Target RCS time sequence feature extraction method and device

Abstract

The invention provides a target RCS time sequence feature extraction method and device. The method comprises the steps that target RCS time sequences of a target in different motion states are acquired; superposing the target RCS time sequence, and calculating to obtain a frequency value corresponding to a preset frequency cumulative contribution amount; determining a cepstrum frequency adopted by cepstrum analysis according to the frequency value; and obtaining a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency. According to the scheme, target RCS time sequence feature extraction in different states is achieved, and then the motion state of the target is recognized based on the target RCS time sequence.

Description

Technical Field

[0001] The invention relates to the technical field of radar target recognition, and in particular to a method and device for extracting target RCS time series features. Background Art

[0002] Since radar has all-day, all-weather, and penetrable detection capabilities, drone target recognition technology based on radar detection has become a research hotspot in recent years. In radar target recognition, the amount of information is related to the signal bandwidth. The larger the bandwidth, the more information. Although the amount of information in narrowband bandwidth is limited, it has a simple structure, low cost, and is easy to develop. Therefore, narrowband radar information is still an important research object for space target recognition. The radar cross section (RCS) of the target is the main information that narrowband radar can obtain from the target. RCS has rich information, but it is difficult to extract.

[0003] At present, there are many papers on the problem of UAV RCS feature extraction and recognition. In practice, the flight status of air and space targets will change with the mission, thus causing modulation of RCS. In addition, Matthew et al. published a paper in 2017 on IET Radar, Sonar &

[0004] An article published by Navigation mentioned that when classifying and identifying drones with payloads, it was found that there were obvious differences in the spectral characteristics of drones in flight and hovering, but no quantitative research was done. Therefore, since the electromagnetic scattering characteristics are jointly determined by the physical properties and attitude motion characteristics of the target, the dynamic RCS characteristics of the target under different motion states change differently, which is of great significance for target recognition applications in actual complex environments. Therefore, a method for extracting the time series features of the target RCS under different motion states is urgently needed. Summary of the invention

[0005] The present invention provides a method and device for extracting target RCS time series features, which realizes the extraction of target RCS time series features under different states, and further identifies the motion state of the target based on the target RCS time series.

[0006] In a first aspect, the present invention provides a method for extracting target RCS time series features, comprising:

[0007] Obtain the target RCS time series under different motion states;

[0008] The target RCS time series is superimposed to calculate the frequency value corresponding to the preset frequency cumulative contribution;

[0009] Determine the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0010] A cepstrum feature extraction parameter matrix of the target in each motion state is obtained according to the cepstrum frequency.

[0011] Optionally, the time series corresponding to the target RCS time series of the same target in different motion states are all the same and the sampling frequencies are all the same.

[0012] Optionally, the preset frequency cumulative contribution is 50%.

[0013] Optionally, the frequency value is determined by the following formula:

[0014] f b =(K' / K)·f s

[0015]

[0016] Wherein, K is the number of Fourier transforms of the target RCS time series; f s is the frequency value corresponding to the Kth Fourier transform in the target RCS time series; b is the frequency value, is the frequency value corresponding to the K'th Fourier transform in the target RCS time series; s(n) is the target RCS time series under the nth motion state; N is the number of motion state categories; FFT() represents Fourier transform; α is the cumulative contribution of the preset frequency.

[0017] Optionally, the cepstrum frequency is calculated by the following formula:

[0018]

[0019] Among them, f re is the cepstrum frequency; f b is the frequency value; f is the frequency value corresponding to the target RCS time series.

[0020] Optionally, obtaining a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency includes:

[0021] For each target RCS time series in the motion state, the following steps are performed:

[0022] The target RCS time series is subjected to preprocessing, frame processing and smoothing processing in sequence to obtain several frame sub-time series;

[0023] Performing discrete Fourier transform on the sub-time series to obtain spectrum energy distribution;

[0024] A triangular filter bank including a plurality of triangular filters is set according to the cepstrum frequency;

[0025] After filtering the spectrum energy distribution through the triangular filter group, calculating the logarithm of the energy output by each triangular filter, and then performing discrete cosine transformation to obtain the cepstrum coefficients of the sub-time series;

[0026] Generate a cepstral coefficient matrix under the motion state based on each of the sub-time series;

[0027] The first-order difference parameters of the cepstral coefficient matrix are extracted and combined to obtain the cepstral feature extraction parameter matrix under the motion state.

[0028] Optionally, the number of triangular filters in the triangular filter group is 22 to 26;

[0029] The frequency response of each triangular filter in the triangular filter bank is determined by the following formula:

[0030]

[0031]

[0032] Among them, H m (f re ) is the frequency response of the m-th triangular filter; f(m) is the cepstrum center frequency value corresponding to the m-th triangular filter; m∈[0,M-1], M is the number of the triangular filters.

[0033] In a second aspect, the present invention provides a target RCS time series feature extraction device, comprising:

[0034] An acquisition module is used to acquire the target RCS time series of the target in different motion states;

[0035] A cepstrum processing module is used to superimpose the target RCS time series, calculate the frequency value corresponding to the preset frequency cumulative contribution, and determine the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0036] A feature extraction module is used to obtain a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency.

[0037] In a third aspect, an embodiment of the present invention further provides a computing device, comprising a memory and a target processor, wherein the memory stores a computer program, and when the target processor executes the computer program, the method described in any first aspect of this specification is implemented.

[0038] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, enables the computer to execute the method described in any first aspect of this specification.

[0039] In a fifth aspect, an embodiment of the present invention further provides a computer program product, comprising computer instructions, which, when executed by a processor, implement the steps of the method described in any first aspect of this specification.

[0040] The present invention provides a method and device for extracting target RCS time series features, which obtains the target RCS time series under different motion states, and then determines the frequency value corresponding to the preset frequency cumulative contribution by superimposing all target RCS time series, and determines the cepstrum frequency during cepstrum analysis based on the frequency value, so as to extract the cepstrum feature parameter matrix under different motion states based on the cepstrum frequency. In this way, the features under different motion states are obtained through cepstrum analysis, and the motion state of the current radar target can be determined by identifying the features, thereby further inferring the mission intention of the radar target, and providing a theoretical basis for subsequent situation awareness. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 creative work.

[0042] Figure 1 is a flow chart of a method for extracting target RCS time series features provided by one embodiment of the present invention;

[0043] Figure 2 It is a target RCS time series and cepstrum feature extraction parameter matrix of a drone in a hovering state provided by an embodiment of the present invention;

[0044] Figure 3 It is a target RCS time series and cepstrum feature extraction parameter matrix of a drone in a self-rotating state provided by an embodiment of the present invention;

[0045] Figure 4 It is a target RCS time series and cepstrum feature extraction parameter matrix of a drone in a hovering state provided by an embodiment of the present invention;

[0046] Figure 5 It is a target RCS time series and cepstrum feature extraction parameter matrix of a UAV in a horizontal straight flight state provided by an embodiment of the present invention;

[0047] Figure 6 It is a target RCS time series and cepstrum feature extraction parameter matrix of a drone in a longitudinal straight flight state provided by an embodiment of the present invention;

[0048] Figure 7 is a hardware architecture diagram of a computing device provided by an embodiment of the present invention;

[0049] Figure 8 It is a structural schematic diagram of a target RCS time series feature extraction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0051] Since the existing narrowband statistical features cannot meet the needs of actual dynamic scenes, it is urgent to expand the feature dimension. Since the electromagnetic scattering characteristics are determined by the physical properties and posture motion characteristics of the target, analyzing the change law of the target dynamic RCS characteristics generated by different target motion states is of great significance for target recognition applications in actual complex environments.

[0052] Based on this, Figure 1 As shown, an embodiment of the present invention provides a method for extracting target RCS time series features, the method comprising:

[0053] Step 100, obtaining the target RCS time series of the target in different motion states;

[0054] Step 102, superimposing the target RCS time series, and calculating the frequency value corresponding to the preset frequency cumulative contribution;

[0055] Step 104, determining the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0056] Step 106: Obtain a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency.

[0057] In the present invention, the method obtains the target RCS time series under different motion states, and then determines the frequency value corresponding to the preset frequency cumulative contribution by superimposing all the target RCS time series, and determines the cepstrum frequency during cepstrum analysis based on the frequency value, so as to extract the cepstrum feature parameter matrix under different motion states based on the cepstrum frequency. In this way, the features under different motion states are obtained through cepstrum analysis, and the motion state of the current radar target can be determined by identifying the features, thereby further inferring the mission intention of the radar target, and providing a theoretical basis for subsequent situational awareness.

[0058] It should be noted that the target referred to in the present invention may be an aerospace target such as a drone.

[0059] With respect to step 100 , the target RCS time series of the same target in different motion states have the same time series and the same sampling frequency, so that the target RCS time series can be superimposed according to the time series.

[0060] For step 102, the target RCS time series is superimposed to calculate the frequency value corresponding to the preset frequency cumulative contribution, and the frequency value is determined by the following formula:

[0061] f b =(K' / K)·f s

[0062]

[0063] Wherein, K is the number of Fourier transforms of the target RCS time series; f s is the frequency value corresponding to the Kth Fourier transform in the target RCS time series; b is the frequency value, is the frequency value corresponding to the K'th Fourier transform in the target RCS time series; s(n) is the target RCS time series under the nth motion state; N is the number of motion state categories; FFT() represents Fourier transform; α is the cumulative contribution of the preset frequency.

[0064] It should be noted that 0<α<1. Although a plurality of K's can be determined based on the above formula, K' in the present invention refers to the minimum value that makes the above inequality hold.

[0065] In the present invention, the frequency value corresponding to the preset frequency cumulative contribution is used as a criterion for constructing the cepstrum frequency scale to obtain an improved cepstrum frequency, thereby further improving the accuracy of feature extraction and the effectiveness of the extracted feature information.

[0066] In a preferred embodiment, the preset frequency cumulative contribution is 50%, that is, α=50%.

[0067] In the present invention, the cepstrum feature extraction parameter matrix obtained by using the preset frequency cumulative contribution of 50% contains more and more comprehensive target motion information, so when the motion state is identified based on the cepstrum feature extraction parameter matrix, the recognition accuracy is higher.

[0068] In step 104, the cepstrum frequency used in the cepstrum analysis is determined according to the frequency value, and the cepstrum frequency is calculated by the following formula:

[0069]

[0070] Among them, f re is the cepstrum frequency; f b is the frequency value; f is the frequency value corresponding to the target RCS time series.

[0071] Specifically, the frequency value corresponding to the target RCS time series is the linear frequency, and step 104 converts the linear frequency into the cepstrum frequency used in the cepstrum analysis by using the above formula. For example, f is the first frequency value corresponding to the first Fourier change of the target RCS time series, then f re is the cepstrum frequency corresponding to the first frequency value after conversion; if the corresponding f is the Kth frequency value corresponding to the Kth Fourier change of the target RCS time series, then f re is the cepstrum frequency corresponding to the Kth frequency value after conversion.

[0072] With respect to step 106, obtaining a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency includes:

[0073] For each target RCS time series in the motion state, the following steps are performed:

[0074] The target RCS time series is subjected to preprocessing, frame processing and smoothing processing in sequence to obtain several frame sub-time series;

[0075] Performing discrete Fourier transform on the sub-time series to obtain spectrum energy distribution;

[0076] A triangular filter bank including a plurality of triangular filters is set according to the cepstrum frequency;

[0077] After filtering the spectrum energy distribution through the triangular filter group, calculating the logarithm of the energy output by each triangular filter, and then performing discrete cosine transformation to obtain the cepstrum coefficients of the sub-time series;

[0078] Generate a cepstral coefficient matrix under the motion state based on each of the sub-time series;

[0079] The first-order difference parameters of the cepstral coefficient matrix are extracted and combined to obtain the cepstral feature extraction parameter matrix under the motion state.

[0080] It should be noted that the preprocessing includes normalizing the target RCS time series, dividing the normalized target RCS time series into frames according to a preset frequency division step, and then performing smoothing to obtain a sub-time series, wherein the frequency band lengths of the sub-time series are the same.

[0081] Specifically, for each target RCS time series in n motion states, normalization and frame processing are performed to obtain I sub-time series; and each individual sub-time series is smoothed using a Hamming window W(l) to minimize the discontinuity of each frame signal and improve the frequency leakage. The expression of the Hamming window W(l) is as follows:

[0082]

[0083] Where L is the length of the sub-time series;

[0084] For the i-th (1≤i≤I) sub-time series, a discrete Fourier transform is performed to obtain the spectrum energy distribution; a set of triangular filter groups are set according to the improved cepstrum frequency, and the logarithm of the energy output by each triangular filter is calculated, and then a set of coefficients M(i) is obtained by discrete cosine transform, that is, the cepstrum coefficients of the i-th frame signal, and then the above steps are repeated to obtain the cepstrum coefficient matrix M corresponding to all I-frame sub-time series under the n-th motion state n ; Extract the cepstral coefficient matrix M n The first-order difference parameter ΔM n , and the cepstrum feature extraction parameter matrix M' of the target RCS time series under the nth motion state is obtained by merging n , M' n =[M n ΔM n ].

[0085] In the present invention, the target RCS time series is Fourier transformed by cepstrum analysis to obtain the spectrum energy distribution, and then the energy output after filtering is logarithmized and discrete cosine transformed to obtain cepstrum coefficients, which have a certain anti-interference ability to noise.

[0086] In a preferred embodiment, the number of triangular filters in the triangular filter bank is 22 to 26;

[0087] The frequency response of each triangular filter in the triangular filter bank is determined by the following formula:

[0088]

[0089]

[0090] Among them, H m (f re ) is the frequency response of the m-th triangular filter; f(m) is the cepstrum center frequency value corresponding to the m-th triangular filter; m∈[0,M-1], M is the number of triangular filters.

[0091] In the present invention, a triangular filter bank is used to smooth the spectrum and eliminate the effect of harmonics, highlighting the formant of the original speech. Since the tone or pitch of the audio is not present in the cepstral coefficient matrix parameters, the recognition system characterized by the cepstral coefficient matrix is ​​not affected by the tone of the input audio. The present invention preferably limits the number of triangular filters to 22 to 26, which can ensure that the triangular filter bank fully exerts its function and further reduce the amount of calculation.

[0092] In a specific embodiment, Figures 2 to 6 The target RCS time series and cepstral feature extraction parameter matrix of the drone in different motion states are shown. The target RCS time series length is 2.5s, the cepstral feature extraction parameter matrix has 21 frames and a dimension of 12, so the size of the parameter matrix M' is 24×21. Figures 2 to 6 It can be seen that the cepstrum feature extraction parameter matrices corresponding to different motion states are different. In order to quantitatively analyze the feature differences, the Pearson Correlation Coefficient (PCC) is selected as the measure, and the expression of this measure is

[0093]

[0094] Where Cov(·) represents the covariance operation, σ Mm , σ Mn Respectively represent M m 、M nThe standard deviation of , m is the mth motion state, and n is the nth motion state. The similarity calculation results obtained by using this expression are as follows: the similarity value of the same motion state is 1.0; the similarity value between the hovering state and the rotation state is 0.4029, the similarity value between the hovering state and the hovering state is 0.4369, the similarity value between the hovering state and the lateral direct flight state is 0.3427, the similarity value between the hovering state and the longitudinal direct flight state is 0.4601, the similarity value between the rotation state and the hovering state is 0.0645, the similarity value between the rotation state and the lateral direct flight state is 0.2431, the similarity value between the rotation state and the longitudinal direct flight state is 0.2659, the similarity value between the hovering state and the lateral direct flight state is 0.2249, the similarity value between the hovering state and the longitudinal direct flight state is 0.2466, and the similarity value between the lateral direct flight state and the longitudinal direct flight state is 0.2649. From the above calculation results, it can be seen that the maximum similarity of features in different motion states is 0.4601, and most features are weakly correlated (ρ<0.4), indicating that the improved cepstrum feature extraction parameter matrix can effectively characterize the motion law of the UAV.

[0095] It should be noted that Figures 2 to 6 The upper figures in the figure are schematic diagrams of the target RCS time series, and the vertical axis is the RCS value (dB); the lower figures are the cepstrum feature extraction parameter matrix (X-band HH polarization).

[0096] In a preferred embodiment, after obtaining the cepstrum feature extraction parameter matrix, the method further includes:

[0097] Obtain the RCS time series of the target to be detected in the current time period;

[0098] The target RCS time series is superimposed to calculate the frequency value corresponding to the preset frequency cumulative contribution;

[0099] Determine the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0100] Obtaining a cepstrum feature extraction parameter matrix of the target to be detected according to the cepstrum frequency;

[0101] The similarity between the cepstrum feature extraction parameter matrix to be detected and the cepstrum feature extraction parameter matrix of the current known motion state is calculated respectively, and the similarity value is output; wherein the current known motion state includes a hovering state, a rotation state, a circling state, a lateral direct flight state, and a longitudinal direct flight state;

[0102] Determine whether there is a similarity value greater than a preset threshold;

[0103] If so, the current known motion state corresponding to the similarity value is determined as the current motion state of the target in the current time period.

[0104] As described in the previous example, the preset threshold value ranges from (0.46, 1], preferably 0.7, 0.8, etc. Specifically, any of the above-mentioned target RCS time series feature extraction methods is used to obtain the above-mentioned cepstrum feature extraction parameter matrix to be detected.

[0105] In the present invention, since there are differences in the cepstral feature extraction parameter matrices under different motion states, by comparing the similarity between the cepstral feature extraction parameter matrix to be detected and the cepstral feature extraction parameter matrix of the known motion state, the current motion state corresponding to the cepstral feature extraction parameter matrix to be detected can be obtained, and the recognition of the target motion state based on the target RCS time series can be realized, which is of great significance for target recognition applications in actual complex environments.

[0106] like Figure 7 , Figure 8 As shown, an embodiment of the present invention provides a target RCS time series feature extraction device. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. From the hardware level, Figure 7 As shown in FIG. 1 , a hardware architecture diagram of a computing device where a target RCS time series feature extraction device is provided in an embodiment of the present invention is provided. Figure 7 In addition to the processor, memory, network interface, and non-volatile memory shown in the figure, the computing device in which the device is located in the embodiment may also generally include other hardware, such as a forwarding chip responsible for processing messages, etc. Taking software implementation as an example, Figure 8 As shown, as a device in a logical sense, the CPU of the computing device in which it is located reads the corresponding computer program in the non-volatile memory into the memory and runs it. This embodiment provides a target RCS time series feature extraction device, which includes:

[0107] An acquisition module 800 is used to acquire a target RCS time series of a target in different motion states;

[0108] The cepstrum processing module 802 is used to superimpose the target RCS time series, calculate the frequency value corresponding to the preset frequency cumulative contribution, and determine the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0109] The feature extraction module 804 is used to obtain a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency.

[0110] In some specific implementations, the acquisition module 800 may be used to execute the above step 100 , the cepstrum processing module 802 may be used to execute the above steps 102 and 104 , and the feature extraction module 804 may be used to execute the above step 106 .

[0111] In one embodiment of the present invention, the time series corresponding to the target RCS time series of the same target in different motion states are the same and the sampling frequency is the same.

[0112] In one embodiment of the present invention, the cepstrum processing module 302 is further configured to perform the following operations:

[0113] The frequency value is determined by the following formula:

[0114] f b =(K' / K)·f s

[0115]

[0116] Wherein, K is the number of Fourier transforms of the target RCS time series; f s is the frequency value corresponding to the Kth Fourier transform in the target RCS time series; b is the frequency value, is the frequency value corresponding to the K'th Fourier transform in the target RCS time series; s(n) is the target RCS time series under the nth motion state; N is the number of motion state categories; FFT() represents Fourier transform; α is the cumulative contribution of the preset frequency, α=50%.

[0117] In one embodiment of the present invention, the cepstrum processing module 302 is further configured to perform the following operations:

[0118] The cepstrum frequency is determined by the following formula:

[0119]

[0120] Among them, f re is the cepstrum frequency; f b is the frequency value; f is the frequency value corresponding to the target RCS time series.

[0121] In one embodiment of the present invention, the feature extraction module 304 is further configured to perform the following operations:

[0122] For each target RCS time series in the motion state, the following steps are performed:

[0123] The target RCS time series is subjected to preprocessing, frame processing and smoothing processing in sequence to obtain several frame sub-time series;

[0124] Performing discrete Fourier transform on the sub-time series to obtain spectrum energy distribution;

[0125] A triangular filter group including a plurality of triangular filters is set according to the cepstrum frequency; the number of triangular filters in the triangular filter group is 22 to 26;

[0126] After filtering the spectrum energy distribution through the triangular filter group, calculating the logarithm of the energy output by each triangular filter, and then performing discrete cosine transformation to obtain the cepstrum coefficients of the sub-time series;

[0127] Generate a cepstral coefficient matrix under the motion state based on each of the sub-time series;

[0128] The first-order difference parameters of the cepstral coefficient matrix are extracted and combined to obtain the cepstral feature extraction parameter matrix under the motion state.

[0129] In one embodiment of the present invention, an identification module is further included, and the identification module is used to perform the following operations:

[0130] Obtain the RCS time series of the target to be detected in the current time period;

[0131] The target RCS time series is superimposed to calculate the frequency value corresponding to the preset frequency cumulative contribution;

[0132] Determine the cepstrum frequency used in the cepstrum analysis according to the frequency value;

[0133] Obtaining a cepstrum feature extraction parameter matrix of the target to be detected according to the cepstrum frequency;

[0134] The similarity between the cepstrum feature extraction parameter matrix to be detected and the cepstrum feature extraction parameter matrix of the current known motion state is calculated respectively, and the similarity value is output; wherein the current known motion state includes a hovering state, a rotation state, a circling state, a lateral direct flight state, and a longitudinal direct flight state;

[0135] Determine whether there is a similarity value greater than a preset threshold;

[0136] If so, the current known motion state corresponding to the similarity value is determined as the current motion state of the target in the current time period.

[0137] It is understood that the structure illustrated in the embodiment of the present invention does not constitute a specific limitation on a target RCS time series feature extraction device. In other embodiments of the present invention, a target RCS time series feature extraction device may include more or fewer components than shown in the figure, or combine some components, or split some components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0138] The information interaction, execution process and other contents between the modules in the above-mentioned device are based on the same concept as the embodiment of the method of the present invention. For the specific contents, please refer to the description in the embodiment of the method of the present invention, and no further description is given here.

[0139] An embodiment of the present invention further provides a computing device, including a memory and a target processor, wherein a computer program is stored in the memory, and when the target processor executes the computer program, a target RCS time series feature extraction method in any embodiment of the present invention is implemented.

[0140] An embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the processor executes a target RCS time series feature extraction method in any embodiment of the present invention.

[0141] An embodiment of the present application also provides a computer program product, which includes a computer program. A processor of a computer device reads the computer program from a computer-readable storage medium, and the processor executes the computer program, so that the computer device executes a target RCS time series feature extraction method described in any of the above embodiments.

[0142] Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above-mentioned embodiments is stored, and a computer (or CPU or MPU) of the system or device can be enabled to read and execute the program code stored in the storage medium.

[0143] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0144] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0145] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling the operating system on the computer to complete part or all of the actual operations based on the instructions of the program code.

[0146] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or expansion module is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0147] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "comprise a ..." do not exclude the presence of other identical factors in the process, method, article or device including the elements.

[0148] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk, etc., various media that can store program codes.

[0149] 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 embodiments of the present invention.

Claims

1. A target RCS time series feature extraction method, characterized in that: include: Obtain the target RCS time series under different motion states; The target RCS time series is superimposed to calculate the frequency value corresponding to the preset frequency cumulative contribution; Determine the cepstrum frequency used in the cepstrum analysis according to the frequency value; A cepstrum feature extraction parameter matrix of the target in each motion state is obtained according to the cepstrum frequency.

2. The method according to claim 1, characterized in that The time series corresponding to the target RCS time series of the same target in different motion states are the same and the sampling frequency is the same; and / or, The preset frequency cumulative contribution is 50%.

3. The method according to claim 1, characterized in that: The frequency value is determined by the following formula: f b =(K' / K)·f s Wherein, K is the number of Fourier transforms of the target RCS time series; f s is the frequency value corresponding to the Kth Fourier transform in the target RCS time series; b is the frequency value, is the frequency value corresponding to the K'th Fourier transform in the target RCS time series; s(n) is the target RCS time series under the nth motion state; N is the number of motion state categories; FFT() represents Fourier transform; α is the cumulative contribution of the preset frequency.

4. The method according to claim 1, characterized in that The cepstrum frequency is calculated by the following formula: Among them, f re is the cepstrum frequency; f b is the frequency value; f is the frequency value corresponding to the target RCS time series.

5. The method according to any one of claims 1 to 4, characterized in that: The step of obtaining a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency includes: For each target RCS time series in the motion state, the following steps are performed: The target RCS time series is subjected to preprocessing, frame processing and smoothing processing in sequence to obtain several frame sub-time series; Performing discrete Fourier transform on the sub-time series to obtain spectrum energy distribution; A triangular filter bank including a plurality of triangular filters is set according to the cepstrum frequency; After filtering the spectrum energy distribution through the triangular filter group, calculating the logarithm of the energy output by each triangular filter, and then performing discrete cosine transformation to obtain the cepstrum coefficients of the sub-time series; Generate a cepstral coefficient matrix under the motion state based on each of the sub-time series; The first-order difference parameters of the cepstral coefficient matrix are extracted and combined to obtain the cepstral feature extraction parameter matrix under the motion state.

6. The method according to claim 5, characterized in that The number of triangular filters in the triangular filter group is 22 to 26; The frequency response of each triangular filter in the triangular filter bank is determined by the following formula: Among them, H m (f re ) is the frequency response of the m-th triangular filter; f(m) is the cepstrum center frequency value corresponding to the m-th triangular filter; m∈[0,M-1], M is the number of the triangular filters.

7. A target RCS time series feature extraction device, characterized in that: include: An acquisition module is used to acquire the target RCS time series of the target in different motion states; A cepstrum processing module is used to superimpose the target RCS time series, calculate the frequency value corresponding to the preset frequency cumulative contribution, and determine the cepstrum frequency used in the cepstrum analysis according to the frequency value; A feature extraction module is used to obtain a cepstrum feature extraction parameter matrix of the target in each motion state according to the cepstrum frequency.

8. A computing device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to execute the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions, which, when executed by a processor, implement the steps of the method according to any one of claims 1 to 6.

Images