A sea surface target recognition method using time domain fractional order entropy

Abstract

The application relates to a sea surface target recognition method using a time domain fractional order entropy, and belongs to the field of radar signal processing, and aims to solve the problems of poor adaptability, poor noise resistance and weak target separability of an entropy feature in the prior art. The method comprises the following steps: collecting radar echo complex data and constructing a target amplitude time sequence; segmenting the sequence by a sliding window and dividing a training set and a test set; calculating a fractional order Shannon entropy feature through a fractional order parameter; optimizing a global parameter of the time domain FSE feature based on a Bhattacharyya distance; calculating a feature mean value by using the optimal parameter, constructing a discriminant direction factor and a discriminant quantity, and setting a decision threshold to complete target classification. Through adaptive adjustment of the fractional order parameter and parameter optimization, the feature noise resistance and inter-class separability are improved, the method is suitable for a short-time observation scene, and the recognition accuracy of ships and channel buoys is significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing, specifically relating to a method for identifying sea surface targets using time-domain fractional entropy. Background Technology

[0002] Surface target identification is a core technology for marine surveillance and waterway management. Among various surface targets, waterway buoys are characterized by slow speed and lack of power, making their motion susceptible to wave disturbances and exhibiting a drifting characteristic. Although waterway buoys are smaller than ships, their radar echo characteristics are highly similar to those of ship targets in scenarios with low radar resolution, short-term observation, or ship-side illumination. Complex sea clutter environments are characterized by strong correlation, non-stationarity, and non-Gaussianity. Extracting stable features to distinguish ships from floating targets has become a technical challenge in this field. Currently, information entropy-based signal complexity features are the main application features under conditions such as short-term observation, with traditional Shannon entropy being widely used. However, traditional Shannon entropy has significant limitations: it lacks adaptability and cannot match the statistical characteristics of signals under different sea states and signal-to-noise ratios; in low signal-to-noise ratio environments, random noise exacerbates fluctuations in probability distribution estimation, making it sensitive to disturbances and exhibiting insufficient feature stability, which seriously affects identification performance.

[0003] The invention patent with publication number CN119471609A discloses a method for identifying sea surface targets using temporal approximate entropy features. Its unique feature lies in the following steps: Step 1: Target echo data buffering. Target detection is performed using echo data after pulse compression processing. The target temporal echo sequence is extracted and buffered according to the target's distance cell information. Step 2: Target temporal approximate entropy feature extraction. Echo data temporal analysis involves segmenting the target temporal echo sequence using a sliding window. Data within each segment is reconstructed, and the corresponding temporal series approximate entropy is calculated. The obtained entropy values ​​are then integrated to form a target feature matrix. Step 3: Temporal approximate entropy feature optimization. For the two parameters in the feature extraction process, namely the dimension m of data reconstruction and the tolerance coefficient r in the threshold F, the optimal parameter range is determined based on the degree of influence of different parameters on the separability of the two types of targets, thereby achieving feature optimization and improving recognition performance. Step 4: Target recognition is performed using temporal approximate entropy features. The support vector machine classification algorithm is used to complete the recognition of two types of targets, and the recognition performance is tested. The feature matrix composed of multiple targets is divided into training and test sets, and the hyperparameters of the model are optimized, including the penalty factor c and the RBF kernel function parameter gamma, to obtain better recognition results. Finally, the feature vectors of the test set are input into the model for testing to obtain the target recognition results. This existing technology has the following defects: insufficient feature adaptability, unable to adapt to complex environments; poor noise resistance; insufficient target separability, unable to achieve accurate recognition. These are the shortcomings of the existing technology.

[0004] In view of this, it is very necessary to provide a sea surface target identification method that applies temporal fractional entropy to solve the above-mentioned defects in the prior art. Summary of the Invention

[0005] To address the technical problems of insufficient feature adaptability of Shannon entropy in existing technologies, which makes it unable to adapt to complex environments, poor noise resistance, and insufficient target separability, thus failing to achieve accurate identification, this invention provides a sea surface target identification method using temporal fractional entropy to solve the above-mentioned technical problems.

[0006] In a first aspect, the present invention provides a method for identifying sea surface targets using temporal fractional entropy, comprising:

[0007] Step S1: The step of collecting radar data and preprocessing, collecting radar echo complex data and representing it as a radar echo complex matrix, constructing the target amplitude time series, and providing accurate and target-characteristic basic data support for subsequent feature extraction;

[0008] Radar echo complex data includes ship target data and channel buoy target data;

[0009] The complex matrix of radar echo is ,in, For the number of pulses in the slow time dimension, The number of distance units in the fast time dimension;

[0010] The target amplitude time series includes the target amplitude time series of ships and the target amplitude time series of channel buoys;

[0011] The time series of target amplitudes for channel buoys is represented as follows: ,in, Represents the modulo value. is the distance unit where the channel buoy target is located, and n is the length of the pulse sequence; Indicates the distance threshold of the target buoy in the channel. Above, the amplitude sequence that varies with the pulse signal;

[0012] The time series of ship target amplitude is represented as ;

[0013] in, Let n be the distance unit where the ship target is located, and n be the length of the pulse sequence. The distance from the center of the target ship to the gate. The final center distance gate for coherent processing time;

[0014] Preferably, the center distance gate of the ship target and the distance gate of the channel buoy target are obtained by constant false alarm rate (CFAR) detection.

[0015] The mathematical expression is:

[0016]

[0017] in, For neighborhood windows, Distance between candidate centers This is the distance from the window size. The average energy of each distance gate;

[0018] The mathematical expression is:

[0019]

[0020] in, This is the set of pulse indices for the k-th coherent processing time. This refers to the coherent processing time.

[0021] Step S2: The step of segmenting the target amplitude time series using a sliding window, dividing the target amplitude time series into training and test sets;

[0022] The time series of ship target amplitude and the time series of channel buoy target amplitude are segmented using a sliding window. The i-th segment of the time series of ship target amplitude or the i-th segment of the time series of channel buoy target amplitude is represented as follows:

[0023]

[0024] in, L represents the length of the time series of the target amplitude of a ship or the target amplitude of a channel buoy, s represents the window length of the sliding window, and s represents the step size of the sliding window. , To round down;

[0025] The segmented time series of ship target amplitude and channel buoy target amplitude are arranged in chronological order, with the first half used as the training set and the second half as the test set.

[0026] Step S3: Extracting fractional Shannon entropy features. By using fractional parameters, fractional Shannon entropy is calculated to improve the feature's adaptability to different signal characteristics and its noise resistance stability.

[0027] Histogram statistics are performed on the segmented target amplitude time series in step S2, dividing the target amplitude time series into B histogram bins. The number of target amplitude sequences in the i-th histogram bin is... The probability distribution corresponding to the target amplitude sequence within the i-th histogram bin. The mathematical expression is:

[0028]

[0029]

[0030] Through fractional parameters Calculate the fractional Shannon entropy The mathematical expression is:

[0031]

[0032] The normalized fractional Shannon entropy mathematical expression is:

[0033]

[0034] in, Let B be the probability distribution, and B be the number of histogram bins. The value range is [0,2], and L is the window length of the sliding window.

[0035] When α=1, it is the classic Shannon entropy; when α<1, the fractional Shannon entropy is less sensitive to random noise, and the fractional Shannon entropy feature is more stable under low SNR conditions, while when α>1, it can better emphasize low probability events.

[0036] Step S4: The parameter optimization step involves global parameter optimization of the temporal FSE features based on the Bach distance to obtain the optimal parameters of the temporal FSE features, maximizing the inter-class separability of ships and channel buoy targets, and laying the foundation for accurate identification.

[0037] The mathematical expression for Bach's distance (BD) is:

[0038]

[0039] Wherein, under the fractional-order parameter α, the time-domain FSE feature x of the ship target and the time-domain FSE feature y of the channel buoy both follow a Gaussian distribution, that is, , , Let x and y be the means of a Gaussian distribution, respectively. , Let x and y be the variances of the Gaussian distributions they follow, respectively.

[0040] In the training set, a grid search is performed on the fractional-order parameter α to calculate the Bádov distance (BD) value corresponding to each fractional-order parameter α. The fractional-order parameter corresponding to the maximum Bádov distance (BD) value is selected as the optimal fractional-order parameter. .

[0041] Preferably, a grid search is performed on the fractional-order parameter α, with the search range set to 0.1-2 and the step size set to 0.1.

[0042] Step S5: The step of determining the sample target classification and decision threshold. Based on the optimal parameters in step S4, the mean value of the sample target time-domain FSE features is calculated, the discrimination direction factor and discrimination quantity are constructed, the sample target classification tendency is determined, the decision threshold is set, the sample target classification is determined, the classification logic is unified, and the consistency and reliability of the recognition results are ensured.

[0043] The mathematical expression for the mean of the time-domain FSE features of a ship target is:

[0044]

[0045] in, To train the temporal FSE features of a concentrated set of ship targets;

[0046] The mathematical expression for the mean of the time-domain FSE characteristic of a channel buoy is:

[0047]

[0048] in, To train the temporal FSE characteristics of channel buoys;

[0049] Discrimination Direction Factor The mathematical expression is:

[0050]

[0051] in, The mean of the time-domain FSE features of the ship target. The mean of the time-domain FSE characteristics of the channel buoy;

[0052] The mathematical expression for the discriminant s is as follows:

[0053]

[0054] Among them, the determination of the sample target in the optimal fractional order parameter will be performed. The time-domain FSE feature is denoted as H. The sample targets include ship targets or channel buoy targets;

[0055] When the discriminant s > the preset threshold, the sample target is determined to be a vessel; when the discriminant s < the preset threshold, the sample target is determined to be a waterway buoy.

[0056] The mathematical expression for the empirical false alarm rate is set as follows:

[0057]

[0058] in, The empirical false alarm rate is given by t, where t is the threshold. This represents the number of times a channel buoy sample is misidentified as a vessel when the threshold is t. The number of samples for training concentrated channel buoys;

[0059] The decision threshold is set to the maximum threshold that satisfies the empirical false alarm rate, and the mathematical expression is:

[0060]

[0061] in, The threshold for judgment, To train on the maximum permissible false alarm rate of one channel buoy misidentifying a vessel on average, ;

[0062] When the discriminant s ≥ the decision threshold At that time, the sample target was classified as a ship;

[0063] Discriminant s < decision threshold At that time, the sample target was classified as a channel buoy;

[0064] The mathematical expression for calculating the accuracy of target classification for samples is:

[0065]

[0066] Among them, TP, TN, FP, and FN are four scenarios for target determination in the sample. TP, TN, FP, and FN are respectively: ship target determined as ship, channel buoy target determined as channel buoy, channel buoy target determined as ship, and ship target determined as channel buoy.

[0067] Secondly, the technical solution of the present invention also provides a sea surface target identification using time-domain fractional entropy, including a radar data acquisition and preprocessing module, a sliding window segmentation module for the target amplitude time series, a fractional Shannon entropy feature extraction module, a parameter optimization module, and a target classification and decision threshold module.

[0068] The radar data acquisition and preprocessing module acquires radar echo complex data and represents it as a radar echo complex matrix to construct the target amplitude time series.

[0069] A sliding window segmentation module for the target amplitude time series is used to segment the target amplitude time series into training and test sets.

[0070] Extract the fractional Shannon entropy feature module and calculate the fractional Shannon entropy using fractional parameters;

[0071] The parameter optimization module performs global parameter optimization on the temporal FSE features based on Bach distance to obtain the optimal parameters for the temporal FSE features.

[0072] The module for determining the classification of sample targets and setting a decision threshold calculates the mean of the time-domain FSE features of the sample targets based on the optimal parameters in the parameter optimization module, constructs the discrimination direction factor and discrimination quantity, determines the classification tendency of the sample targets, sets the decision threshold, and determines the classification of the sample targets.

[0073] The beneficial effects of this invention are as follows: This invention provides a sea surface target identification method using fractional-order temporal entropy. By introducing fractional-order parameters and combining them with Bach distance for global optimization, it solves the problem of insufficient adaptability of traditional entropy-based features. Relying on the inherent characteristics of fractional-order Shannon entropy, it significantly reduces sensitivity to random noise, maintaining stable feature representation capabilities even in low signal-to-noise ratio environments, and significantly improving anti-interference performance. The use of sliding window segmentation processing, coupled with targeted target amplitude time series construction, effectively adapts to the sample utilization needs of short-term observation scenarios. By unifying classification logic through the discrimination of direction factors and setting decision thresholds, it significantly improves the identification accuracy of ships and channel buoy targets. It has low computational complexity, and the parameter optimization and threshold setting logic is clear, easy to operate, and convenient for engineering deployment, efficiently meeting practical application needs.

[0074] Furthermore, the design principle of this invention is reliable, the structure is simple, and it has a very wide range of application prospects. Attached Figure Description

[0075] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0076] Figure 1 This is a flowchart of a sea surface target identification method using temporal fractional entropy provided by the present invention.

[0077] Figure 2 This is a block diagram of a sea surface target recognition system that applies time-domain fractional entropy, as provided by the present invention. Detailed Implementation

[0078] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0079] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0080] Example 1:

[0081] like Figure 1 As shown, this embodiment of the invention provides a method for identifying sea surface targets using temporal fractional entropy, comprising the following steps:

[0082] Step S1: The step of acquiring radar data and preprocessing, acquiring radar echo complex data and representing it as a radar echo complex matrix, and constructing the target amplitude time series;

[0083] Radar echo complex data includes ship target data and channel buoy target data;

[0084] The complex matrix of radar echo is ,in, For the number of pulses in the slow time dimension, The number of distance units in the fast time dimension;

[0085] In this embodiment, the pulse repetition frequency (PRF) is 2 kHz.

[0086] The target amplitude time series includes the target amplitude time series of ships and the target amplitude time series of channel buoys;

[0087] The time series of target amplitudes for channel buoys is represented as follows: ,in, Represents the modulo value. is the distance unit where the channel buoy target is located, and n is the length of the pulse sequence; Indicates the distance threshold of the target buoy in the channel. Above, the amplitude sequence that varies with the pulse signal;

[0088] The time series of ship target amplitude is represented as ;

[0089] in, Let n be the distance unit where the ship target is located, and n be the length of the pulse sequence. The distance from the center of the target ship to the gate. The final center distance gate for coherent processing time (CPI);

[0090] The center distance gate of the ship target and the distance gate of the channel buoy target are determined by the Constant False Alarm Rate (CFAR) detection.

[0091] Set the distance gate with the highest average energy as the final center distance gate for this coherent processing time. The mathematical expression is:

[0092]

[0093] in, For neighborhood windows, Distance between candidate centers This is the distance from the window size. The average energy of each distance gate;

[0094] In this embodiment, , , One distance unit.

[0095] The mathematical expression is:

[0096]

[0097] in, This is the set of pulse indices for the k-th coherent processing time. This refers to the coherent processing time.

[0098] In this embodiment, let the pulse index set of the k-th CPI be . ,like .

[0099] Step S2: The step of segmenting the target amplitude time series using a sliding window, dividing the target amplitude time series into training and test sets;

[0100] The time series of ship target amplitude and the time series of channel buoy target amplitude are segmented using a sliding window. The i-th segment of the time series of ship target amplitude or the i-th segment of the time series of channel buoy target amplitude is represented as follows:

[0101]

[0102] in, L represents the length of the time series of the target amplitude of a ship or the target amplitude of a channel buoy, s represents the window length of the sliding window, and s represents the step size of the sliding window. , To round down;

[0103] In this embodiment, L=1024 and S=512.

[0104] The segmented time series of ship target amplitude and channel buoy target amplitude are arranged in chronological order, with the first half used as the training set and the second half as the test set.

[0105] Step S3: Extracting fractional Shannon entropy features by calculating fractional Shannon entropy using fractional parameters;

[0106] Histogram statistics are performed on the segmented target amplitude time series in step S2, dividing the target amplitude time series into B histogram bins. The number of target amplitude sequences in the i-th histogram bin is... The probability distribution corresponding to the target amplitude sequence within the i-th histogram bin. The mathematical expression is:

[0107]

[0108]

[0109] Through fractional parameters Calculate the fractional Shannon entropy The mathematical expression is:

[0110]

[0111] The normalized fractional Shannon entropy mathematical expression is:

[0112]

[0113] in, Let B be the probability distribution, and B be the number of histogram bins. The value range is [0,2], and L is the window length of the sliding window.

[0114] When α=1, it is the classic Shannon entropy; when α<1, the fractional Shannon entropy is less sensitive to random noise, and the fractional Shannon entropy feature is more stable under low SNR conditions, while when α>1, it can better emphasize low probability events.

[0115] Step S4: The parameter optimization step involves performing global parameter optimization on the temporal FSE features based on the Bach distance to obtain the optimal parameters for the temporal FSE features.

[0116] The mathematical expression for Bach's distance (BD) is:

[0117]

[0118] Wherein, under the fractional-order parameter α, the time-domain FSE feature x of the ship target and the time-domain FSE feature y of the channel buoy both follow a Gaussian distribution, that is, , , Let x and y be the means of a Gaussian distribution, respectively. , Let x and y be the variances of the Gaussian distributions they follow, respectively.

[0119] In this embodiment, the larger the Barthel distance BD, the less the overlap between the temporal FSE characteristics of the ship target and the temporal FSE characteristics of the channel buoy, and the stronger their separability.

[0120] In the training set, a grid search is performed on the fractional-order parameter α to calculate the Bádov distance (BD) value corresponding to each fractional-order parameter α. The fractional-order parameter corresponding to the maximum Bádov distance (BD) value is selected as the optimal fractional-order parameter. .

[0121] A grid search is performed on the fractional-order parameter α, with the search range set to 0.1-2 and the step size set to 0.1.

[0122] Step S5: The step of determining the sample target classification and decision threshold. Based on the optimal parameters in step S4, calculate the mean value of the time-domain FSE features of the sample target, construct the discrimination direction factor and discrimination quantity, determine the sample target classification tendency, set the decision threshold, and determine the sample target classification.

[0123] The mathematical expression for the mean of the time-domain FSE features of a ship target is:

[0124]

[0125] in, To train the temporal FSE features of a concentrated set of ship targets;

[0126] The mathematical expression for the mean of the time-domain FSE characteristic of a channel buoy is:

[0127]

[0128] in, To train the temporal FSE characteristics of channel buoys;

[0129] Discrimination Direction Factor The mathematical expression is:

[0130]

[0131] in, The mean of the time-domain FSE features of the ship target. The mean of the time-domain FSE characteristics of the channel buoy;

[0132] The mathematical expression for the discriminant s is as follows:

[0133]

[0134] Among them, the determination of the sample target in the optimal fractional order parameter will be performed. The time-domain FSE feature is denoted as H. The sample targets include ship targets or channel buoy targets;

[0135] When the discriminant s > the preset threshold, the sample target is determined to be a vessel; when the discriminant s < the preset threshold, the sample target is determined to be a waterway buoy.

[0136] The mathematical expression for the empirical false alarm rate is set as follows:

[0137]

[0138] in, The empirical false alarm rate is given by t, where t is the threshold. This represents the number of times a channel buoy sample is misidentified as a vessel when the threshold is t. The number of samples for training concentrated channel buoys;

[0139] The decision threshold is set to the maximum threshold that satisfies the empirical false alarm rate, and the mathematical expression is:

[0140]

[0141] in, The threshold for judgment, To train on the maximum permissible false alarm rate of one channel buoy misidentifying a vessel on average, ;

[0142] When the discriminant s ≥ the decision threshold At that time, the sample target was classified as a ship;

[0143] Discriminant s < decision threshold At that time, the sample target was classified as a channel buoy;

[0144] The mathematical expression for calculating the accuracy of target classification for samples is:

[0145]

[0146] Among them, TP, TN, FP, and FN are four scenarios for target determination in the sample. TP, TN, FP, and FN are respectively: ship target determined as ship, channel buoy target determined as channel buoy, channel buoy target determined as ship, and ship target determined as channel buoy.

[0147] Example 2:

[0148] like Figure 2 As shown, this embodiment also provides a sea surface target identification method using fractional entropy in the time domain, including a radar data acquisition and preprocessing module 1, a sliding window segmentation module for the target amplitude time series 2, a fractional Shannon entropy feature extraction module 3, a parameter optimization module 4, and a target classification and decision threshold module 5.

[0149] Radar data acquisition and preprocessing module 1 acquires radar echo complex data and represents it as a radar echo complex matrix to construct the target amplitude time series;

[0150] Module 2 for sliding window segmentation of the target amplitude time series uses a sliding window to segment the target amplitude time series, dividing it into training and test sets;

[0151] Module 3 extracts the fractional Shannon entropy features and calculates the fractional Shannon entropy using fractional parameters;

[0152] Parameter optimization module 4 performs global parameter optimization on the temporal FSE features based on Bach distance to obtain the optimal parameters of the temporal FSE features;

[0153] The sample target classification and decision threshold module 5, based on the optimal parameters in the parameter optimization module 4, calculates the mean value of the time-domain FSE features of the sample target, constructs the discrimination direction factor and discrimination quantity, determines the sample target classification tendency, sets the decision threshold, and determines the sample target classification.

[0154] In this technical solution, the time-domain FSE feature refers to the fractional-order Shannon entropy feature.

[0155] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. The methods disclosed in the embodiments are described simply because they correspond to the systems disclosed in the embodiments; relevant details can be found in the method section.

[0156] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0157] In the embodiments provided by this invention, it should be understood that the disclosed systems, methods, and approaches can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between systems or units may be electrical, mechanical, or other forms.

[0158] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0159] In addition, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit.

[0160] Similarly, in the various embodiments of the present invention, each processing unit can be integrated into a functional module, or each processing unit can exist physically, or two or more processing units can be integrated into a functional module.

[0161] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

[0162] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0163] The above-disclosed embodiments are merely preferred embodiments of the present invention, but the present invention is not limited thereto. Any non-creative variations that can be conceived by those skilled in the art, as well as any improvements and modifications made without departing from the principles of the present invention, should fall within the protection scope of the present invention.

Claims

1. A method for identifying sea surface targets using temporal fractional entropy, characterized in that, Includes the following steps: Step S1: The step of acquiring radar data and preprocessing, acquiring radar echo complex data and representing it as a radar echo complex matrix, and constructing the target amplitude time series; Step S2: The step of segmenting the target amplitude time series using a sliding window, dividing the target amplitude time series into training and test sets; Wherein, the time series of the target amplitude of the i-th segment of the vessel or the time series of the target amplitude of the i-th segment of the channel buoy are represented as: in, L represents the length of the time series of the target amplitude of a ship or the target amplitude of a channel buoy, s represents the window length of the sliding window, and s represents the step size of the sliding window. , To round down; The segmented time series of ship target amplitude and channel buoy target amplitude are arranged in chronological order, with the first half used as the training set and the second half as the test set. Step S3: Extracting fractional Shannon entropy features by calculating fractional Shannon entropy using fractional parameters; Histogram statistics are performed on the segmented target amplitude time series in step S2, dividing the target amplitude time series into B histogram bins. The number of target amplitude sequences in the i-th histogram bin is... The probability distribution corresponding to the target amplitude sequence within the i-th histogram bin. The mathematical expression is: Through fractional parameters Calculate the fractional Shannon entropy The mathematical expression is: The normalized fractional Shannon entropy mathematical expression is: in, Let B be the probability distribution, and B be the number of histogram bins. The value range is [0,2], and L is the window length of the sliding window; Step S4: The parameter optimization step involves performing global parameter optimization on the temporal FSE features based on the Bach distance to obtain the optimal parameters for the temporal FSE features. Step S5: The step of determining the sample target classification and decision threshold. Based on the optimal parameters in step S4, calculate the mean value of the time-domain FSE features of the sample target, construct the discrimination direction factor and discrimination quantity, determine the sample target classification tendency, set the decision threshold, and determine the sample target classification.

2. The sea surface target identification method using temporal fractional entropy according to claim 1, characterized in that, The radar echo complex data includes ship target data and channel buoy target data; The complex matrix of radar echo is ,in, For the number of pulses in the slow time dimension, The number of distance units in the fast time dimension; The target amplitude time series includes the target amplitude time series of ships and the target amplitude time series of channel buoys.

3. The sea surface target identification method using time-domain fractional entropy according to claim 2, characterized in that, The time series of target amplitudes for channel buoys is represented as follows: ,in, Represents the modulo value. is the distance unit where the channel buoy target is located, and n is the length of the pulse sequence; Indicates the distance threshold of the target buoy in the channel. Above, the amplitude sequence that varies with the pulse signal; The time series of ship target amplitude is represented as ; in, Let n be the distance unit where the ship target is located, and n be the length of the pulse sequence. The distance from the center of the target ship to the gate. The final center distance gate for coherent processing time; The mathematical expression is: in, For neighborhood windows, Distance between candidate centers This is the distance from the window size. The average energy of each distance gate; The mathematical expression is: in, This is the set of pulse indices for the k-th coherent processing time. This refers to the coherent processing time.

4. The sea surface target identification method using time-domain fractional entropy according to claim 1, characterized in that, The mathematical expression for Bach's distance (BD) is: Wherein, under the fractional-order parameter α, the time-domain FSE feature x of the ship target and the time-domain FSE feature y of the channel buoy both follow a Gaussian distribution, that is, , , Let x and y be the means of a Gaussian distribution, respectively. , Let x and y be the variances of the Gaussian distributions they follow, respectively. In the training set, a grid search is performed on the fractional-order parameter α to calculate the Bádov distance (BD) value corresponding to each fractional-order parameter α. The fractional-order parameter corresponding to the maximum Bádov distance (BD) value is selected as the optimal fractional-order parameter. .

5. The sea surface target identification method using time-domain fractional entropy according to claim 1, characterized in that, The mathematical expression for the mean of the time-domain FSE features of a ship target is: in, To train the temporal FSE features of a concentrated set of ship targets; The mathematical expression for the mean of the time-domain FSE characteristic of a channel buoy is: in, To train the temporal FSE characteristics of channel buoys.

6. The sea surface target identification method using time-domain fractional entropy according to claim 5, characterized in that, Discrimination Direction Factor The mathematical expression is: in, The mean of the time-domain FSE features of the ship target. The mean of the time-domain FSE characteristics of the channel buoy; The mathematical expression for the discriminant s is as follows: Among them, the determination of the sample target in the optimal fractional order parameter will be performed. The time-domain FSE feature is denoted as H. The sample targets include ship targets or channel buoy targets; When the discriminant s > the preset threshold, the sample target is determined to be a vessel; when the discriminant s < the preset threshold, the sample target is determined to be a waterway buoy.

7. The sea surface target identification method using time-domain fractional entropy according to claim 6, characterized in that, The mathematical expression for the empirical false alarm rate is set as follows: in, The empirical false alarm rate is given by t, where t is the threshold. This represents the number of times a channel buoy sample is misidentified as a vessel when the threshold is t. The number of samples for training concentrated channel buoys; The decision threshold is set to the maximum threshold that satisfies the empirical false alarm rate, and the mathematical expression is: in, The threshold for judgment, To train on the maximum permissible false alarm rate of one channel buoy misidentifying a vessel on average, .

8. A method for identifying sea surface targets using temporal fractional entropy according to claim 7, characterized in that, When the discriminant s ≥ the decision threshold At that time, the sample target was classified as a ship; Discriminant s < decision threshold At that time, the sample target was classified as a channel buoy; The mathematical expression for calculating the accuracy of target classification for samples is: Among them, TP, TN, FP, and FN are four scenarios for target determination in the sample. TP, TN, FP, and FN are respectively: ship target determined as ship, channel buoy target determined as channel buoy, channel buoy target determined as ship, and ship target determined as channel buoy.

Images