echo time series obtained under the K types of antenna speeds to obtain the periodic strength PS characteristics of the signal.

[0017] Furthermore, in the sample entropy analysis method in step S03, the sample entropy value of the i-th or all sequences in the K power echo time series obtained under the K types of antennas is obtained to obtain the sample entropy value SE feature reflecting the strength of the signal periodicity.

[0018] Compared with the prior art, the present invention has the following beneficial effects:

[0019] Power time series data is acquired through single and multiple antenna scanning speed modes, characteristic indicators are extracted by combining multiple periodic detection algorithms, and the threshold judgment mechanism is used to accurately identify wind turbine targets. This has the advantages of improving detection accuracy, reducing radar false alarm rate, and ensuring maritime navigation safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0021] Figure 1 This is a schematic diagram of a method for detecting offshore wind turbines using a maritime radar according to the present invention. DETAILED DESCRIPTION

[0022] The technical solutions of the present invention will be described clearly and completely below with reference to the embodiments. It is obvious that the embodiments described are only some of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without creative effort are intended to fall within the scope of protection of the present invention.

[0023] In existing technologies, research on radar echo characteristics of offshore wind turbines lags behind actual construction needs. Traditional maritime radars mainly rely on target echo power detection. However, due to the large geometric size, complex structure and rotating blade characteristics of wind turbines, the echo signal characteristics are easily complex, which can easily lead to misjudgment or missed detection of targets. Existing methods have not established an effective analysis mechanism for the periodic characteristics of such targets, making it difficult to achieve accurate identification in dense sea environments. For example, ship radars near a wind farm in a certain sea area often misjudge wind turbines as moving ships or sea clutter, resulting in frequent warnings from the collision avoidance system or collision risks.

[0024] To address these issues, the reliable extraction of periodic features from non-stationary signals is crucial. Research has found that rotating wind turbine blades produce periodic features in radar echoes within specific frequency bands. However, due to interference from sea clutter and ship motion, the signal characteristics obtained in short-duration, single-antenna scanning modes are not stable enough. By analyzing time series of single echo power accumulated over a certain period of time, or multiple echo power series at different antenna scanning speeds, it was found that both single scans of a certain duration and multiple scans at different antenna scanning speeds can enhance the stability and information content of the echo signal, making it easier to highlight the periodic characteristics of the radar echo signal. Further verification revealed that combining multiple signal analysis methods can effectively suppress noise interference and effectively extract the signal's periodic features. This results in a technical approach for acquiring time series using single and multiple antenna scanning speed modes and jointly analyzing periodic indicators using multiple algorithms. It is desirable for the antenna to have at least two adjustable mechanical scanning speeds, a basic feature of existing maritime radars, but this method also takes into account the applicability of antennas with only a single rotational speed.

[0025] Therefore, see Figure 1 As shown, the present application proposes a method for detecting offshore wind turbines using a maritime radar, comprising: after the radar is turned on and preheated, scanning the same azimuth range gate at a single or multiple mechanical scanning speeds, and recording the echo power sequence with a certain duration at a single scanning speed, or the echo power sequence obtained at multiple antenna scanning speeds. For the echo power sequence with a certain duration obtained at a single antenna scanning speed, three methods, namely wavelet analysis, empirical mode decomposition, and sample entropy, are used to analyze the periodicity of the detection sequence, extract the peak signal-to-noise ratio, periodicity intensity, and sample entropy index, and comprehensively combine the three periodicity indexes to determine the target attributes through threshold setting; for multiple echo power sequences obtained at multiple antenna scanning speeds, wavelet analysis, empirical mode analysis, or sample entropy analysis are used to analyze the periodicity of the detection sequence, extract at least one of the peak signal-to-noise ratio, periodicity intensity, or sample entropy index of the signal, and determine the target attributes through threshold setting.

[0026] Adjusting the antenna's mechanical scanning speed involves varying the number of power echo acquisitions by changing the antenna's rotation rate, enabling better analysis and determination of target attributes. For example, using two different speeds within the 20-40 rpm range can enhance the periodicity of power echoes acquired at different antenna speeds. In other words, the different periodic characteristics of two sets of echoes acquired at two different antenna speeds further enhance quantitative determination of target attributes. Time series construction involves acquiring the echo power of at least N = 32 transmitted pulses for a given azimuth and range gate during a single scan at a fixed antenna speed. This means that a coherent interval for that azimuth and range scan should contain at least 32 pulses. For example, using N = 64 allows for the acquisition of a longer echo power sequence for that azimuth and range gate, making the echo sequence's periodicity more pronounced, facilitating target attribute determination. In the periodic feature detection method, wavelet analysis extracts different frequency components of the signal through multi-scale decomposition, obtains the signal's periodic peak signal-to-noise ratio (PSNR) to evaluate the signal's periodicity strength, empirical mode decomposition obtains the signal's intrinsic mode function through adaptive decomposition to evaluate the signal's periodicity strength, and sample entropy analysis evaluates the signal's periodicity strength through sequence complexity. All three can be used independently or in combination for feature extraction, depending on the specific circumstances described below.

[0027] Specifically, the number of antenna speeds selected determines the number of echo sequences; when the antenna only uses a single speed scan to obtain power echoes, or when analyzing the echo sequence obtained from a certain speed under multiple speed scans, the number of echo sequences is 1; in order to enhance the accuracy of target attribute determination, the three indicators corresponding to the three periodic analysis methods mentioned above are used for comprehensive judgment, and when the indicator exceeds the preset threshold, it is determined to be a wind turbine. When the selected antenna speed is two or more, there are two or more echo power sequences, and the sequences are analyzed for periodicity separately, for example, one of the three periodic analysis methods mentioned above is used to obtain the corresponding sequence periodicity indicator; for example, discrete wavelet transform is used to decompose components of different frequency bands, and the energy distribution of each component is calculated to determine the main period; or multiple IMF components are obtained through empirical mode decomposition, and components with significant periodicity are screened to calculate the energy ratio; or the sample entropy value can be calculated to evaluate the regularity of the sequence, and the lower the entropy value, the stronger the periodicity; the same indicators under two or more scanning speeds are compared, and when the indicator exceeds the preset threshold, it is determined to be a wind turbine;

[0028] Compared with existing technologies, traditional methods rely solely on the echo power after pulse accumulation to achieve target detection, and are unable to distinguish between marine targets such as wind turbines and ships. The present method utilizes the periodic characteristics of the radar power echo of the rotating wind turbine and the adjustable rotation speed of the radar antenna to enhance the significance of the periodic characteristics. Combined with the independent and cross-validation mechanisms of multiple algorithms, it effectively overcomes misjudgments caused by interference such as sea clutter. For example, an experiment showed that the traditional method had an error rate of up to 42% in level 4 sea conditions. Under two selected antenna rotation speeds, the present method can reduce the error rate to 10% by determining the periodic peak power characteristics through the wavelet transform method.

[0029] Through the above technical solution, this application realizes the reliable extraction of the unique periodic characteristics of offshore wind turbine echoes, and solves the problem of target misidentification in complex sea conditions. This method is compatible with both incoherent and coherent radar systems, and improves detection robustness through multi-dimensional signal analysis, providing effective technical support for ship collision avoidance systems and wind farm supervision.

[0030] The present application further proposes that the method is applicable to a maritime radar operating frequency of approximately 9.4 GHz;

[0031] The operating frequency refers to the oscillation frequency of the electromagnetic waves emitted by the radar when propagating in free space. Specifically, a solid-state transmitter or a magnetron transmitter can be used to generate the electromagnetic wave signal. The frequency is set within the X-band range. The X-band refers to the microwave frequency band between 8-12GHz. This frequency band has high spatial resolution and moderate atmospheric attenuation characteristics, and is suitable for accurate detection of sea surface targets at medium and short distances.

[0032] Specifically, the 9.4 GHz operating frequency was selected as the core parameter of maritime radar. Its wavelength is approximately 3.2 cm, which can effectively capture the periodic echo modulation caused by the rotation of offshore wind turbine blades. Since wind turbine blades are typically on the order of tens of meters in size, the interaction between their rotation period and radar waves will form a relatively significant periodic feature in the echo power sequence.

[0033] Compared with existing technologies, some other maritime radars typically operate in the S-band or C-band. For example, some shore-based radars use frequencies below 3 GHz to improve long-range detection capabilities. However, these lower-frequency bands have longer wavelengths, are less sensitive to subtle echo changes caused by the rotation of wind turbine blades, and are susceptible to interference from sea surface multipath. The 9.4 GHz X-band frequency strikes a balance between resolution and anti-interference capabilities, making it more suitable for capturing the unique periodic echo characteristics of wind turbines.

[0034] Through the above technical solution, the present application can improve the radar's detection sensitivity to periodic echoes from offshore wind turbines and reduce the probability of misjudgment due to environmental noise or target motion complexity. The frequency selection is also compatible with the hardware configuration of existing shipborne navigation radars, achieving target recognition without additional equipment costs, while adapting to detection needs under different sea conditions.

[0035] The present application further proposes that in step S02, the larger the value of N is, the longer the time series S is. iN The more obvious the periodicity of (i=1,2,...,K);

[0036] N refers to the number of pulse echoes recorded within the coherence interval, which can be achieved by increasing the radar scan time or adjusting the pulse repetition frequency. This can enhance the ability to extract periodic features by accumulating more data points. A predetermined range gate refers to a fixed distance unit within the radar detection range. This can be achieved through the range gate gating technology in radar signal processing, which is used to focus the analysis of echo signals at a specific location.

[0037] Specifically, when the radar scans the same azimuth and a given range gate, by extending the number of pulse echo records within the coherent interval, for example increasing N from 32 to 64 or higher, random noise interference can be effectively suppressed while enhancing the cumulative effect of periodic signals. Since the rotation of wind turbine blades has a stable periodic characteristic, when the N value increases, the energy proportion of the periodic component in the time series will increase significantly, making it easier to detect in subsequent time-frequency analysis or sample entropy calculation;

[0038] Compared with existing technologies, traditional methods usually use short time series of fixed length for periodic analysis, which is prone to misjudgment or missed detection due to insufficient data. However, this method can adapt to the detection needs of different sea conditions and wind turbine sizes by dynamically adjusting the value of N. Especially in complex electromagnetic environments, increasing the value of N can significantly improve the signal-to-noise ratio of periodic indicators.

[0039] Through the above technical solution, the present application solves the problem of unstable periodic feature extraction in short time series. By optimizing the data acquisition length, the ability to resolve periodic echoes of wind turbines is improved, thereby reducing the probability of misjudgment due to noise interference or insufficient data, and enhancing the reliability of radar detection results.

[0040] This application further proposes that the antenna preferentially selects the lowest scanning speed, at a fixed pulse repetition frequency, to obtain pulse echoes within multiple coherent intervals and construct a longer time series;

[0041] The lowest scanning speed refers to the lower limit of the antenna's mechanical rotation speed range, which can be achieved using a typical rotation speed of 20 rpm. A lower scanning speed allows more pulse echo acquisition windows to be covered within the same time, thereby increasing the time span of data acquisition. Constructing a longer time series refers to extending the number of pulse echo records within a single coherent interval. This can be achieved by continuously acquiring N pulse echo powers within multiple coherent intervals and splicing them together. Longer data sequences can improve the resolution capability of the periodicity detection algorithm.

[0042] Specifically, during the detection process, when the radar antenna operates at the lowest scanning speed, the rate of change of its azimuth angle decreases, so that the target in the same azimuth can be illuminated multiple times within the radar beam dwell time. For example, at the first scanning speed, it may take a long time for the antenna to complete a 360° scan. At this time, the pulse echo of a specific range gate is continuously recorded, and more pulse samples can be accumulated in a single coherent interval. By superimposing or splicing the data sequences in multiple coherent intervals, the length of the final time series is significantly increased. This expanded data sequence can more fully reflect the periodic echo modulation characteristics caused by the rotation of the wind turbine blades, thereby providing a more reliable input signal for subsequent time-frequency analysis or sample entropy calculation.

[0043] Compared with existing technologies, traditional radar detection methods typically use a fixed scanning speed or prioritize higher scanning speeds to improve scanning efficiency, but this sacrifices the detection accuracy of periodic signals. This solution actively selects the lowest scanning speed, effectively extending the observation time of target echo signals while ensuring scanning coverage. This technical approach overcomes the difficulty of extracting periodic features caused by the short data sequence in traditional methods and is particularly suitable for scenarios with long rotation periods of wind turbine blades.

[0044] Through the above technical solution, the present application can significantly enhance the radar's ability to capture the periodic echo characteristics of offshore wind turbines, reduce the risk of misjudgment due to insufficient data sequence length, and provide a more stable data basis for subsequent periodic index calculations, thereby improving the accuracy of offshore wind turbine identification.

[0045] This application further proposes a method for detecting offshore wind turbines using maritime radar, which uses a time-frequency analysis method based on wavelet analysis in periodic detection analysis and uses discrete wavelet transform to analyze the time series S iN (i=1,2,...,K,K≥1) Perform analysis at multiple frequency resolutions to obtain time-frequency features at different scales;

[0046] Discrete wavelet transform refers to a mathematical transformation method that decomposes a time series into different frequency subbands. Specifically, it can be implemented using the Mallat algorithm, which decomposes the signal step by step through a low-pass and high-pass filter bank. This transform can effectively separate the high-frequency and low-frequency components in the signal and accurately capture transient characteristics in non-stationary signal analysis.

[0047] Multi-frequency resolution analysis refers to the hierarchical processing of signals through a multi-level decomposition structure. Specifically, it can be implemented using orthogonal wavelet basis functions, such as the Daubechies wavelet system. This method can observe signal characteristics at different scales, making weak periodic components prominent in specific resolution layers.

[0048] The time-frequency feature refers to the joint distribution characteristics of the signal in the time and frequency dimensions. It can be specifically achieved through the detection of the maximum value of the modulus of the wavelet coefficient. This feature can characterize the time-varying law of the periodic components in the signal and provide a quantitative basis for the periodic echo caused by the rotation of the wind turbine blades.

[0049] Specifically, in the radar echo signal processing process, the discrete wavelet transform is first performed on the collected single or multiple time series. By setting the number of decomposition layers to 4-6, the original signal is decomposed into a set of approximate coefficients and detail coefficients. In the decomposition process of each layer, the boundary extension method is used to avoid edge effects. Subsequently, the time-frequency energy of the detail coefficients of each layer is calculated to extract the scale components with significant periodicity. By comparing the stability of the time-frequency characteristics at different scanning speeds, the stable periodic signal generated by the rotation of the wind turbine blades can be effectively distinguished from the random fluctuations of the environmental noise.

[0050] Compared with existing technologies, traditional periodicity detection methods such as Fourier transform can only perform analysis in the global frequency domain and cannot locate the time distribution of periodic components. However, the multi-resolution characteristics of discrete wavelet transform can simultaneously capture the local characteristics of the signal in the time domain and the distribution pattern in the frequency domain. It is particularly suitable for analyzing non-stationary periodic signals caused by the rotation of wind turbine blades. Compared with the short-time Fourier transform with a fixed window length, this method can more accurately match the periodic characteristics of the target signal through adaptive scale selection.

[0051] Through the above technical solution, the present application can effectively extract the periodic echo characteristics generated by the rotation of offshore wind turbine blades and accurately identify weak periodic signals in the complex sea clutter background. This method enhances the anti-noise interference capability through multi-scale analysis, solves the problem of insufficient resolution of traditional methods in non-stationary signal processing, and significantly improves the detection reliability of offshore wind turbines.

[0052] This application further proposes to use an adaptive decomposition method based on empirical mode decomposition in periodic detection analysis to convert the collected single or multiple time series S iN(i=1,2,...,K,K≥1) are decomposed into multiple intrinsic mode functions, and the target characteristics are determined by analyzing the periodic characteristics of the intrinsic mode functions;

[0053] Among them, empirical mode decomposition refers to an adaptive signal decomposition method, which can specifically use the empirical mode decomposition algorithm to decompose non-stationary signals into multiple intrinsic mode functions, which is suitable for processing nonlinear and non-stationary signals. The intrinsic mode function refers to a component signal that satisfies the number of extreme points equal to the number of zero crossing points or differs by at most one, and the local mean is zero. Specifically, it can be iteratively extracted through a screening process and used to characterize oscillation modes of different time scales in the signal. Periodic feature analysis refers to identifying the dominant periodic component by calculating the spectrum or autocorrelation function of the intrinsic mode function. Specifically, the fast Fourier transform or autocorrelation function peak detection method can be used to extract features related to the rotation period of the wind turbine blades from the decomposed components;

[0054] Specifically, in the process of radar echo signal processing, the acquired time series is first subjected to empirical mode decomposition to obtain multiple intrinsic mode functions. Each intrinsic mode function corresponds to an oscillation component of a different frequency, including a periodic component caused by the rotation of the wind turbine blades. By calculating the spectrum or autocorrelation function of each component, the component with significant periodicity can be identified. Since the rotation of the wind turbine blades will produce a stable periodic echo, the corresponding intrinsic mode function will show a periodic feature consistent with the rotation frequency, while the components corresponding to random noise or non-periodic interference signals do not have such features. By screening the components with significant periodicity, the wind turbine target can be effectively distinguished from other interference signals.

[0055] Compared with existing technologies, traditional methods such as Fourier transform or wavelet analysis require preset basis functions when processing non-stationary signals, making it difficult to adaptively match signal characteristics. However, empirical mode decomposition does not require preset basis functions and can achieve adaptive decomposition based on the signal's own characteristics. It is particularly suitable for signals with time-varying characteristics such as wind turbine echoes. In addition, the sample entropy method in existing technologies is not sensitive enough to weak periodic signals. By analyzing the periodic characteristics of the intrinsic mode function, it can effectively extract the periodic components hidden in the noise.

[0056] Through the above technical solution, the present application can separate the periodic components related to wind turbines from complex radar echoes, improve the detection capability of weak periodic signals, thereby accurately identifying offshore wind turbine targets and avoiding misjudgment due to noise interference or non-periodic signals.

[0057] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for detecting offshore wind turbines using a maritime radar, characterized in that: include: S01: Turn on the maritime radar and preheat it. The maritime radar operates in the X-band and performs a 360-degree scan of the surrounding environment through the mechanical scanning method of the antenna; S02: The speed of the maritime radar antenna is adjustable, and K (K≥1) speeds are set; when the antenna is at the i (1≤i≤K) At a certain scanning speed, for a given range gate at a certain azimuth, the N pulse echo powers within a coherent interval are recorded to form a time series S with a length of N. iN ; S03: For the above i-th time series S iN Perform periodic detection and analysis of non-stationary signals, the detection and analysis adopts a time-frequency method based on wavelet analysis, an adaptive method based on empirical mode decomposition, and an entropy value analysis method based on samples; or, for the K power sequences S under the above K antenna speeds, iN (i=1, 2, ..., K) respectively perform periodic analysis of non-stationary signals, wherein the periodic detection analysis adopts a time-frequency analysis method based on wavelet analysis or an adaptive decomposition method based on empirical mode decomposition or a sample entropy analysis method; S04: Through the periodic detection analysis of step S03, from the i-th time series S iN Obtain three periodic indicators including peak signal-to-noise ratio PSNR, periodic strength PS, and sample entropy SE, or obtain K power sequences S under the above K antenna speeds iN At least one periodicity indicator including peak signal-to-noise ratio (PSNR), periodicity strength (PS), and sample entropy (SE) in (i=1, 2, ..., K); S05: For the i-th time series S iN The three periodic indicators are jointly determined; or the K power time series S obtained under the K types of antenna speeds are iN (i=1,2,...,K) and the same periodicity index is obtained for judgment; S06: When the periodicity index is greater than or equal to a preset threshold, the target in the specific distance gate is determined to be an offshore wind turbine; otherwise, it is not determined to be an offshore wind turbine.

2. The method for detecting offshore wind turbines using a maritime radar according to claim 1, wherein: The operating frequency of the maritime radar is 9.4 GHz.

3. The method for detecting offshore wind turbines using a maritime radar according to claim 1, wherein: The mechanical scanning speed of the antenna is in the range of 20-40 rpm.

4. The method for detecting offshore wind turbines using a maritime radar according to claim 1, wherein: In the step S02, the larger the value of N is, the greater the time series S is. iN The more obvious the periodic characteristics are.

5. The method for detecting offshore wind turbines using a maritime radar according to claim 1, wherein: In step S02, the antenna preferentially selects the lowest scanning speed to obtain multiple pulse echoes within the coherent interval and construct a longer time series to facilitate the extraction of periodic features.

6. The method for detecting offshore wind turbines using a maritime radar according to claim 1, characterized in that: The time-frequency analysis method based on wavelet analysis in step S03 uses discrete wavelet transform to perform time-frequency analysis on the i-th or all sequences of the K power echo time series obtained under the K types of antenna speeds to obtain the periodic characteristics of the signal peak signal-to-noise ratio PSNR.

7. The method for detecting offshore wind turbines using a maritime radar according to claim 1, characterized in that: The adaptive decomposition method based on empirical mode decomposition in step S03 is used to perform modal decomposition and IMF analysis on the i-th or all sequences of the K power echo time series obtained under the K types of antenna speeds to obtain the periodic strength PS characteristics of the signal.

8. The method for detecting offshore wind turbines using a maritime radar according to claim 1, wherein: In the sample entropy analysis method in step S03, the sample entropy value of the i-th or all sequences in the K power echo time series obtained under the K types of antennas is obtained to obtain the sample entropy value SE feature reflecting the strength of the signal periodicity.

Images