Radar target detection method, system, device and medium based on two feature fusion

Abstract

The application provides a radar target detection method, system, device and medium based on two-feature fusion, and belongs to the technical field of radar signal processing. The method comprises the following steps: constructing training data by using existing radar target detection data, extracting phase linearity based on a training data extraction process, and extracting the relative peak height of the Doppler spectrum by using a relative peak height extraction process; based on the phase linearity and the relative peak height, a linear dimension reduction method with compact distribution is used to obtain a fusion matrix of two features; reference distance unit data and to-be-detected distance unit data are obtained in real time, the phase linearity extraction process and the relative peak height extraction process are respectively performed, and the fusion matrix is used to generate fusion features; based on the fusion features, the parameters of a generalized extreme value distribution are estimated, a detection statistic is constructed, a preset false alarm probability is combined to calculate an adaptive threshold, the target is judged by comparing the detection statistic with the threshold, and the constant false alarm rate detection under clutter suppression is realized.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology, and more specifically relates to a radar target detection method, system, device and medium based on two feature fusion. Background Technology

[0002] Detection of weak targets in strong sea clutter has always been a widely studied topic in the radar field. Radar target detectors can be broadly classified into energy-based detectors and feature-based detectors. Energy-based detectors utilize the differences in position and / or energy accumulation between targets and clutter in various domains to enhance the signal-to-clutter ratio of target cells, thereby improving target detection performance. Feature-based detectors, on the other hand, utilize the differences in certain characteristics between targets and clutter, extracting these characteristics to increase the contrast between target cells and clutter cells, thus improving detection performance.

[0003] In recent years, research on feature-based detectors has been increasing. While some features exhibit high sensitivity to low signal-to-clutter ratios, most features do not necessarily possess this desirable property, especially under short time-series conditions. Therefore, employing a multi-feature joint detector for radar target detection can significantly improve the detection performance of weak targets on the sea surface by fully utilizing the information in high-dimensional space.

[0004] Currently, the decision space of multi-feature joint detectors is typically solved using convex hull or concave hull learning algorithms. However, the accuracy of the resulting decision space largely depends on the number of multi-feature samples under target-free conditions, the geometry of the feature sample distribution in the multi-dimensional feature space, and the accuracy of the convex hull / concave hull algorithm itself. Experimental data analysis shows that the existence of decision space estimation errors severely impacts the detection performance of multi-feature joint detectors. An effective solution is feature dimensionality reduction, particularly fusing multiple features into a single feature for detection. In this case, the decision space degenerates into a detection threshold. Under the same data volume, the estimation error of the detection threshold will be significantly reduced compared to estimating the decision space. Summary of the Invention

[0005] To address the above problems, the present invention aims to provide a radar target detection method, system, device, and medium based on two-feature fusion. By fusing phase linearity and relative peak height as two features, and utilizing linear dimensionality reduction and generalized extreme value distribution modeling, constant false alarm rate radar target detection under clutter suppression is achieved, thereby improving detection performance.

[0006] To achieve the above objectives, the present invention employs the following technical solution:

[0007] In a first aspect, embodiments of this application provide a radar target detection method based on two-feature fusion, comprising:

[0008] Training data is constructed using existing radar target detection data. Phase linearity is extracted based on the training data extraction process, and the relative peak height of the Doppler spectrum is extracted using the relative peak height extraction process.

[0009] Based on the phase linearity and relative peak height, the fusion matrix of the two features is obtained using a linear dimensionality reduction method with compact distribution.

[0010] The reference distance cell data and the distance cell data to be detected are acquired in real time. The phase linearity extraction process and the relative peak height extraction process are executed respectively, and the fusion matrix is ​​used to generate fusion features.

[0011] Based on the fusion features, the generalized extreme value distribution parameters are estimated and the detection statistics are constructed. An adaptive threshold is calculated by combining the preset false alarm probability. The target decision is completed by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression.

[0012] In an optional implementation, the phase linearity extraction process includes:

[0013] The first in the training data indivual The entanglement phase of the data is represented as:

[0014]

[0015] Let the difference between adjacent winding phases be . Adjust it using the following formula. arrive Between, to eliminate Jump:

[0016]

[0017] The true phase after untangling is represented as:

[0018]

[0019] Let the phase difference between two adjacent phases of the true phase sequence be... Then the phase linearity is expressed as:

[0020]

[0021] in, Then the normalized phase linearity is:

[0022] .

[0023] In an optional implementation, the relative peak height extraction process includes:

[0024] Let the first distance unit be the first distance unit. indivual Data is Then the Doppler amplitude spectrum can be expressed by the following formula:

[0025]

[0026] in, For Doppler frequency, The pulse repetition period;

[0027] Let the peak value of the Doppler amplitude spectrum be The relative peak height can then be expressed as:

[0028]

[0029] in, The extent of the Doppler sidelobe. for The number of Doppler frequency points within the range;

[0030] The normalized relative peak height is:

[0031] .

[0032] In an optional implementation, obtaining the fusion matrix of the two features based on the phase linearity and relative peak height using a linear dimensionality reduction method with distribution compactness includes:

[0033] Based on the phase linearity and relative peak height, a feature dataset is generated. ;

[0034] in, Let be the number of samples in the feature dataset, and be the th . Feature data samples Includes phase linearity features and relative peak height characteristics ,Right now:

[0035]

[0036] After linear dimensionality reduction, the fused feature dataset can be obtained:

[0037]

[0038] set up , They represent the fusion features respectively. Given the expected value and variance, the kurtosis is calculated as follows:

[0039]

[0040] in, Indicates taking the expected value;

[0041] set up The probability distribution function is The interquartile range is then expressed as:

[0042]

[0043] in, , They are respectively The lower quartile and upper quartile;

[0044] The objective function is defined as follows:

[0045]

[0046] in, ;

[0047] exist Traverse within the interval When the objective function When the minimum value is reached, the corresponding fusion matrix This is the optimal weighted vector for the two features PL and RPH.

[0048] In an optional implementation, the real-time acquisition of reference range cell data and target range cell data includes:

[0049] The sea radar is set to operate in coherent processing mode, with the number of coherent pulse trains being [number missing]. After orthogonal demodulation and matched filtering, the I / Q data in the range cell to be detected is denoted as... The I / Q data in the reference range cell is denoted as ;

[0050] in, Indicates the distance unit to be detected. The subscript represents the number of reference distance cells. Indicates the first One reference distance cell;

[0051] The target I / Q data that may exist in the range cell to be detected are denoted as... The clutter I / Q data in the range cell to be detected and the reference range cell are respectively denoted as and .

[0052] In an optional implementation, estimating the generalized extreme value distribution parameters and constructing the detection statistic based on the fusion features includes:

[0053] The probability density function of the fused features is modeled using a generalized extreme value distribution, as follows:

[0054]

[0055] in It is a fusion of feature variables. For position parameters, For scale parameters, For shape parameters;

[0056] The shape parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0057]

[0058] in, , , , ;

[0059] from Obtained from each reference distance cell fusion feature samples The sequence is obtained by arranging the sequences in ascending order. ;

[0060] The scale parameter is obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0061]

[0062] The location parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0063]

[0064] in, It is the Gamma function;

[0065] Based on the form of the probability density function of the GEV distribution of the fused features under the condition of no target data, the following detection statistic is designed:

[0066]

[0067] When fusion features The number of samples tends to infinity, making , , At that time, the detection statistic can be derived. The asymptotic probability density function is:

[0068] .

[0069] In an optional implementation, the step of calculating the adaptive threshold by combining a preset false alarm probability includes:

[0070] The false alarm probability is calculated using the following formula. Corresponding detection thresholds:

[0071] .

[0072] Secondly, embodiments of this application also provide a radar target detection system based on two-feature fusion, comprising:

[0073] The data extraction module is used to construct training data using existing radar target detection data, extract phase linearity based on the training data extraction process, and extract the relative peak height of the Doppler spectrum using the relative peak height extraction process.

[0074] The fusion rule generation module is used to obtain the fusion matrix of the two features based on the phase linearity and relative peak height using a linear dimensionality reduction method with distribution compactness;

[0075] The measured data fusion module is used to acquire reference distance cell data and distance cell data to be detected in real time, and to perform the extraction process of phase linearity and the extraction process of relative peak height respectively, and to generate fusion features using the fusion matrix;

[0076] The target detection module is used to estimate the generalized extreme value distribution parameters and construct the detection statistics based on the fused features. It calculates the adaptive threshold by combining the preset false alarm probability and completes the target decision by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression.

[0077] Thirdly, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the radar target detection method based on two-feature fusion as described in any of the above.

[0078] Fourthly, embodiments of this application also provide a storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the radar target detection method based on two-feature fusion as described in any of the above claims.

[0079] As can be seen from the above technical solutions, the present invention has the following advantages:

[0080] The radar target detection method based on two-feature fusion provided in this application improves target detection performance by fusing phase linearity and relative peak height, two key features in radar target detection. First, a linear dimensionality reduction technique based on distribution compactness is used to reduce the dimensionality of both phase linearity and relative peak height. This process, through linear transformation, fuses the two features into a more representative fused feature, effectively reducing the feature dimensionality and improving the efficiency of subsequent processing. Second, a generalized extreme value distribution is used to model the fused feature. By modeling the probability density function, the statistical characteristics of the fused feature can be described more accurately, providing a foundation for the construction of subsequent detection statistics. Finally, based on the generalized extreme value distribution model, detection statistics are constructed, and an adaptive threshold is calculated by combining a preset false alarm probability. In this process, by comparing the detection statistics with the threshold, accurate target determination is achieved, effectively suppressing clutter interference and improving the accuracy and reliability of radar target detection.

[0081] This application delves into two key features in radar target detection: phase linearity and relative peak height. It utilizes linear dimensionality reduction techniques, such as Principal Component Analysis (PCA) or Linear Discriminant Analysis (LDA), to reduce and fuse these features. This approach generates more discriminative fused features, effectively reducing redundant information and improving the feature representation capability and accuracy of radar target detection.

[0082] This application employs the Generalized Extreme Value Distribution (GEV) to model the fused features. Through probability density function estimation and parameter fitting, it more accurately describes the statistical properties of the fused features. This modeling process considers the possible distribution patterns of the fused features, enhancing the robustness and adaptability of detection. In complex clutter environments, GEV modeling enables the detection statistics to have stronger resistance to clutter interference, improving the reliability and stability of detection.

[0083] This application combines a preset false alarm probability with constant false alarm rate (CFAR) detection technology to automatically adjust the detection threshold according to environmental changes. By estimating the statistical characteristics of background clutter in real time and calculating the adaptive threshold accordingly, constant false alarm rate radar target detection is achieved. This detection method effectively suppresses clutter interference and noise effects, ensuring the stability and consistency of detection performance, and enabling the method to maintain excellent detection results under different environments and conditions. Attached Figure Description

[0084] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0085] Figure 1 This is a flowchart illustrating the radar target detection method based on two-feature fusion provided in this application.

[0086] Figure 2 The principle block diagram of the radar target detection method based on two feature fusion provided in this application.

[0087] Figure 3 A two-dimensional feature data point map provided for this application.

[0088] Figure 4 Histogram of fusion features provided for this application.

[0089] Figure 5 A schematic diagram showing the fitting results of the probability density function of the fused feature data provided in this application.

[0090] Figure 6 This is a schematic diagram of the radar target detection system based on two-feature fusion provided in this application.

[0091] Figure 7 A schematic diagram of the structure of the electronic device provided in this application. Detailed Implementation

[0092] The various embodiments of this disclosure will be described more fully in the detailed steps of the radar target detection method based on two-feature fusion described below. This disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of this disclosure to the specific embodiments disclosed herein, but rather this disclosure should be understood to cover all adjustments, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of this disclosure.

[0093] In the following, the terms “comprising” or “may include”, which may be used in various embodiments of this disclosure, indicate the presence of the disclosed functions, operations, or elements, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in various embodiments of this disclosure, the terms “comprising,” “having,” and their cognates are intended only to indicate a particular feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or the possibility of adding one or more combinations of the foregoing.

[0094] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0095] Please see Figure 1 The diagram shows a flowchart of a radar target detection method based on two-feature fusion in a specific embodiment. The method includes:

[0096] S1: Construct training data using existing radar target detection data, extract phase linearity based on the training data using the phase linearity extraction process, and extract the relative peak height of the Doppler spectrum using the relative peak height extraction process.

[0097] S2: Based on the phase linearity and relative peak height, the fusion matrix of the two features is obtained by using a linear dimensionality reduction method with compact distribution.

[0098] S3: Real-time acquisition of reference range cell data and target range cell data using signal model, execution of phase linearity extraction process and relative peak height extraction process respectively, and generation of fusion features using fusion matrix.

[0099] S4: Based on the fusion features, estimate the generalized extreme value distribution parameters and construct the detection statistics. Combine the preset false alarm probability to calculate the adaptive threshold. Use the comparison between the detection statistics and the threshold to complete the target decision, so as to achieve constant false alarm detection under clutter suppression.

[0100] In this embodiment, the radar target detection method based on two feature fusion is divided into two stages: training and detection. For example... Figure 2 As shown, the purpose of the training phase is to obtain the fusion rules for the two features PL and RPH using the training data and a linear dimensionality reduction method based on distribution compactness. The training data refers to a large amount of I / Q data without target range cells obtained under the same observation conditions. The detection phase, based on the range cell data to be detected and the reference range cell data, uses the fusion rules to form fusion features, and on this basis, constructs detection statistics to complete the threshold detection.

[0101] In this embodiment, by cleverly fusing the complementary features of phase linearity and relative peak height, and utilizing the linear dimensionality reduction technique with compact distribution to achieve effective feature fusion, and by combining the generalized extreme value distribution to model the fused features, an efficient detection statistic is constructed. Thus, constant false alarm rate radar target detection is achieved in complex clutter environments, greatly improving the accuracy, robustness and adaptability of target detection.

[0102] In one embodiment of the present invention, based on the phase linearity extraction process and the relative peak height extraction process disclosed in steps S1 and S3, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0103] In this invention, the phase linearity extraction process and the relative peak height extraction process disclosed are essentially feature extraction processes. The basic principle is to use the N I / Q data obtained in each distance cell to extract two features: phase linearity and relative peak height of the Doppler spectrum, and to normalize the two features relative to their respective maximum and minimum values.

[0104] Phase linearity describes The phase linearity in long I / Q data, abbreviated as PL, is based on the idea that the phase linearity increases when a target is present in the I / Q data sequence. Based on this idea, the phase linearity extraction process of this invention includes:

[0105] 1. Take the training data or the first distance unit in each distance unit. indivual The entanglement phase of the data is represented as:

[0106]

[0107] Let the difference between adjacent winding phases be . Adjust it using the following formula. arrive Between, to eliminate Jump:

[0108]

[0109] The true phase after untangling is represented as:

[0110]

[0111] 2. Let the phase difference between two adjacent phases of the true phase sequence be... Then the phase linearity is expressed as:

[0112]

[0113] in, Then the normalized phase linearity is:

[0114] .

[0115] The relative peak height of the Doppler spectrum describes what indivual The energy concentration of data in the Doppler domain, abbreviated as Relative Peak Height (RPH), is based on the idea that target echo energy is more concentrated in the Doppler domain than sea clutter energy. Based on this idea, the relative peak height extraction process of this invention includes:

[0116] 1. Let the first distance unit be the first distance unit. indivual Data is Then the Doppler amplitude spectrum can be expressed by the following formula:

[0117]

[0118] in, For Doppler frequency, The pulse repetition period;

[0119] 2. Let the peak value of the Doppler amplitude spectrum be... The relative peak height can then be expressed as:

[0120]

[0121] in, The extent of the Doppler sidelobe. for The number of Doppler frequency points within the range;

[0122] The normalized relative peak height is:

[0123] .

[0124] In one embodiment of the present invention, based on step S2, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0125] In step S2, a linear dimensionality reduction method with compact distribution is disclosed. The linear dimensionality reduction problem of this invention is to reduce the two-dimensional feature space formed by PL and RPH to a one-dimensional feature space, which is equivalent to searching for a straight line in the two-dimensional feature space such that the objective function is optimal when the two-dimensional feature data is projected onto this line. The line search process involves traversing all possible direction vectors. To achieve this, in which This represents the angle with the horizontal axis; the key lies in the design of the objective function.

[0126] In radar target detection, a compact probability density function for clutter data is generally more beneficial for improving detection performance. Therefore, this invention introduces kurtosis and interquartile range (IQR) to measure the compactness of the probability density function of the clutter fusion features after linear dimensionality reduction. Kurtosis reflects the sharpness of the probability density function at the mean, while IQR reflects the concentration of the data.

[0127] Based on the above principles, the linear dimensionality reduction method for distribution compactness disclosed in this invention specifically includes the following process:

[0128] Based on the phase linearity and relative peak height, a feature dataset is generated. Among them, phase linearity and relative peak height are generated by feature extraction from training data composed of observation data of targetless distance units.

[0129] Referring to the feature dataset mentioned above, Let be the number of samples in the feature dataset, and be the th . Feature data samples Includes phase linearity features and relative peak height characteristics ,Right now:

[0130]

[0131] After linear dimensionality reduction, the fused feature dataset can be obtained:

[0132]

[0133] set up , They represent the fusion features respectively. Given the expected value and variance, the kurtosis is calculated as follows:

[0134]

[0135] in, Indicates taking the expected value;

[0136] set up The probability distribution function is The interquartile range is then expressed as:

[0137]

[0138] in, , They are respectively The lower quartile and upper quartile;

[0139] The objective function is defined as follows:

[0140]

[0141] in, ;

[0142] exist Traverse within the interval When the objective function When the minimum value is reached, the corresponding direction vector This is the optimal weighted vector for the two features PL and RPH.

[0143] When the measured data is processed using the linear dimensionality reduction method described above for compact distribution, refer to Figure 3 The given two-dimensional feature data point plot shows the measured data, and the line corresponding to the optimal weighted vector is marked on the plot. (Reference) Figure 4 The histogram of fused features is shown, where blue represents data without the target and red represents data with the target. It can be seen that the Pearson correlation coefficient between the PL and RPH features extracted based on data without the target is -0.12, and its absolute value is less than 1 / e, indicating a weak correlation between the two features.

[0144] In one embodiment of the present invention, based on the signal model disclosed in step S3, a possible embodiment will be given below, and its specific implementation will be described in a non-limiting manner.

[0145] In step S3, the reference range cell data and the range cell data to be detected are acquired in real time using a signal model, including:

[0146] The sea radar is set to operate in coherent processing mode, with the number of coherent pulse trains being [number missing]. After orthogonal demodulation and matched filtering, the I / Q data in the range cell to be detected is denoted as... ,in, This represents the range cell to be detected; the I / Q data in the reference range cell is denoted as... ,

[0147] in The subscript represents the number of reference distance cells. Indicates the first There are one reference range cell, and there is no target echo in the reference range cell.

[0148] The target I / Q data that may exist in the range cell to be detected are denoted as... The clutter I / Q data in the range cell to be detected and the reference range cell are respectively denoted as and The two are statistically independent and identically distributed.

[0149] Therefore, the object detection problem can be expressed as a binary hypothesis testing problem as follows:

[0150]

[0151] in, This indicates that there is no target in the range cell to be detected. This indicates that a target exists within the range cell to be detected.

[0152] It should be noted that, using the feature extraction method and the linear dimensionality reduction method with distribution compactness disclosed in the above embodiments, the binary hypothesis testing problem can be formulated as follows:

[0153]

[0154] in, This represents the fused feature formed after linear dimensionality reduction of the two-dimensional feature space composed of PL and RPH features; This represents the fusion characteristics of target clutter in the range cell to be detected. This represents the clutter fusion characteristics within the reference range cell. Under the assumption that clutter statistics are statistically independent and identically distributed in each range cell, the clutter fusion characteristics obtained from each range cell are... They are also statistically independent and identically distributed.

[0155] In one embodiment of the present invention, based on step S4, the following will provide a possible embodiment and describe its specific implementation in a non-limiting manner.

[0156] This embodiment discloses a generalized extreme value distribution modeling and detection statistic design process, which specifically includes the following two parts:

[0157] 1. Modeling of generalized extreme value distribution of fused features:

[0158] This embodiment uses the Generalized Extreme Value Distribution (GEV) to model the probability density function of the fused features generated after dimensionality reduction of the target unit features. The GEV distribution is a three-parameter distribution; compared to single-parameter or two-parameter distributions, the high degrees of freedom of GEV allow it to flexibly fit the tail characteristics of different types of distributions. The distribution function and probability density function of the GEV distribution are shown below:

[0159]

[0160] in It is a fusion feature variable; The location parameter determines the central location of the distribution; The scale parameter determines the extent of the distribution's expansion; The shape parameter determines the shape and tail characteristics of the distribution;

[0161] The shape parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0162]

[0163] in, , , , ;

[0164] from Obtained from each reference distance cell fusion feature samples The sequence is obtained by arranging the sequences in ascending order. ;

[0165] The scale parameter is obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0166]

[0167] The location parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is:

[0168]

[0169] in, It is the Gamma function.

[0170] Figure 5 against Figure 4 The histogram of the fused feature data generated after dimensionality reduction of the target unit features is first normalized to obtain the probability density function. Then, the GEV distribution, exponential distribution, Rayleigh distribution, log-normal distribution, and Weibull distribution are used for fitting, and the mean square error of the theoretical and empirical values ​​of the tail of the probability density function is calculated.

[0171] The results show that the GEV distribution has the best fitting effect, as shown in Table 1 below.

[0172] Table 1: Comparison of Fitting Results

[0173]

[0174] 2. Design of detection statistics:

[0175] Based on the form of the probability density function of the GEV distribution of the fused features under the condition of no target data, the following detection statistic is designed:

[0176]

[0177] When fusion features The number of samples tends to infinity, making , , At that time, the detection statistic can be derived. The asymptotic probability density function is:

[0178] .

[0179] The above equation shows that, in the absence of target data, feature fusion... When the sample size approaches infinity, the detection statistic... The asymptotic probability density function is independent of any background parameters, i.e., the detection statistic... It is asymptotically close to CFAR.

[0180] Therefore, the false alarm probability can be calculated using the following formula. Corresponding detection thresholds:

[0181] .

[0182] like Figure 6 As shown, the following are embodiments of the radar target detection system based on two-feature fusion provided in this disclosure. This system and the radar target detection method based on two-feature fusion in the above embodiments belong to the same inventive concept. For details not described in detail in the embodiments of the radar target detection system based on two-feature fusion, please refer to the embodiments of the radar target detection method based on two-feature fusion described above.

[0183] A radar target detection system based on two-feature fusion includes:

[0184] The data extraction module is used to construct training data using existing radar target detection data, extract phase linearity based on the training data using the phase linearity extraction process, and extract the relative peak height of the Doppler spectrum using the relative peak height extraction process.

[0185] The fusion rule generation module is used to obtain the fusion matrix of the two features based on the phase linearity and relative peak height using a linear dimensionality reduction method with distribution compactness.

[0186] The measured data fusion module is used to acquire reference range cell data and target range cell data in real time using the signal model, perform phase linearity extraction and relative peak height extraction processes respectively, and generate fusion features using the fusion matrix.

[0187] The target detection module is used to estimate the generalized extreme value distribution parameters and construct the detection statistics based on the fused features. It calculates the adaptive threshold by combining the preset false alarm probability and completes the target decision by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression.

[0188] The radar target detection system based on two-feature fusion provided in this embodiment deeply mines and utilizes two key features in radar target detection—phase linearity and relative peak height—and effectively fuses them using linear dimensionality reduction technology, forming more discriminative fused features and significantly improving detection performance. Simultaneously, by modeling the fused features using a generalized extreme value distribution, its statistical characteristics are described more accurately, enhancing the robustness and adaptability of the detection. Combined with a preset false alarm probability calculation adaptive threshold, constant false alarm rate (CEPR) radar target detection is achieved, effectively suppressing clutter interference and noise effects, and ensuring the stability and reliability of detection performance.

[0189] Figure 7 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0190] The radar target detection method based on two-feature fusion provided in this application can be applied to electronic devices. Those skilled in the art will understand that the electronic device structure involved in the embodiments of this invention does not constitute a limitation on the electronic device. An electronic device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements. In the embodiments of this invention, the electronic device includes, but is not limited to, laptop computers, desktop computers, workbenches, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described and / or claimed herein.

[0191] Electronic devices may include processors, external memory interfaces, internal memory, universal serial bus (USB) interfaces, charging management modules, power management modules, batteries, wireless communication modules, audio modules, speakers, microphones, sensor modules, buttons, cameras, displays, and SIM card interfaces, etc.

[0192] A processor may include one or more processing units, such as: a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural network processing unit (NPU). Different processing units may be independent devices or integrated into one or more processors.

[0193] The processor can serve as the nerve center and command center of an electronic device. The controller can generate operation control signals based on the instruction opcode and timing signals to control the fetching and execution of instructions.

[0194] The processor may also include memory for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. This memory can store instructions or data that the processor has just used or that are used repeatedly. If the processor needs to use the instruction or data again, it can retrieve it directly from this memory. This avoids repeated accesses, reduces processor latency, and thus improves system efficiency.

[0195] An external storage interface (ESI) can be used to connect external memory cards, such as microSD cards, to expand the storage capacity of electronic devices. The external memory card communicates with the processor through the ESI to perform data storage functions, such as saving music and video files on the external memory card.

[0196] Internal memory can be used to store computer executable program code, which includes instructions. The processor executes various functional applications and data processing of electronic devices by running the instructions stored in internal memory. Internal memory can include a program storage area and a data storage area. Internal memory can include high-speed random access memory, and can also include non-volatile memory, such as at least one disk storage device, flash memory device, universal flash storage (UFS), etc.

[0197] Wireless communication functionality in electronic devices can be achieved through antennas, wireless communication modules, modem processors, and baseband processors.

[0198] Wireless communication modules can provide solutions for wireless communication applications in electronic devices, including wireless local area networks (WLANs) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), and infrared (IR) technologies.

[0199] Electronic devices can implement audio functions through audio modules, speakers, receivers, microphones, headphone jacks, and application processors.

[0200] Electronic devices can achieve shooting functions through ISPs, cameras, video codecs, GPUs, displays, and application processors.

[0201] Electronic devices can achieve display functions through GPUs, displays, and application processors.

[0202] A GPU is a microprocessor for image processing, connected to the display screen and application processor. GPUs are used to perform mathematical and geometric calculations for graphics rendering. A processor may include one or more GPUs, which execute program instructions to generate or modify display information.

[0203] A display screen is used to display images, videos, etc. A display screen includes a display panel.

[0204] The aforementioned electronic device realizes the radar target detection method based on two feature fusion proposed in this application. By using linear dimensionality reduction technology to fuse phase linearity and relative peak height, and combining generalized extreme value distribution modeling and constant false alarm rate detection technology, it achieves the beneficial effects of improving radar target detection accuracy, enhancing robustness and adaptability, and ensuring stable detection performance.

[0205] The storage medium provided in this application stores a program product capable of implementing a radar target detection method based on two feature fusion.

[0206] Radar target detection methods based on two-feature fusion include:

[0207] Training data is constructed using existing radar target detection data. Based on the training data, phase linearity is extracted using the phase linearity extraction process, and the relative peak height of the Doppler spectrum is extracted using the relative peak height extraction process.

[0208] Based on the phase linearity and relative peak height, the fusion matrix of the two features is obtained using a linear dimensionality reduction method with compact distribution.

[0209] The reference range cell data and the range cell data to be detected are acquired in real time using a signal model. The phase linearity extraction process and the relative peak height extraction process are executed respectively, and the fusion feature is generated using the fusion matrix.

[0210] Based on the fusion features, the generalized extreme value distribution parameters are estimated and the detection statistics are constructed. An adaptive threshold is calculated by combining the preset false alarm probability. The target decision is completed by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression.

[0211] In some possible implementations, the radar target detection method based on two-feature fusion of this disclosure can be implemented as a program product including program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure.

[0212] The storage medium disclosed herein may be any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0213] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A radar target detection method based on two-feature fusion, characterized in that, include: Training data is constructed using existing radar target detection data. Phase linearity is extracted based on the training data extraction process, and the relative peak height of the Doppler spectrum is extracted using the relative peak height extraction process. Based on the phase linearity and relative peak height, the fusion matrix of the two features is obtained using a linear dimensionality reduction method with compact distribution. The reference distance cell data and the distance cell data to be detected are acquired in real time. The phase linearity extraction process and the relative peak height extraction process are executed respectively, and the fusion matrix is ​​used to generate fusion features. Based on the fusion features, the generalized extreme value distribution parameters are estimated and the detection statistics are constructed. An adaptive threshold is calculated by combining the preset false alarm probability. The target decision is completed by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression. The method of obtaining the fusion matrix of the two features based on the phase linearity and relative peak height using a linear dimensionality reduction method with distribution compactness includes: Based on the phase linearity and relative peak height, a feature dataset is generated. ; in, Let be the number of samples in the feature dataset, and be the th . Feature data samples Includes phase linearity features and relative peak height characteristics ,Right now: After linear dimensionality reduction, a fused feature dataset is constructed: set up , They represent the fusion features respectively. Given the expected value and variance, the kurtosis is calculated as follows: in, Indicates taking the expected value; set up The probability distribution function is The interquartile range is then expressed as: in, , They are respectively The lower quartile and upper quartile; The objective function is defined as follows: in, ; exist Traverse within the interval When the objective function When the minimum value is reached, the corresponding fusion matrix This is the optimal weighted vector for the two features PL and RPH.

2. The radar target detection method based on two-feature fusion according to claim 1, characterized in that, The phase linearity extraction process includes: The first in the training data indivual The entanglement phase of the data is represented as: Let the difference between adjacent winding phases be . Adjust it using the following formula. arrive Between, to eliminate Jump: The true phase after untangling is represented as: Let the phase difference between two adjacent phases of the true phase sequence be... Then the phase linearity is expressed as: in, Then the normalized phase linearity is: 。 3. The radar target detection method based on two-feature fusion according to claim 2, characterized in that, The extraction process for the relative peak height includes: Let the first distance unit be the first distance unit. indivual Data is The Doppler amplitude spectrum is then expressed by the following formula: in, For Doppler frequency, The pulse repetition period; Let the peak value of the Doppler amplitude spectrum be The relative peak height can then be expressed as: in, The extent of the Doppler sidelobe. for The number of Doppler frequency points within the range; The normalized relative peak height is: 。 4. The radar target detection method based on two-feature fusion according to claim 3, characterized in that, The real-time acquisition of reference distance cell data and distance cell data to be detected includes: The sea radar is set to operate in coherent processing mode, with the number of coherent pulse trains being [number missing]. After orthogonal demodulation and matched filtering, the I / Q data in the range cell to be detected is denoted as... The I / Q data in the reference range cell is denoted as ; in, Indicates the distance unit to be detected. The subscript represents the number of reference distance cells. Indicates the first One reference distance cell; The target I / Q data that may exist in the range cell to be detected are denoted as... The clutter I / Q data in the range cell to be detected and the reference range cell are respectively denoted as and .

5. The radar target detection method based on two-feature fusion according to claim 4, characterized in that, The estimation of generalized extreme value distribution parameters and construction of detection statistics based on fusion features include: The probability density function of the fused features is modeled using a generalized extreme value distribution, as follows: in It is a fusion of feature variables. For position parameters, For scale parameters, For shape parameters; The shape parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is: in, , , , ; from Obtained from each reference distance cell fusion feature samples The sequence is obtained by arranging the sequences in ascending order. ; The scale parameter is obtained by using the probabilistic weighted moment estimation method. The estimator expression is: The location parameters are obtained by using the probabilistic weighted moment estimation method. The estimator expression is: in, It is the Gamma function; Based on the form of the probability density function of the GEV distribution of the fused features under the condition of no target data, the following detection statistic is designed: When fusion features The number of samples tends to infinity, making , , At that time, the detection statistic was derived. The asymptotic probability density function is: 。 6. The radar target detection method based on two-feature fusion according to claim 5, characterized in that, The adaptive threshold calculation based on the preset false alarm probability includes: The false alarm probability is calculated using the following formula. Corresponding detection thresholds: 。 7. A radar target detection system based on two-feature fusion, characterized in that, The system employs the radar target detection method based on two-feature fusion as described in any one of claims 1 to 6; The system includes: The data extraction module is used to construct training data using existing radar target detection data, extract phase linearity based on the training data extraction process, and extract the relative peak height of the Doppler spectrum using the relative peak height extraction process. The fusion rule generation module is used to obtain the fusion matrix of the two features based on the phase linearity and relative peak height using a linear dimensionality reduction method with distribution compactness; The measured data fusion module is used to acquire reference distance cell data and distance cell data to be detected in real time, and to perform the extraction process of phase linearity and the extraction process of relative peak height respectively, and to generate fusion features using the fusion matrix; The target detection module is used to estimate the generalized extreme value distribution parameters and construct the detection statistics based on the fused features. It calculates the adaptive threshold by combining the preset false alarm probability and completes the target decision by comparing the detection statistics with the threshold, so as to achieve constant false alarm detection under clutter suppression.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the radar target detection method based on two-feature fusion as described in any one of claims 1 to 6.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the radar target detection method based on two-feature fusion as described in any one of claims 1 to 6.

Images