A sea surface floating target detection method, device, equipment and medium

Abstract

The application provides a sea surface floating target detection method, device, equipment and medium, and belongs to the technical field of target detection. The method comprises the following steps: acquiring a reference signal and a to-be-detected signal in the process of performing sea surface floating target detection by a radar; performing feature extraction on the reference signal and the to-be-detected signal respectively to obtain reference signal features corresponding to the reference signal and to-be-detected signal features corresponding to the to-be-detected signal; respectively determining a joint probability density value corresponding to the reference signal features and a joint probability density value corresponding to the to-be-detected signal features; determining a target detection threshold according to the joint probability density value corresponding to the reference signal features and a preset false alarm rate; and comparing the target detection threshold and the joint probability density value corresponding to the to-be-detected signal features to realize target detection. The application considers the correlation between features, can improve the accuracy of the feature space by constructing the joint probability density value of the reference signal features, can better distinguish targets and clutter, and improves the target detection performance.

Description

Technical Field

[0001] The present invention relates to the field of target detection technology, and in particular to a method, device, equipment and medium for detecting sea surface floating targets. Background Art

[0002] Detection of small floating targets on the sea surface is of great significance in various fields and can provide support for small vessel detection, maritime search and rescue, coastal defense security, and natural target discovery.

[0003] With a growing understanding of the physical properties and scattering mechanisms of clutter and targets, detection methods based on distinguishable features between clutter and targets have been proposed. Feature-based detection methods construct a feature space of clutter and targets and design corresponding binary classifiers for target detection. Due to the imbalance between target and clutter samples, existing binary classifiers are primarily anomaly detectors based on density and clutter spatial shape, determining anomalies by constructing a clutter sample space.

[0004] However, the existing binary classifier algorithms based on density and clutter space shape require prior knowledge of the feature space shape. When the prior knowledge is inaccurate, it is difficult to accurately represent the clutter space, which will lead to a decrease in detection performance. Summary of the Invention

[0005] The present invention provides a method, device, equipment and medium for detecting floating targets on the sea surface, which solves the defect that the existing technology is difficult to accurately characterize the clutter space, thereby causing a decrease in detection performance.

[0006] The present invention provides a method for detecting floating targets on the sea surface, comprising:

[0007] Acquire an echo signal of a radar during a process of detecting a floating target on the sea surface, wherein the echo signal includes a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, and the reference signal includes clutter;

[0008] Extracting features of the reference signal and the signal to be detected respectively to obtain reference signal features corresponding to the reference signal and signal features corresponding to the signal to be detected;

[0009] Determining the joint probability density value corresponding to the reference signal feature and the joint probability density value corresponding to the test signal feature respectively;

[0010] Determining a target detection threshold based on a joint probability density value corresponding to the reference signal feature and a preset false alarm rate;

[0011] The target detection threshold is compared with the joint probability density value corresponding to the signal feature to be detected to achieve target detection.

[0012] As an embodiment, determining the joint probability density value corresponding to the reference signal feature includes:

[0013] Determine the edge probability density value corresponding to each of the reference signal features based on a kernel density estimation algorithm;

[0014] Determining a marginal cumulative distribution function value corresponding to each reference signal feature according to a marginal probability density value corresponding to each reference signal feature;

[0015] The joint probability density value corresponding to each of the reference signal features is determined based on the empirical Bornstein copula function and the marginal cumulative distribution function value corresponding to each of the reference signal features.

[0016] As an embodiment, determining the joint probability density value corresponding to the feature of the signal to be measured includes:

[0017] The joint probability density value corresponding to each of the signal features to be measured is determined based on a kernel density estimation algorithm.

[0018] As an embodiment, the expression of the joint probability density value corresponding to all the reference signal features is as follows:

[0019]

[0020] in, , represents the cumulative distribution of signal feature edges, Indicates the i The marginal cumulative distribution function value corresponding to the reference signal feature, is the observation vector, i =1,2, is the joint empirical cumulative distribution, m is the order of the empirical Bernstein copula function, q and l is the subscript parameter of the summation function, and are the first-order derivatives of the binomial distribution functions of the two features respectively.

[0021] As an embodiment, the extracting features of the reference signal and the signal to be detected respectively to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected includes:

[0022] Performing arbitrary domain transformation on the reference signal and the signal to be detected respectively to obtain the reference signal and the signal to be detected after domain transformation;

[0023] Respective domain transformation is carried out on the reference signal and the to-be-detected signal, and corresponding domain features are extracted to obtain reference signal features corresponding to the reference signal and to-be-detected signal features corresponding to the to-be-detected signal.

[0024] As an embodiment, the target detection threshold is determined according to the joint probability density value corresponding to the reference signal features and a preset false alarm rate, which comprises:

[0025] The product of the false alarm rate and the signal length of the reference signal is determined.

[0026] The minimum value between the product and the joint probability density value is taken as the target detection threshold.

[0027] As an embodiment, the target detection threshold is compared with the joint probability density value corresponding to the to-be-detected signal features to realize target detection, which comprises:

[0028] If the joint probability density value corresponding to the to-be-detected signal features is not less than the target detection threshold, the to-be-detected signal features are determined as clutter.

[0029] If the joint probability density value corresponding to the to-be-detected signal features is less than the target detection threshold, the to-be-detected signal features are determined as target.

[0030] The application further provides a sea surface floating target detection device, which comprises:

[0031] A signal acquisition module is configured to acquire echo signals of a radar in a sea surface floating target detection process, wherein the echo signals comprise reference signals corresponding to a radar reference unit and to-be-detected signals corresponding to a radar detection unit, and the reference signals comprise clutter.

[0032] A feature extraction module is configured to extract features from the reference signals and the to-be-detected signals respectively to obtain reference signal features corresponding to the reference signals and to-be-detected signal features corresponding to the to-be-detected signals.

[0033] A first determination module is configured to determine joint probability density values corresponding to the reference signal features and the to-be-detected signal features respectively.

[0034] A second determination module is configured to determine a target detection threshold according to the joint probability density value corresponding to the reference signal features and a preset false alarm rate.

[0035] A target detection module is configured to compare the target detection threshold with the joint probability density value corresponding to the to-be-detected signal features to realize target detection.

[0036] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a method for detecting a floating target on the sea surface as described above is implemented.

[0037] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for detecting a floating target on the sea surface as described above is implemented.

[0038] The present invention provides a method, device, equipment and medium for detecting floating targets on the sea surface, which obtains an echo signal of a radar during the detection of floating targets on the sea surface, wherein the echo signal includes a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, and the reference signal includes clutter; feature extraction is performed on the reference signal and the signal to be detected, respectively, to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected; a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the signal to be detected feature are determined respectively; a target detection threshold is determined based on the joint probability density value corresponding to the reference signal feature and a preset false alarm rate; and the target detection threshold is compared with the joint probability density value corresponding to the signal to be detected feature to achieve target detection. By constructing the joint probability density value corresponding to the reference signal feature, the present invention takes into account the correlation between features in the process of constructing the feature space of clutter, thereby improving the accuracy of the feature space, helping to better distinguish targets from clutter, and improving target detection performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is one of the flow charts of a method for detecting floating targets on the sea surface provided by the present invention.

[0041] Figure 2 This is the second flow chart of a method for detecting floating targets on the sea surface provided by the present invention.

[0042] Figures 3a-3d Schematic diagrams 1 to 4 comparing the performance results of the present invention and the prior art on different data.

[0043] Figure 4It is a structural schematic diagram of a sea surface floating target detection device provided by the present invention.

[0044] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0045] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0046] It should be noted that all actions of acquiring signals, information or data in the present invention are performed in compliance with the corresponding local data protection laws and policies and with authorization from the corresponding device owner.

[0047] In the field of feature detection of floating targets on the sea surface, existing feature-based detection research can be mainly divided into the following two types: one focuses on extracting physically interpretable features in different domains, such as energy in the time domain and distribution of spectrum in the frequency domain; the other focuses on proposing better binary classifiers to achieve accurate classification, such as density-based anomaly detection algorithms and convex hull-based detector designs.

[0048] The algorithms used by existing binary classifiers need to assume the distribution of the feature space. When the actual distribution of the feature space deviates from the prior, it is difficult to accurately characterize the clutter space. In addition, the existing algorithms do not take into account the correlation between features, resulting in a decrease in the detection performance of the binary classifier.

[0049] To overcome the problem of degraded detection performance caused by existing feature detection of floating targets on the sea surface, which fails to consider the deviation between the true distribution of features and the prior and the correlation between features, the present invention provides a method, apparatus, device and medium for detecting floating targets on the sea surface. These methods combine clutter features, facilitate the establishment of an accurate clutter feature distribution model, improve the quality of feature space construction in multi-feature target detection, and thus improve target detection performance. The present invention is described in detail below with reference to the accompanying drawings.

[0050] Figure 1 This is one of the flow charts of a method for detecting floating targets on the sea surface provided by the present invention. Figure 1As shown, the present invention provides a method for detecting floating targets on the sea surface, which can be applied to the field of detecting floating targets on the sea surface, and can also be applied to the field of detecting small floating targets on the sea surface. If applied to the field of detecting small floating targets on the sea surface, the target of the technical solution is adaptively changed to a small target. The method provided by the present invention includes steps S100 to S500.

[0051] Step S100: Acquire an echo signal of a radar during detection of a floating target on the sea surface. The echo signal includes a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit. The reference signal includes clutter.

[0052] Constant False Alarm Rate Detector (CFAR) is a radar target detection algorithm that first specifies a range Doppler unit as a detection unit, sets a circle of guard units around the detection unit, and then sets a circle of reference units around the guard unit.

[0053] The radar reference unit provided in the embodiment of the present invention refers to the reference unit in the constant false alarm detection algorithm, the radar detection unit refers to the detection unit in the constant false alarm detection algorithm, and clutter refers to other interference signals or noise other than the signal reflected from the target object during the radar detection process.

[0054] Step S200 , performing feature extraction on the reference signal and the signal to be detected respectively, to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected.

[0055] The embodiment of the present invention does not limit the feature extraction algorithm, as long as separable features can be extracted from the reference signal and the signal to be detected. The reference signal is used to construct a feature space of clutter.

[0056] Step S300: determining the joint probability density value corresponding to the reference signal feature and the joint probability density value corresponding to the signal feature to be measured.

[0057] Step S400: determining a target detection threshold according to a joint probability density value corresponding to the reference signal feature and a preset false alarm rate.

[0058] Step S500 : comparing the target detection threshold with the joint probability density value corresponding to the signal feature to be detected to achieve target detection.

[0059] It can be understood that the present application considers the correlation between the features in the process of constructing the feature space of the clutter by constructing the joint probability density value corresponding to the feature of the reference signal, can improve the accuracy of the feature space, helps to better distinguish the target from the clutter, and improves the target detection performance. In addition, the present application uses the probability density value corresponding to the feature for target detection, instead of directly using the feature for target detection, which can further improve the accuracy of target detection.

[0060] On the basis of the above-mentioned embodiments, as an optional embodiment, the feature extraction is performed on the reference signal and the to-be-detected signal respectively to obtain the reference signal feature corresponding to the reference signal and the to-be-detected signal feature corresponding to the to-be-detected signal, and the method comprises the following steps.

[0061] In step S210, arbitrary domain transformation is performed on the reference signal and the to-be-detected signal respectively to obtain the domain-transformed reference signal and the domain-transformed to-be-detected signal.

[0062] The time domain transformation or the frequency domain transformation or the time-frequency domain transformation can be performed on the reference signal and the to-be-detected signal respectively, and the embodiments of the present application are not limited in this regard and are universal.

[0063] In step S220, the corresponding domain feature extraction is performed on the domain-transformed reference signal and the domain-transformed to-be-detected signal respectively to obtain the reference signal feature corresponding to the reference signal and the to-be-detected signal feature corresponding to the to-be-detected signal.

[0064] Optionally, for the transformation mode of the reference signal and the to-be-detected signal, the corresponding feature extraction algorithm is used to perform feature extraction on the domain-transformed reference signal and the domain-transformed to-be-detected signal.

[0065] For the time domain transformation, the time domain features such as signal amplitude, correlation time and envelope feature are extracted; for the frequency domain transformation, the frequency spectrum features such as main frequency, frequency spectrum width and harmonic feature are extracted by using fast Fourier transform, or the frequency domain energy features are extracted by using power spectrum density; for the time-frequency domain transformation, the time-frequency features are extracted by using the time-frequency feature extraction algorithm, such as short-time Fourier transform algorithm.

[0066] It can be understood that the present application extracts the features of the echo signal by using the arbitrary domain transformation and the corresponding domain feature extraction algorithm in order to construct the feature space of the clutter and perform target detection, which is universal and has higher flexibility.

[0067] On the basis of the above-mentioned embodiments, as an optional embodiment, the joint probability density value corresponding to the reference signal feature is determined, and the method comprises the following steps.

[0068] Step S310 : Considering the influence of noise on the estimation result, the marginal probability density value corresponding to each reference signal feature is determined based on a kernel density estimation algorithm. The kernel density estimation algorithm can reduce the influence of noise on the estimation result.

[0069] Let the reference signal be x p , the reference signal characteristics are expressed as , , the reference signal features can be understood as the features of the clutter. Correspondingly, the expression for determining the edge probability density value corresponding to each of the reference signal features using the kernel density estimation algorithm is as follows:

[0070] ;

[0071] in, For the i The marginal probability density value corresponding to the reference signal feature, h is the kernel density estimation bandwidth, N is the signal length of the reference signal, K is the density kernel function used for kernel density estimation, Representative i Reference signal characteristics No. n samples, is the observation vector, i =1,2.

[0072] Step S320 : determining a marginal cumulative distribution function value corresponding to each reference signal feature according to the marginal probability density value corresponding to each reference signal feature.

[0073] The expression of the marginal cumulative distribution function value corresponding to each of the reference signal features is as follows:

[0074] ;

[0075] in, Indicates the i The marginal cumulative distribution function value corresponding to the characteristics of the reference signal, i =1,2.

[0076] Step S330: Determine the joint probability density value corresponding to each reference signal feature based on the empirical Bernstein copula function and the marginal cumulative distribution function value corresponding to each reference signal feature. The empirical Bernstein copula function is also known as the empirical Bernstein copula function. Compared to other copula extraction methods, the empirical Bernstein copula function is more flexible and better able to characterize complex correlations between features.

[0077] The expression of the joint probability density value corresponding to all the reference signal features is as follows:

[0078] ;

[0079] in, , represents the cumulative distribution of signal feature edges, Indicates the i The marginal cumulative distribution function value corresponding to the reference signal feature, is the observation vector, i =1,2, is the joint empirical cumulative distribution, m is the order of the empirical Bernstein copula function, q and l is the subscript parameter of the summation function, and are the first-order derivatives of the binomial distribution functions of the two features respectively.

[0080] After determining the marginal probability density values ​​corresponding to all the reference signal features, the joint probability density value corresponding to the reference signal features is obtained according to the marginal probability density values ​​corresponding to all the reference signal features, and the expression is as follows:

[0081]

[0082] in, It is k The joint probability density value of each feature of the reference signal, It is k The reference signal i The marginal probability density value of each feature.

[0083] Optionally, determining a joint probability density value corresponding to the feature of the signal to be measured includes:

[0084] Step S340: Determine the joint probability density value corresponding to each of the signal features to be measured based on a kernel density estimation algorithm. The calculation method can refer to step S310. It should be noted that the embodiment of the present invention does not limit the execution order of step S310 and step S340.

[0085] It can be understood that the present invention uses the empirical Bernstein Copula function to combine the marginal distributions of various clutter features to form a multi-feature joint probability density, which can improve the accuracy of the clutter feature distribution model, help to better distinguish targets from clutter, and improve detection performance.

[0086] Based on the above embodiment, as an optional embodiment, determining the target detection threshold according to the joint probability density value corresponding to the reference signal feature and a preset false alarm rate includes the following steps.

[0087] Step S410: Determine the product of the false alarm rate and the signal length of the reference signal. In signal detection theory, the false alarm rate (FAR) refers to the probability of mistakenly determining a noise or background signal as a target signal.

[0088] Step S420: taking the minimum value between the product and the joint probability density value as the target detection threshold.

[0089] The expression of target detection threshold is as follows:

[0090]

[0091] in, is the target detection threshold, is the joint probability density value corresponding to all reference signal features, is the false alarm rate.

[0092] Optionally, the comparing the target detection threshold and the joint probability density value corresponding to the feature of the signal to be detected to achieve target detection includes the following steps.

[0093] Step S510: If the joint probability density value corresponding to the signal feature to be measured is not less than the target detection threshold, the signal feature to be measured is determined to be clutter.

[0094] Step S520: If the joint probability density value corresponding to the signal feature to be detected is less than the target detection threshold, the signal feature to be detected is determined to be a target.

[0095] It can be understood that the present invention uses the probability density function value as the detection statistic for detection, which can effectively improve the detection accuracy.

[0096] The present invention is described in detail below with reference to a preferred embodiment.

[0097] Figure 2 This is a second flow chart of a method for detecting floating targets on the sea surface provided by the present invention. Figure 2 As shown, the feature detection method for small floating targets on the sea surface based on the empirical Bernstein Copula function of the present invention includes a training process and an application process. The training process is used to determine the joint probability density function of each clutter feature to more accurately describe the clutter space and determine the detection threshold. The training process includes steps 1 to 7, and the application process includes steps 8 and 9.

[0098] The input parameters of the present invention include the reference signal x corresponding to the radar reference unit p , the signal to be detected x corresponding to the radar detection unit.

[0099] Step 1: Domain transform on reference cell signal x p The present application is not limited to a specific domain transform method and is universal.

[0100] Step 2: Extract features from the transformed domain reference cell signal Similarly, the present application is not limited to a specific feature.

[0101] Step 3: Obtain the marginal probability density function of each feature of the clutter using kernel density estimation. See step S310 for the specific expression.

[0102] Step 4: Obtain the marginal cumulative distribution function of each feature of the clutter using kernel density estimation. See step S320 for the expression.

[0103] Step 5: Obtain the joint probability density function using the empirical Bernstein Copula function. See step S330 for the expression.

[0104] Step 6: Obtain the joint probability density function value of each reference sample. See step S330 for the expression.

[0105] Step 7: Determine the constant false alarm threshold according to the quantile. See step S420 for the expression.

[0106] Step 8: Calculate the feature vector composed of all the features of the to-be-detected signal x corresponding to the to-be-detected signal Use the probability density function value corresponding to the feature vector as the detection statistic .

[0107] Step 9: If is less than , it is a target, otherwise it is clutter.

[0108] Figure 3a is a performance result comparison diagram of the present application and the prior art on HH polarized data, Figure 3b is a performance result comparison diagram of the present application and the prior art on HV polarized data, Figure 3c is a performance result comparison diagram of the present application and the prior art on VH polarized data, Figure 3d is a performance result comparison diagram of the present application and the prior art on VV polarized data. The prior art refers to a small target detection method in marine clutter based on three features. In the figure, "-o-" is the detection result obtained by using the prior detection method for different features, and "-*" is the detection result obtained by using the present application method for different features. As can be seen from the figure, the present application has better performance.

[0109] ​​In summary, the present invention relates to the field of feature detection of small targets floating on the sea surface. To address the problem that traditional feature detection methods assume feature independence, resulting in inaccurate feature space construction and low detection performance, a feature detection method for small targets floating on the sea surface based on the empirical Bernstein Copula function is proposed. Given a given feature, the method can use kernel density estimation to obtain the marginal distribution of each feature, and use the empirical Bernstein Copula function to accurately construct a joint probability density distribution. This method can accurately construct the feature space of clutter and use the probability density function value as a detection statistic for detection. The present invention has higher accuracy than existing similar methods and is a highly promising feature detection method for small targets floating on the sea surface. This method addresses the drawback of existing feature detection technologies for small targets floating on the sea surface, which directly designs a classifier in the feature space after extracting distinguishable features from target clutter, without designing an algorithm based on the correlation between features.

[0110] A sea surface floating target detection device provided by the present invention is described below. The sea surface floating target detection device described below and the sea surface floating target detection method described above can correspond to each other.

[0111] Figure 4 This is a schematic diagram of the structure of a sea surface floating target detection device provided by the present invention. Figure 4 As shown, the present invention also provides a sea surface floating target detection device, which includes the following modules.

[0112] The signal acquisition module 410 is used to acquire the echo signal of the radar during the process of detecting floating targets on the sea surface. The echo signal includes a reference signal corresponding to the radar reference unit and a signal to be detected corresponding to the radar detection unit. The reference signal includes clutter.

[0113] A feature extraction module 420 is configured to extract features from the reference signal and the signal to be detected, respectively, to obtain reference signal features corresponding to the reference signal and signal features corresponding to the signal to be detected;

[0114] A first determination module 430 is configured to respectively determine a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the test signal feature;

[0115] A second determination module 440 is configured to determine a target detection threshold based on a joint probability density value corresponding to the reference signal feature and a preset false alarm rate;

[0116] The target detection module 450 is configured to compare the target detection threshold with the joint probability density value corresponding to the feature of the signal to be detected to achieve target detection.

[0117] As an embodiment, the first determining module 430 is further configured to:

[0118] Determine the edge probability density value corresponding to each of the reference signal features based on a kernel density estimation algorithm;

[0119] Determining a marginal cumulative distribution function value corresponding to each reference signal feature according to a marginal probability density value corresponding to each reference signal feature;

[0120] The joint probability density value corresponding to each of the reference signal features is determined based on the empirical Bornstein copula function and the marginal cumulative distribution function value corresponding to each of the reference signal features.

[0121] As an embodiment, the first determining module 430 is further configured to:

[0122] The joint probability density value corresponding to each of the signal features to be measured is determined based on a kernel density estimation algorithm.

[0123] As an embodiment, the expression of the joint probability density value corresponding to all the reference signal features is as follows:

[0124]

[0125] in, , represents the feature marginal cumulative distribution, Indicates the i The marginal cumulative distribution function value corresponding to the reference signal feature, is the observation vector, i =1,2, is the joint empirical cumulative distribution, m is the order of the empirical Bernstein copula function, q and l is the subscript parameter of the summation function, and are the first-order derivatives of the binomial distribution functions of the two features respectively.

[0126] As an embodiment, the feature extraction module 420 is further configured to:

[0127] Performing arbitrary domain transformation on the reference signal and the signal to be detected respectively to obtain the reference signal and the signal to be detected after domain transformation;

[0128] Corresponding domain features are extracted from the reference signal and the signal to be detected after domain transformation, respectively, to obtain reference signal features corresponding to the reference signal and signal features to be detected corresponding to the signal to be detected.

[0129] As an embodiment, the second determining module 440 is further configured to:

[0130] determining a product of the false alarm rate and a signal length of the reference signal;

[0131] The minimum value between the product and the joint probability density value is used as the target detection threshold.

[0132] As an embodiment, the target detection module 450 is further configured to:

[0133] If the joint probability density value corresponding to the signal feature to be measured is not less than the target detection threshold, the signal feature to be measured is determined to be clutter;

[0134] If the joint probability density value corresponding to the signal feature to be detected is less than the target detection threshold, the signal feature to be detected is determined to be a target.

[0135] It should be noted that the sea surface floating target detection device provided by the present invention can execute a sea surface floating target detection method described in any of the above embodiments during specific operation, and has the technical effect corresponding to the method, which will not be elaborated in this embodiment.

[0136] Figure 5 An example of a physical structure diagram of an electronic device is shown below. Figure 5 As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communications bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other via the communications bus 540. The processor 510 may call logic instructions in the memory 530 to execute a method for detecting floating targets on the sea surface. The method includes: obtaining an echo signal from a radar during the process of detecting floating targets on the sea surface, wherein the echo signal includes a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, wherein the reference signal includes clutter; performing feature extraction on the reference signal and the signal to be detected to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected; determining a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the signal to be detected feature; determining a target detection threshold based on the joint probability density value corresponding to the reference signal feature and a preset false alarm rate; and comparing the target detection threshold with the joint probability density value corresponding to the signal to be detected feature to achieve target detection.

[0137] Furthermore, the logic instructions in the aforementioned memory 530 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0138] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a method for detecting floating targets on the sea surface provided by the above methods, the method including: obtaining an echo signal of a radar during the process of detecting floating targets on the sea surface, the echo signal including a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, the reference signal including clutter; performing feature extraction on the reference signal and the signal to be detected respectively to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected; determining a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the signal to be detected feature respectively; determining a target detection threshold based on the joint probability density value corresponding to the reference signal feature and a preset false alarm rate; and comparing the target detection threshold with the joint probability density value corresponding to the signal to be detected feature to achieve target detection.

[0139] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute a method for detecting floating targets on the sea surface provided by the above-mentioned methods, the method comprising: obtaining an echo signal of a radar during the process of detecting floating targets on the sea surface, the echo signal comprising a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, the reference signal comprising clutter; performing feature extraction on the reference signal and the signal to be detected respectively to obtain a reference signal feature corresponding to the reference signal and a signal to be detected feature corresponding to the signal to be detected; determining a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the signal to be detected feature respectively; determining a target detection threshold based on the joint probability density value corresponding to the reference signal feature and a preset false alarm rate; and comparing the target detection threshold with the joint probability density value corresponding to the signal to be detected feature to achieve target detection.

[0140] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0141] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0142] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for detecting floating targets on the sea surface, characterized in that: include: Acquire an echo signal of a radar during a process of detecting a floating target on the sea surface, wherein the echo signal includes a reference signal corresponding to a radar reference unit and a signal to be detected corresponding to a radar detection unit, and the reference signal includes clutter; Extracting features of the reference signal and the signal to be detected respectively to obtain reference signal features corresponding to the reference signal and signal features corresponding to the signal to be detected; Determining the joint probability density value corresponding to the reference signal feature and the joint probability density value corresponding to the test signal feature respectively; Determine the product of a preset false alarm rate and the signal length of the reference signal, and use the minimum value between the product and the joint probability density value corresponding to the reference signal feature as the target detection threshold; The target detection threshold is compared with the joint probability density value corresponding to the signal feature to be detected to achieve target detection.

2. A method for detecting floating targets on the sea surface according to claim 1, characterized in that: Determining a joint probability density value corresponding to the reference signal feature includes: Determine the edge probability density value corresponding to each of the reference signal features based on a kernel density estimation algorithm; Determining a marginal cumulative distribution function value corresponding to each reference signal feature according to a marginal probability density value corresponding to each reference signal feature; The joint probability density value corresponding to each of the reference signal features is determined based on the empirical Bornstein copula function and the marginal cumulative distribution function value corresponding to each of the reference signal features.

3. A method for detecting floating targets on the sea surface according to claim 2, characterized in that: Determining a joint probability density value corresponding to the feature of the signal to be measured includes: The joint probability density value corresponding to each of the signal features to be measured is determined based on a kernel density estimation algorithm and an empirical Bernstein link function.

4. A method for detecting floating targets on the sea surface according to claim 2 or 3, characterized in that: The marginal probability density values ​​corresponding to all the reference signal features are expressed as follows: in, represents the cumulative distribution of signal feature edges, represents the marginal cumulative distribution function value corresponding to the i-th reference signal feature, ξ i,v is the observation vector, i=1,2, is the joint empirical cumulative distribution, m is the order of the empirical Bernstein copula function, q and l are the index parameters of the summation function, and are the first-order derivatives of the binomial distribution functions of the two features respectively.

5. A method for detecting floating targets on the sea surface according to claim 1, characterized in that: The extracting features of the reference signal and the signal to be detected respectively to obtain reference signal features corresponding to the reference signal and signal features corresponding to the signal to be detected includes: Performing arbitrary domain transformation on the reference signal and the signal to be detected respectively to obtain the reference signal and the signal to be detected after domain transformation; Corresponding domain features are extracted from the reference signal and the signal to be detected after domain transformation, respectively, to obtain reference signal features corresponding to the reference signal and signal features to be detected corresponding to the signal to be detected.

6. A method for detecting floating targets on the sea surface according to claim 1, characterized in that: The comparing the target detection threshold with the joint probability density value corresponding to the feature of the signal to be detected to achieve target detection includes: If the joint probability density value corresponding to the signal feature to be measured is not less than the target detection threshold, the signal feature to be measured is determined to be clutter; If the joint probability density value corresponding to the signal feature to be detected is less than the target detection threshold, the signal feature to be detected is determined to be a target.

7. A device for detecting floating targets on the sea surface, characterized in that: include: A signal acquisition module is used to acquire the echo signal of the radar during the process of detecting floating targets on the sea surface, wherein the echo signal includes a reference signal corresponding to the radar reference unit and a signal to be detected corresponding to the radar detection unit, and the reference signal includes clutter; A feature extraction module is used to extract features of the reference signal and the signal to be detected respectively, to obtain reference signal features corresponding to the reference signal and signal features corresponding to the signal to be detected; A first determination module is used to respectively determine a joint probability density value corresponding to the reference signal feature and a joint probability density value corresponding to the test signal feature; a second determination module, configured to determine a product of a preset false alarm rate and a signal length of the reference signal, and use a minimum value between the product and a joint probability density value corresponding to the reference signal feature as a target detection threshold; The target detection module is used to compare the target detection threshold with the joint probability density value corresponding to the signal feature to achieve target detection.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for detecting floating targets on the sea surface as claimed in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, a method for detecting floating targets on the sea surface as claimed in any one of claims 1 to 6 is implemented.

Images