Anti-unmanned aerial vehicle method and system based on sound detection

Abstract

The invention relates to the technical field of sound wave detection, in particular to an anti-unmanned aerial vehicle method and system based on sound detection, and the method comprises the steps: obtaining the frequency data and phase data of a sound wave signal in each short-time window; obtaining a distribution characteristic index of the phase difference according to the phase difference between the same frequency data in each short-time window and the adjacent short-time window; obtaining a distortion characteristic index of the phase difference according to the phase difference and the amplitude difference between the same frequency data in each short-time window and the adjacent short-time window; clustering the phase difference to obtain a cluster, and obtaining an interval signal segment; according to the frequency information in each interval signal segment and the frequency data in each cluster, obtaining the matching degree of the interval signal segment and the cluster; and the ITD decomposition algorithm is optimized, signal reconstruction is carried out to obtain an optimal sound wave signal, and the optimal sound wave signal is used for obtaining unmanned aerial vehicle positioning. The method is high in signal processing precision, and is effectively used for unmanned aerial vehicle sound wave identification, positioning and interception.

Description

Technical Field

[0001] The present invention relates to the technical field of acoustic wave detection, and in particular to an anti-UAV method and system based on acoustic detection. Background Art

[0002] In today's information and intelligent society, drones, as a type of aircraft that does not require manned operation, have shown great potential and value in many fields such as military, civil, and scientific research in recent years. The anti-drone method based on acoustic detection uses the propagation characteristics of sound waves in the air to detect, identify, and track drones by transmitting and receiving sound wave signals of specific frequencies. However, accurate sound wave signals need to be detected in this process. Since the detection sound waves of drones propagate longitudinally, there will be obvious differences in air density, wind disturbance, temperature and humidity in the longitudinal space, which may cause irregular attenuation and deformation during the propagation of sound waves, greatly reducing its detection accuracy. The existing ITD algorithm is used to decompose and reconstruct the sound wave signal, which can remove a certain degree of noise to a certain extent. However, for the irregular distortion that occurs during the propagation of sound waves, the ITD feature decomposition cannot obtain a more accurate signal decomposition result, which makes the accuracy of the reconstructed signal low. Summary of the invention

[0003] In order to solve the technical problem that the existing ITD feature decomposition cannot obtain a more accurate signal decomposition result for the irregular distortion that occurs during the propagation of sound waves, the purpose of the present invention is to provide an anti-UAV method and system based on acoustic detection. The technical solutions adopted are as follows:

[0004] In a first aspect, the present invention provides an anti-UAV method based on acoustic detection, the method comprising:

[0005] When the UAV flies over the monitoring area, the UAV's acoustic wave signal is collected, and the frequency data and phase data of the acoustic wave signal in each short-time window are obtained;

[0006] According to the distribution of the phase data difference between each short-time window and the same frequency data in the adjacent short-time window, the distribution characteristic indexes corresponding to different types of phase differences are obtained;

[0007] According to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, the distribution of the difference change trend is analyzed to obtain the distortion characteristic index corresponding to different types of phase differences;

[0008] The phase differences are clustered according to the distribution characteristic index and the distortion characteristic index to obtain clusters, and the sound wave signal is divided into interval signal segments using the ITD decomposition algorithm; the matching degree between each interval signal segment and each cluster is obtained according to the distribution of the frequency information of the sound wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster;

[0009] The ITD decomposition algorithm is optimized using the matching degree, and signal reconstruction is performed according to the signal decomposition result to obtain a preferred sound wave signal, and the preferred sound wave signal is used to obtain the positioning of the UAV.

[0010] Preferably, the distribution characteristic indicators corresponding to different types of phase differences are obtained according to the distribution of the phase data difference between each short-time window and the same frequency data in the adjacent short-time window, specifically including:

[0011] For each short-time window and the next adjacent short-time window, each phase difference data is determined based on the difference between the phase data corresponding to the same frequency data; wherein the same phase difference data is the same type of phase difference;

[0012] According to the distribution of the number of time series features of each type of phase difference, a distribution feature index corresponding to each type of phase difference is obtained.

[0013] Preferably, obtaining the distribution characteristic index corresponding to each type of phase difference according to the distribution of the number of time series characteristics of each type of phase difference specifically includes:

[0014] The total number of short-time windows in which the same type of phase difference appears continuously is recorded as the number of time series features of each type of phase difference;

[0015] Based on the ratio between the maximum value of the number of time series features of each type of phase difference and the mean value of the number of time series features of all other types of phase differences, the distribution feature index corresponding to each type of phase difference is determined.

[0016] Preferably, the method analyzes the distribution of the difference change trend according to the difference in phase data and amplitude difference between the same frequency data in each short-time window and the adjacent short-time window to obtain the distortion characteristic indicators corresponding to different types of phase differences, specifically including:

[0017] The short-time window where the phase difference of each type is obtained is recorded as the characteristic short-time window;

[0018] According to the amplitude difference between the frequency data corresponding to the characteristic short-time window and the adjacent characteristic short-time window under the same type of phase difference, combined with the corresponding phase difference data, the characteristic difference value corresponding to each phase difference data under the same type is obtained;

[0019] For any type of phase difference, fitting is performed based on the characteristic difference value corresponding to each phase difference data to determine the characteristic difference function of the phase difference of the any type;

[0020] According to the change trend of each data point in the characteristic difference function of each type of phase difference, the distortion characteristic index corresponding to each type of phase difference is obtained.

[0021] Preferably, the amplitude difference between the frequency data corresponding to the characteristic short-time window and the adjacent characteristic short-time window under the same type of phase difference is combined with the corresponding phase difference data to obtain the characteristic difference value corresponding to each phase difference data under the same type, specifically including:

[0022] For any type of phase difference, the amplitude difference data of each phase difference data between the amplitudes of the corresponding frequency data in two adjacent characteristic short-time windows corresponding to each phase difference data is calculated;

[0023] The Euclidean norm of each phase difference data and the corresponding amplitude difference data is used as the feature difference value of each two adjacent feature short-time windows; the feature difference values ​​of the two adjacent feature short-time windows correspond to the feature difference value corresponding to one phase difference data.

[0024] Preferably, obtaining the distortion characteristic index corresponding to each type of phase difference according to the change trend of each data point in the characteristic difference function of each type of phase difference specifically includes:

[0025] For any type of phase difference, the slope value of each data point is calculated based on the characteristic difference function, and the mean of the slope values ​​corresponding to all data points on the characteristic difference function is negatively normalized to obtain the distortion characteristic index corresponding to the phase difference of this type.

[0026] Preferably, clustering the phase differences according to the distribution characteristic index and the distortion characteristic index to obtain cluster clusters specifically includes:

[0027] The Euclidean norm between the distribution characteristic index and the distortion characteristic index corresponding to each type of phase difference is used as the characteristic measurement parameter of each type of phase difference;

[0028] Based on the difference distance between the characteristic measurement parameters of different types of phase differences, each type of phase difference is clustered to obtain a cluster cluster.

[0029] Preferably, the matching degree between each interval signal segment and each cluster cluster is obtained according to the distribution of the frequency information of the sound wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster cluster, specifically including:

[0030] Record any interval signal segment as a selected interval signal segment, and record any cluster as a selected cluster;

[0031] Calculate the mean value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment to obtain the first eigenvalue of the selected interval signal segment; calculate the entropy value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment to obtain the second eigenvalue of the selected interval signal segment;

[0032] Calculating the mean value of the frequency data corresponding to all phase differences in the selected cluster to obtain the first eigenvalue of the selected cluster; calculating the entropy value of the frequency data corresponding to all phase differences in the selected cluster to obtain the second eigenvalue of the selected cluster;

[0033] The absolute value of the difference between the first eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the first difference, and the absolute value of the difference between the second eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the second difference; the product of the first difference and the second difference is negatively correlated normalized to obtain the matching degree between the selected interval signal segment and the selected cluster cluster.

[0034] Preferably, the optimizing the ITD decomposition algorithm by using the matching degree specifically includes:

[0035] For any cluster, the average value of the amplitude data corresponding to all phase differences in the cluster is calculated to obtain the amplitude characteristic value of the cluster;

[0036] Arrange all clusters in order from small to large according to the amplitude characteristic value of each cluster, decompose and process the acoustic wave signal using the ITD decomposition algorithm, and in the decomposition process of each interval signal segment, adjust the original baseline function in turn according to the arrangement order using the matching degree between the interval signal segment and each cluster, and obtain the modified baseline function of each decomposition of the interval signal segment, wherein the adjustment is a product operation;

[0037] Based on the modified baseline function and the original sound wave signal, combined with the ITD algorithm principle, an ITD decomposition result is obtained.

[0038] In a second aspect, the present invention provides an anti-UAV system based on acoustic detection, which is used to implement the steps of an anti-UAV method based on acoustic detection. The anti-UAV system based on acoustic detection includes:

[0039] A data acquisition module is used to collect the acoustic wave signal of the drone when the drone flies over the monitoring area, and obtain the frequency data and phase data of the acoustic wave signal in each short-time window;

[0040] A distribution characteristic analysis module, used to obtain distribution characteristic indicators corresponding to different types of phase differences according to the distribution of phase data differences between each short-time window and the same frequency data in adjacent short-time windows;

[0041] The distortion feature analysis module is used to analyze the distribution of the difference change trend according to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, and obtain the distortion feature index corresponding to different types of phase differences;

[0042] A feature matching module is used to cluster the phase differences according to the distribution feature index and the distortion feature index to obtain clusters, and to divide the acoustic wave signal by using the ITD decomposition algorithm to obtain interval signal segments; and to obtain the matching degree between each interval signal segment and each cluster according to the distribution of the frequency information of the acoustic wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster;

[0043] The signal optimization module is used to optimize the ITD decomposition algorithm using the matching degree, and to reconstruct the signal according to the signal decomposition result to obtain the optimal sound wave signal, and the optimal sound wave signal is used to obtain the UAV positioning.

[0044] The embodiments of the present invention have at least the following beneficial effects:

[0045] The present invention first performs time-frequency analysis on the collected original sound wave signal to obtain frequency data and phase data in a short-time window, providing a data basis for the subsequent analysis of the phase data performance when the same frequency has a signal delay phenomenon. Then, on the one hand, by analyzing the phase difference distribution of the same frequency data, the signal delay in adjacent short-time windows is extracted, that is, the distribution of phase differences, reflecting the representativeness of each phase difference to a certain type of signal component; on the other hand, by analyzing the phase difference and amplitude difference of the same frequency, and analyzing the trend of the difference change, the distortion characteristics of different delayed signal components in the timing can be reflected, and the distortion characteristic index of each phase difference is calculated to describe the degree to which different delayed signal components are affected by the same distortion mechanism. Furthermore, all types of phase differences are classified in combination with distribution characteristics and distortion characteristics, and signal components affected by the same distortion mechanism are divided into the same cluster. The number of types of distortion mechanisms can be quantified, and then the degree of feature matching between the distortion mechanism signal corresponding to each cluster and the interval signal segment to be decomposed is analyzed. This can guide the subsequent process of decomposing characteristic components using the ITD decomposition algorithm, so that the optimized ITD algorithm can express the distortion information more accurately when decomposing the IMF component, so that the preferred sound wave signal reconstructed using the signal decomposition structure can be clearer and more accurate, minimizing information loss and achieving high signal processing accuracy, which can be effectively used in subsequent drone sound wave identification, positioning and interception. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only 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.

[0047] Figure 1 It is a flowchart of the steps of an anti-UAV method based on acoustic detection provided by the present invention;

[0048] Figure 2 It is a flowchart of the steps of the method for obtaining distribution characteristic indicators corresponding to different types of phase differences provided by the present invention;

[0049] Figure 3 It is a schematic diagram of the corresponding distribution of two adjacent short-time window sound wave signals and phase spectra provided by the present invention;

[0050] Figure 4 It is a flowchart of the steps of the method for obtaining the distortion characteristic index corresponding to each type of phase difference provided by the present invention;

[0051] Figure 5It is a schematic diagram of the corresponding relationship between the sound wave signal and the short-time window provided by the present invention;

[0052] Figure 6 It is a partial distribution diagram of the characteristic difference function provided by the present invention;

[0053] Figure 7 It is a structural block diagram of an anti-UAV system based on acoustic detection provided by the present invention;

[0054] Figure 8 It is a structural schematic diagram of a computer device provided by the present invention. DETAILED DESCRIPTION

[0055] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation, structure, features and effects of an anti-UAV method and system based on acoustic detection proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0056] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0057] The following is a detailed description of a specific solution of an anti-UAV method and system based on acoustic detection provided by the present invention in conjunction with the accompanying drawings.

[0058] See also Figure 1 , which shows a flowchart of a method for countering UAV based on acoustic detection provided by an embodiment of the present invention, the method comprising the following steps:

[0059] Step S100, when the drone flies over the monitoring area, the acoustic wave signal of the drone is collected, and the frequency data and phase data of the acoustic wave signal in each short-time window are obtained.

[0060] When a drone flies over the monitoring area, the rapid rotation of its rotors will generate significant noise. This is the main source of the drone's acoustic signal. Other noises such as motor noise, aerodynamic noise, mechanical vibration, and electronic equipment noise will also be collected by the acoustic detector at the same time.

[0061] When the anti-UAV system actually performs operations, it needs to face complex meteorological conditions. Its sound wave propagation process will be directly affected by environmental factors, especially the air density, wind disturbance, temperature and humidity in different altitudes. There will be obvious differences, which may cause the sound wave signal received at the sound wave detector to be severely distorted. Therefore, how to extract accurate UAV sound wave signals from the sound wave detector is one of the core issues to improve the accuracy of UAV identification and positioning.

[0062] When sound waves pass through air layers with different characteristics, they may produce certain distortions. The distortions in the sound wave signals received by the sound wave detector are often formed multiple times; therefore, the source of the distorted signal lacks explainability, that is, it is impossible to determine the source of the distorted signal, and it is difficult to estimate the real sound wave signal under the distorted signal, as well as the drone sound wave signal that may be contained therein; however, the distorted signals formed multiple times may also leave some traces. Some signal components may have similar characteristics after experiencing multiple distortions at the same time, indicating that the signal propagation environment is uniform and the distortion mechanism is repeated; another part of the signal components have greater characteristic differences after experiencing multiple distortions, which indicates that the signal propagation environment is complex and the distortion mechanism is diversified.

[0063] Based on this, in order to analyze the characteristic distribution of the signal delay of the drone's sound wave signal and whether the signal has undergone a certain degree of distortion, it is first necessary to perform time-frequency analysis on the sound wave signal.

[0064] Specifically, in this embodiment, by deploying an acoustic wave detector in the test area, the acoustic wave detector is used to collect the acoustic wave signal of the drone, that is, a simulated flight of the drone is performed in the test area to generate the acoustic wave signal of the drone, and when the drone flies over the monitoring area, the deployed acoustic wave detector can receive the acoustic wave signal generated by the drone. Among them, the acoustic wave detector is generally composed of multiple microphones or acoustic wave sensors. In this embodiment, a hexagonal microphone array is used to form a coverage network to ensure that the spatial area that needs to be monitored is covered.

[0065] Further, the sound wave signal of the drone is subjected to a short-time Fourier transform to obtain the amplitude spectrum and phase spectrum of the sound wave signal in each short-time window. In this embodiment, the short-time window is set to 0.5 seconds, and the implementer can set it according to the specific implementation scenario. In this embodiment, after the sound wave signal is subjected to a short-time Fourier transform, the frequency data in each short-time window, the amplitude data under each frequency data, and the phase data under each frequency can be obtained.

[0066] Step S200, obtaining distribution characteristic indicators corresponding to different types of phase differences according to the distribution of phase data differences between each short-time window and the same frequency data in adjacent short-time windows.

[0067] It should be noted that when air, as a medium, propagates sound waves, it can be approximately considered as a linear time-invariant system. When different frequency components of sound waves propagate in the air, they will have different phase and amplitude changes due to air conditions. Because the phase is related to frequency, propagation time, etc., sound waves of different frequencies have different phase changes at the same propagation time; at the same time, the absorption and scattering characteristics of air for sound waves of different frequencies are different, resulting in different amplitude changes.

[0068] Group delay is the envelope delay of different frequency components in a signal. In the air system, group delay is the same for all frequency components. This is determined by the physical properties of air. Although the phase and amplitude of each frequency component change individually, as a whole, their time delay is the same. This also means that even if all frequency components are delayed by the same amount in the time domain, their relative relationship does not change and they can still be regarded as components of the original signal. Therefore, the signal delay phenomenon of a certain type of signal component in adjacent short-time windows can be characterized by analyzing the difference in phase, and the frequency components corresponding to the same phase difference are likely to represent the same type of signal components.

[0069] Based on this, this embodiment first analyzes the phase difference changes of the same frequency components in adjacent short-time windows, and then analyzes the distribution characteristics of the phase difference changes in the first aspect, and analyzes the distortion characteristics of the phase difference changes in the second aspect.

[0070] In this embodiment, if Figure 2 As shown, the method for obtaining distribution characteristic indicators corresponding to different types of phase differences can be implemented by step S201 and step S202.

[0071] Step S201, for each short-time window and the next adjacent short-time window, each phase difference data is determined based on the difference between the phase data corresponding to the same frequency data; wherein the same phase difference data is the same type of phase difference.

[0072] Specifically, Figure 3The figure is a schematic diagram of the corresponding distribution of the acoustic wave signal and the phase spectrum of two adjacent short-time windows. The frequency data corresponding to the first short-time window on the left side of the figure includes four types: X1, X2, X3 and X4, and the frequency data corresponding to the acoustic wave signal in the second short-time window on the right side includes four types: X1, X2, X5 and X6. The frequency data of the same type in the two adjacent short-time windows are X1 and X2. Then, between the two adjacent short-time windows, the difference in the phase difference corresponding to the frequency data X1 and the difference in the phase difference corresponding to the frequency data X2 can be calculated. That is, for the two adjacent short-time windows, the phase difference between the phase data of the frequency data X1 in the first short-time window and the phase data of the frequency data X1 in the second short-time window can be calculated to obtain a phase difference data of the frequency data X1. Based on the same reason, the phase difference between the phase data of the frequency data X2 in the first short-time window and the phase data of the frequency data X2 in the second short-time window can be calculated to obtain a phase difference data of the frequency data X2.

[0073] It can be understood that the same value of the phase difference data refers to the same type of phase difference, and one phase difference data corresponds to two adjacent short-time windows and a frequency data with the same value, but the phase data and amplitude data of the frequency data with the same value in two adjacent short-time windows may be different.

[0074] Step S202 , obtaining a distribution characteristic index corresponding to each type of phase difference according to the distribution of the number of time series characteristics of each type of phase difference.

[0075] The phase difference data corresponding to every two adjacent short-time windows reflects the signal delay of the same type of frequency data in these two adjacent short-time windows. If the same type of phase difference data is continuously distributed in time series, it means that the continuous signal delay may belong to the same type of distortion mechanism. The longer the continuous distribution time is, the greater the influence of this type of distortion mechanism on the delay of the sound wave signal.

[0076] Specifically, the total number of short-time windows in which the same type of phase difference appears continuously is obtained and recorded as the number of time series features of each type of phase difference. For example, the a-th type of phase difference contains three phase difference data in consecutive adjacent short-time windows, that is, the first and second short-time windows correspond to the first phase difference data of the a-th type of phase difference, the second and third short-time windows correspond to the second phase difference data of the a-th type, and the third and fourth short-time windows correspond to the third phase difference data of the a-th type of phase difference. Therefore, the value of the number of time series features of the a-th type of phase difference is 4, indicating that the same type of phase difference appears continuously in four consecutive adjacent short-time windows.

[0077] Further, based on the ratio between the maximum number of timing characteristics of each type of phase difference and the mean number of timing characteristics of all other types of phase differences, the distribution characteristic index corresponding to each type of phase difference is determined. As a specific example, the calculation formula for the distribution characteristic index corresponding to the vth type of phase difference can be expressed as:

[0078]

[0079] Among them, Q v represents the distribution characteristic index corresponding to the phase difference of the vth type, D v represents the number of time series features of the phase difference of the vth type, Max represents the maximum value function, Max(D v ) represents the maximum value of the number of time series features of the phase difference of the vth type, μ(D′ v ) represents the mean of the number of temporal features of phase differences of all categories except the vth category, D′ v It represents the number of time series features of phase differences of other types except the vth type, and Norm represents the normalization function, which is used to normalize the comparison value.

[0080] The larger the value of , the more continuous the distribution of the phase difference of the vth type in time series, and the smaller the time series continuity of the phase differences of other types, which further indicates that the acoustic signal interval corresponding to the phase difference of the vth type is continuously affected by the same type of distortion mechanism, and the degree of influence is higher.

[0081] The distribution characteristic index corresponding to each type of phase difference characterizes the signal delay situation and the degree of influence of the distortion mechanism on the acoustic signal represented by each type of phase difference.

[0082] Step S300 , analyzing the distribution of the difference change trend according to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, and obtaining distortion characteristic indicators corresponding to different types of phase differences.

[0083] Under normal circumstances, the frequency components corresponding to the same phase difference are more likely to belong to the same type of signal components, because in an ideal linear time-invariant system, if they are the same type of original signal components, their amplitude and phase changes between adjacent short-time windows should have certain regularity. Therefore, when the amplitude and phase difference changes of the frequency data corresponding to the same phase difference are more complex, such signal components are more likely to be distorted signals. Therefore, by analyzing the changing trends and richness of the amplitude differences and phase differences of the frequency data corresponding to the same type of phase differences in adjacent short-time windows, the degree of distortion characteristics of the sound wave signals corresponding to the same type of phase differences can be quantitatively analyzed.

[0084] In this embodiment, if Figure 4 As shown, the method for obtaining the distortion characteristic index corresponding to each type of phase difference can be implemented by steps S301 to S304.

[0085] Step S301, obtaining a short-time window where each type of phase difference is located and recording it as a characteristic short-time window.

[0086] like Figure 5 It is a schematic diagram of the corresponding relationship between the sound wave signal and the short-time window. Figure 5 The middle gray short-time window is the short-time window where the phase difference of the vth type is located, that is, the characteristic short-time window corresponding to the phase difference of the vth type. The white short-time window is the short-time window where there is no phase difference of the vth type. Figure 5 It can be seen that the characteristic short-time window has continuous and adjacent distribution features, and also has discontinuous situations.

[0087] Step S302, according to the amplitude difference between the frequency data corresponding to the characteristic short-time window and the adjacent characteristic short-time window under the same type of phase difference, combined with the corresponding phase difference data, obtain the characteristic difference value corresponding to each phase difference data under the same type.

[0088] In order to provide a data basis for the subsequent analysis of the distortion characteristics of the same type of signal, it is first necessary to comprehensively characterize the corresponding differences in amplitude and phase of the same type of signal delay, and quantify the signal parameter differences corresponding to the current signal delay.

[0089] Specifically, for any type of phase difference, the amplitude difference data of each phase difference data between the amplitudes of the corresponding frequency data in two adjacent characteristic short-time windows corresponding to each phase difference data is calculated; the Euclidean norm of each phase difference data and the corresponding amplitude difference data is used as the characteristic difference value of each two adjacent characteristic short-time windows; the characteristic difference values ​​of the two adjacent characteristic short-time windows correspond to the characteristic difference value corresponding to one phase difference data.

[0090] More specifically, for the vth type of phase difference, the calculation formula for the characteristic difference value of two adjacent characteristic short-time windows corresponding to the i-th phase difference data can be expressed as:

[0091]

[0092] Among them, Z v,i It represents the characteristic difference value of two adjacent characteristic short-time windows corresponding to the i-th phase difference data under the v-th type of phase difference, and A1 represents the phase data of the first characteristic short-time window and the phase data of the second characteristic short-time window in two adjacent characteristic short-time windows corresponding to the i-th phase difference data under the v-th type of phase difference respectively; v,i and A2 v,i They respectively represent the amplitude data of the first characteristic short-time window and the amplitude data of the second characteristic short-time window in the two adjacent characteristic short-time windows corresponding to the i-th phase difference data under the v-th type of phase difference.

[0093] It can be understood that the i-th phase difference data under the v-th type of phase difference has phase data, frequency data and amplitude data corresponding to the i-th phase difference data in two adjacent characteristic short-time windows. It should be noted that in order to avoid the dimension affecting the data analysis results, the phase data and amplitude data in this embodiment are both standardized data, and the method for standardizing the data is a well-known technology and will not be introduced in detail here.

[0094] For the vth type of phase difference, the i-th phase difference data corresponds to a characteristic difference value. Based on the steps of obtaining the phase difference data, it can be known that a characteristic difference value also corresponds to two adjacent short-time windows, that is, a characteristic short-time window. The characteristic difference value of two adjacent characteristic short-time windows corresponding to each phase difference data reflects the characteristic difference of the signal data between the adjacent characteristic short-time windows. It can be understood that for some short-time windows where there is no characteristic short-time window, there is no need to calculate the corresponding characteristic difference value.

[0095] Step S303: for any type of phase difference, fitting is performed based on the characteristic difference value corresponding to each phase difference data to determine the characteristic difference function of the phase difference of the any type.

[0096] In this embodiment, if Figure 6 It is a partial distribution diagram of the characteristic difference function, and shows the corresponding relationship between the characteristic difference value in the characteristic difference function and the short-time window and the sound wave signal. Figure 6The medium gray short-time window is the characteristic short-time window corresponding to the vth type of phase difference, that is, the first, second, third, fourth, seventh, eighth and ninth short-time windows in the figure are all characteristic short-time windows. Therefore, in the figure, the first short-time window and the second short-time window correspond to a characteristic difference value, that is, a data point on the characteristic difference function, and there is no characteristic difference value between the fourth short-time window and the fifth short-time window.

[0097] Specifically, the present embodiment adopts the least square method to perform function fitting, and the horizontal axis in the characteristic difference function graph is the serial number of the characteristic difference value, and the vertical axis is the value of the characteristic difference value. Based on this, by analyzing the change trend of the characteristic difference function graph, the change of the characteristic difference value can be reflected, and then the complexity of the signal difference parameters under the same type of phase difference can be reflected. If the distribution of the signal difference parameters is more complex, it means that the signal delay component corresponding to the type of phase difference is more likely to be a distortion situation.

[0098] Step S304 , obtaining a distortion characteristic index corresponding to each type of phase difference according to a change trend of each data point in a characteristic difference function of each type of phase difference.

[0099] Specifically, for any type of phase difference, the slope value at the location of each data point is calculated based on the characteristic difference function, and the mean of the slope values ​​corresponding to all data points on the characteristic difference function is negatively normalized to obtain the distortion characteristic index corresponding to the phase difference of the type. It can be understood that the derivative value of the characteristic difference function at each data point is the slope value of the corresponding data point.

[0100] As a specific example, the distortion characteristic index W corresponding to the v-th type of phase difference is v The calculation formula can be expressed as: Among them, R v It represents the mean of the slope values ​​of all data points on the characteristic difference function of the phase difference of the vth type, and exp() represents an exponential function with the natural constant e as the base.

[0101] R v It reflects the average change trend of the characteristic difference value of the phase difference of the vth type. The larger its value is, the greater the characteristic difference of the adjacent characteristic short-time window where the phase difference of the vth type is located gradually increases with the change of time series, which further indicates that the distortion mechanism of the signal component represented by the phase difference of this type is also complex, that is, the greater the possibility of distortion of the signal component, and the corresponding distortion characteristic index W v The larger the value of .

[0102] Step S400, clustering the phase differences according to the distribution characteristic index and the distortion characteristic index to obtain clustering clusters, and using the ITD decomposition algorithm to divide the acoustic wave signal to obtain interval signal segments; according to the distribution of the frequency information of the acoustic wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each clustering cluster, the matching degree between each interval signal segment and each clustering cluster is obtained.

[0103] The distribution characteristic index of each type of phase difference characterizes the representativeness of the phase differences belonging to the same type for a certain type of delayed signal components. The distortion characteristic index of each type of phase difference characterizes the performance of the signal components represented by the phase differences belonging to the same type being affected by the distortion mechanism. By combining the characteristic performance of each type of phase difference in these two aspects, the comprehensive performance of each delayed signal component being affected by the distortion mechanism can be quantified. The degree of influence of different distortion mechanisms is different. Based on this, the phase differences can be classified first to screen out the signal components affected by the same distortion mechanism.

[0104] Specifically, the Euclidean norm between the distribution characteristic index and the distortion characteristic index corresponding to each type of phase difference is used as the characteristic measurement parameter of each type of phase difference, which can be expressed as: Where L m The characteristic metric parameter representing the phase difference of the mth species, Q m represents the distribution characteristic index corresponding to the phase difference of the mth type, W m Represents the distortion characteristic index corresponding to the m-th type of phase difference.

[0105] Then, based on the difference distance between the characteristic measurement parameters of different types of phase differences, each type of phase difference is clustered to obtain a cluster. In this embodiment, the K-means clustering algorithm is used to classify all phase differences. The value of the number of clusters K can be determined by the elbow method. The cluster measurement distance between different types of phase differences can be determined by the Euclidean distance between the corresponding characteristic measurement parameters. The specific classification process belongs to the known technology and will not be introduced in detail here. In the obtained clustering results, each cluster represents the signal components with different delays affected by the same distortion mechanism, and the number of clusters can be approximately regarded as the number of distortion mechanism types affected by the distortion mechanism of the current sound wave signal.

[0106] At this point, in combination with step S200 and step S300, this embodiment uses short-time Fourier transform to extract the signal delay in adjacent short-time windows, that is, the distribution of phase differences, and then calculates the distribution characteristics of each phase difference based on the distribution characteristics of different delayed signal components in time series to describe the influence of the same distortion mechanism on different delayed signal components, and also reflects the representativeness of each phase difference for a certain type of signal component. Further, based on the characteristic difference function of the phase difference in time series, the distortion characteristic performance of each phase difference is obtained to describe the characteristics that the characteristic difference of the phase difference is more complex as the time series increases and is affected by the distortion mechanism. Then, all types of phase differences are classified in combination with the distribution characteristics and distortion characteristics, and the signal components affected by the same distortion mechanism are divided into the same cluster cluster, so that the number of types of distortion mechanisms can be quantified. The purpose of estimating the number of types of distortion mechanisms is to provide a pre-input value of the decomposition number for the subsequent decomposition of different components in the acoustic wave signal.

[0107] Based on this, in order to cater to the multi-distortion mechanism of the acoustic signal during the propagation process, the present invention uses the ITD decomposition algorithm to divide the acoustic signal into interval signal segments. The baseline extraction operator CSI commonly used in the ITD decomposition process uses the interval between adjacent extreme points as an estimated intrinsic time scale. CSI uses cubic spline interpolation technology to approximate the baseline of the signal. The specific implementation process is a well-known technology and will not be described in detail. It can be understood that in the preliminary decomposition process using the ITD algorithm, an estimated intrinsic time scale is also an interval signal segment in this embodiment.

[0108] Furthermore, the main process of the ITD decomposition algorithm is to separate the signal components by continuously subtracting the mean envelope (baseline function) within each estimated intrinsic time scale, and finally obtain the intrinsic mode function (IMF) component. At the same time, considering that the acoustic signal is affected by different distortion mechanisms, the data representation of each interval signal segment can be matched according to the degree of the impact, and the most matching combination of the original signal and the signal affected by the distortion can be obtained, so that the signals affected by the same type of distortion mechanism can be divided into the same IMF component through the interval signal segment.

[0109] Based on this, the matching degree between each interval signal segment and each cluster is quantitatively analyzed to reflect the data matching between each estimated intrinsic time scale and each distortion mechanism influence type. This embodiment takes any interval signal segment and any cluster as an example for explanation, that is, any interval signal segment is recorded as a selected interval signal segment, and any cluster is recorded as a selected cluster.

[0110] Specifically, the mean value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment is calculated to obtain the first eigenvalue of the selected interval signal segment; the entropy value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment is calculated to obtain the second eigenvalue of the selected interval signal segment. The mean value of the frequency data corresponding to all the phase differences in the selected cluster cluster is calculated to obtain the first eigenvalue of the selected cluster cluster; the entropy value of the frequency data corresponding to all the phase differences in the selected cluster cluster is calculated to obtain the second eigenvalue of the selected cluster cluster. Among them, the entropy value calculated in this embodiment is information entropy, which is used to measure the degree of information confusion of the frequency data. It should be noted that in order to avoid the dimension from affecting the data analysis results, the frequency data in this embodiment is also standardized data. The absolute value of the difference between the first eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the first difference, and the absolute value of the difference between the second eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the second difference; the product of the first difference and the second difference is negatively correlated and normalized to obtain the matching degree between the selected interval signal segment and the selected cluster cluster.

[0111] In this embodiment, the nth interval signal segment is taken as the selected interval signal segment, the tth cluster is taken as the selected cluster cluster, and then the matching degree between the selected interval signal segment and the selected cluster cluster, that is, the matching degree between the nth interval signal segment and the tth cluster cluster, can be expressed as:

[0112] G(n,t)=1―Norm(|H1 n ―P1 t |×|H2 n ―P2 t |)

[0113] Among them, G(n,t) represents the matching degree between the selected interval signal segment and the selected cluster, that is, the matching degree between the nth interval signal segment and the tth cluster, n represents the nth interval signal segment, and t represents the tth cluster;

[0114] H1 n It represents the mean value of the frequency data corresponding to all sound wave signals in the nth interval signal segment, which is also the first eigenvalue of the nth interval signal segment. n Indicates that there is information entropy of frequency data corresponding to the sound wave signal in the nth interval signal segment, which is also the second eigenvalue of the nth interval signal segment; P1 t P2 represents the mean of the frequency data corresponding to all phase differences in the t-th cluster, which is also the first eigenvalue of the t-th cluster. t It represents the information entropy of the frequency data corresponding to all phase differences in the t-th cluster, which is also the second eigenvalue of the t-th cluster; Norm() represents a linear normalization function, which is used to normalize the data.

[0115] The first eigenvalue represents the average frequency distribution in a set of data. For example, the larger the value of the first eigenvalue corresponding to the cluster, the greater the influence of the distortion mechanism corresponding to the cluster on the high-frequency signal components, and the smaller the value of the first eigenvalue, the greater the influence on the low-frequency signal components. The second eigenvalue represents the complexity of the information distribution in a set of data. For example, the larger the value of the second eigenvalue corresponding to the cluster, the more complex the signal components contained in the distortion mechanism corresponding to the cluster.

[0116] |H1 n ―P1 t | is the first difference, which indicates the difference between the signal interval segmented by the ITD algorithm and the signal parameters in the cluster represented by a distortion mechanism. |H2 n ―P2 t | is the second difference, which indicates the difference between the complexity of the signal parameters in the signal interval segmented by the algorithm and the cluster represented by the distortion mechanism. The smaller the values ​​of the first difference and the second difference, the smaller the difference between the original signal corresponding to the interval signal segment and the distortion mechanism signal represented by the cluster, and the greater the matching degree between the two, that is, the greater the value of the matching degree.

[0117] At this point, the matching degree between each interval signal segment and each cluster can be obtained according to the same method, which characterizes the feature matching between the original signal represented by the interval signal segment and the distortion mechanism signal represented by the cluster.

[0118] Step S500, optimizing the ITD decomposition algorithm by using the matching degree, performing signal reconstruction according to the signal decomposition result to obtain a preferred sound wave signal, and the preferred sound wave signal is used to obtain the positioning of the UAV.

[0119] According to the degree of matching between each interval signal segment and each cluster, the baseline function in the feature decomposition process of each interval signal segment can be optimized to optimize the decomposition process of the ITD algorithm, so as to decompose the signal components of different distortion mechanisms in time series and effectively decompose the signal components affected by distortion, so that the quality of the acoustic wave signal after signal reconstruction is better, which is helpful to accurately locate the position of the drone.

[0120] Specifically, for any cluster, the average value of the amplitude data corresponding to all phase differences in the cluster is calculated to obtain the amplitude eigenvalue of the cluster; all clusters are arranged in order from small to large according to the amplitude eigenvalue of each cluster. The amplitude eigenvalue reflects the average amplitude distribution of the cluster. The smaller the average amplitude distribution, the more likely the distortion mechanism signal component corresponding to the cluster will be decomposed first. Then, the clusters are sorted in order from small to large according to the average amplitude, which can guide the decomposition process of the acoustic signal.

[0121] Then, the sound wave signal is decomposed and processed using the ITD decomposition algorithm. During the decomposition process of each interval signal segment, the original baseline function is adjusted in sequence according to the matching degree between the interval signal segment and each cluster cluster in the order of arrangement to obtain a modified baseline function for each decomposition of the interval signal segment. The adjustment is a multiplication operation. Based on the modified baseline function and the original sound wave signal, combined with the principle of the ITD algorithm, the ITD decomposition result is obtained.

[0122] It should be noted that the decomposition principle of ITD is to separate the signal components by continuously subtracting the mean envelope (baseline function) within each estimated intrinsic time scale. When the baseline function is removed from the original signal, the remaining signal becomes the IMF component. The baseline signal is used as the new input signal, and the baseline function is repeatedly extracted and the IMF components are obtained by subtracting the baseline function until the baseline function is a monotonic function or a straight line.

[0123] In this embodiment, the number of distortion mechanism types obtained by clustering is K, and then the number of decomposition layers of the ITD algorithm is set to K+1, that is, it includes K IMF components and a final trend term signal, and at the same time, for each interval signal segment corresponding to the estimated intrinsic time scale, a baseline function of each layer of decomposition is set. It can be understood that the method for obtaining the baseline function belongs to the well-known technology of the ITD algorithm. For example, the mean envelope can be used as the baseline function, and the implementer can set it according to the specific implementation scenario.

[0124] Specifically, for any interval signal segment corresponding to an estimated intrinsic time scale, when the interval signal segment is subjected to feature decomposition, the product of the degree of matching between the interval signal segment and the first cluster in the arrangement order and the first baseline function is calculated to obtain a modified baseline function corresponding to the first baseline function.

[0125] The first IMF component can be obtained by subtracting the first corrected baseline function from the original acoustic wave signal in the interval signal segment. Since the original baseline function is corrected in this process, the phase difference between the original acoustic wave signal and the corrected baseline function will refer to the matching degree between the interval signal segment and the first cluster cluster, and the first type of distortion mechanism signal component in the arrangement order will be retained in the first IMF component.

[0126] The original sound wave signal in the interval signal segment is subtracted from the first IMF component signal to obtain the residual signal, and the residual signal is used as the original signal again, and the second baseline function is obtained. Then, according to the same acquisition method as the first IMF component, the product of the matching degree between the interval signal segment and the second cluster in the arrangement order and the second baseline function is calculated to obtain the corrected baseline function corresponding to the second baseline function. Further, after subtracting the corrected baseline function corresponding to the second baseline function from the original signal, the second IMF component can be obtained. For the same reason, the second type of distortion mechanism signal component in the arrangement order is retained in the second IMF component.

[0127] By analogy, K IMF component signals and a final trend term signal can be obtained in the end. Each IMF component corresponds to a signal component of a distortion mechanism, and the trend term signal can approximately represent the distortion-free sound wave signal. Then, adaptive smoothing is performed on each IMF component signal, and then the sound wave signal is reconstructed in combination with the trend term signal to obtain a clearer sound wave signal obtained by the microphone array, making the drone positioning information more accurate. Among them, smoothing and reconstructing the signal are well-known technologies and will not be introduced in detail here.

[0128] Finally, using the clear acoustic wave signal obtained through reconstruction, accurate drone positioning can be obtained through acoustic wave detection and other technologies. Then relevant staff can make a decision whether to send out an intercepting drone based on the specific situation. The ground control station provides the intercepting drone with accurate positioning information of the invading drone through the GFS base station, and the intercepting drone completes the interception mission.

[0129] like Figure 7 As shown, the present invention also provides an anti-UAV system based on acoustic detection, which is used to perform the steps of an anti-UAV method based on acoustic detection. The anti-UAV system based on acoustic detection includes:

[0130] A data acquisition module is used to collect the acoustic wave signal of the drone when the drone flies over the monitoring area, and obtain the frequency data and phase data of the acoustic wave signal in each short-time window;

[0131] A distribution characteristic analysis module, used to obtain distribution characteristic indicators corresponding to different types of phase differences according to the distribution of phase data differences between each short-time window and the same frequency data in adjacent short-time windows;

[0132] The distortion feature analysis module is used to analyze the distribution of the difference change trend according to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, and obtain the distortion feature index corresponding to different types of phase differences;

[0133] A feature matching module is used to cluster the phase differences according to the distribution feature index and the distortion feature index to obtain clusters, and to divide the acoustic wave signal by using the ITD decomposition algorithm to obtain interval signal segments; and to obtain the matching degree between each interval signal segment and each cluster according to the distribution of the frequency information of the acoustic wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster;

[0134] The signal optimization module is used to optimize the ITD decomposition algorithm using the matching degree, and to reconstruct the signal according to the signal decomposition result to obtain the optimal sound wave signal, and the optimal sound wave signal is used to obtain the UAV positioning.

[0135] The present application also provides a computer device. Figure 8 , which shows a schematic diagram of the structure of a computer device provided by an embodiment of the present invention, wherein the computer device includes a memory 801, a processor 802, and a computer program 803 stored in the memory 801 and running on the processor 802, wherein when the processor 802 executes the computer program 803, the computer device can execute the aforementioned anti-UAV method based on acoustic detection.

[0136] An embodiment of the present application also provides a computer program product. When the computer program product is run on a computer device, the computer device can execute the above-mentioned anti-UAV method based on acoustic detection.

[0137] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program code is stored. When the computer program code runs on a computer device, the computer device can execute the above-mentioned anti-UAV method based on acoustic detection.

[0138] In the embodiments provided in the present application, it should be understood that the provided computer device, computer program product and computer-readable storage medium are all used to execute the corresponding methods provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the methods provided above and will not be repeated here.

[0139] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An anti-UAV method based on acoustic detection, characterized in that: The method comprises the following steps: When the UAV flies over the monitoring area, the UAV's acoustic wave signal is collected, and the frequency data and phase data of the acoustic wave signal in each short-time window are obtained; According to the distribution of the phase data difference between each short-time window and the same frequency data in the adjacent short-time window, the distribution characteristic indexes corresponding to different types of phase differences are obtained; According to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, the distribution of the difference change trend is analyzed to obtain the distortion characteristic index corresponding to different types of phase differences; The phase differences are clustered according to the distribution characteristic index and the distortion characteristic index to obtain clusters, and the sound wave signal is divided into interval signal segments using the ITD decomposition algorithm; the matching degree between each interval signal segment and each cluster is obtained according to the distribution of the frequency information of the sound wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster; The ITD decomposition algorithm is optimized using the matching degree, and signal reconstruction is performed according to the signal decomposition result to obtain a preferred sound wave signal, and the preferred sound wave signal is used to obtain the positioning of the UAV.

2. The anti-UAV method based on acoustic detection according to claim 1 is characterized in that: The distribution characteristic indicators corresponding to different types of phase differences are obtained according to the distribution of the phase data difference between each short-time window and the same frequency data in the adjacent short-time window, specifically including: For each short-time window and the next adjacent short-time window, each phase difference data is determined based on the difference between the phase data corresponding to the same frequency data; wherein the same phase difference data is the same type of phase difference; According to the distribution of the number of time series features of each type of phase difference, a distribution feature index corresponding to each type of phase difference is obtained.

3. The anti-UAV method based on acoustic detection according to claim 2 is characterized in that: The step of obtaining the distribution characteristic index corresponding to each type of phase difference according to the distribution of the number of time series characteristics of each type of phase difference specifically includes: The total number of short-time windows in which the same type of phase difference appears continuously is recorded as the number of time series features of each type of phase difference; Based on the ratio between the maximum value of the number of time series features of each type of phase difference and the mean value of the number of time series features of all other types of phase differences, the distribution feature index corresponding to each type of phase difference is determined.

4. The anti-UAV method based on acoustic detection according to claim 2 is characterized in that: According to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, the distribution of the difference change trend is analyzed to obtain the distortion characteristic indicators corresponding to different types of phase differences, which specifically include: The short-time window where the phase difference of each type is obtained is recorded as the characteristic short-time window; According to the amplitude difference between the frequency data corresponding to the characteristic short-time window and the adjacent characteristic short-time window under the same type of phase difference, combined with the corresponding phase difference data, the characteristic difference value corresponding to each phase difference data under the same type is obtained; For any type of phase difference, fitting is performed based on the characteristic difference value corresponding to each phase difference data to determine the characteristic difference function of the phase difference of the any type; According to the change trend of each data point in the characteristic difference function of each type of phase difference, the distortion characteristic index corresponding to each type of phase difference is obtained.

5. The anti-UAV method based on acoustic detection according to claim 4 is characterized in that: The step of obtaining the characteristic difference value corresponding to each phase difference data under the same type according to the amplitude difference between the frequency data corresponding to the characteristic short-time window and the adjacent characteristic short-time window under the same type of phase difference, in combination with the corresponding phase difference data, specifically includes: For any type of phase difference, the amplitude difference data of each phase difference data between the amplitudes of the corresponding frequency data in two adjacent characteristic short-time windows corresponding to each phase difference data is calculated; The Euclidean norm of each phase difference data and the corresponding amplitude difference data is used as the feature difference value of each two adjacent feature short-time windows; the feature difference values ​​of the two adjacent feature short-time windows correspond to the feature difference value corresponding to one phase difference data.

6. The anti-UAV method based on acoustic detection according to claim 4 is characterized in that: The step of obtaining the distortion characteristic index corresponding to each type of phase difference according to the change trend of each data point in the characteristic difference function of each type of phase difference specifically includes: For any type of phase difference, the slope value of each data point is calculated based on the characteristic difference function, and the mean of the slope values ​​corresponding to all data points on the characteristic difference function is negatively normalized to obtain the distortion characteristic index corresponding to the phase difference of this type.

7. The anti-UAV method based on acoustic detection according to claim 1 is characterized in that: The clustering of the phase differences according to the distribution characteristic index and the distortion characteristic index to obtain cluster clusters specifically includes: The Euclidean norm between the distribution characteristic index and the distortion characteristic index corresponding to each type of phase difference is used as the characteristic measurement parameter of each type of phase difference; Based on the difference distance between the characteristic measurement parameters of different types of phase differences, each type of phase difference is clustered to obtain a cluster cluster.

8. The anti-UAV method based on acoustic detection according to claim 1 is characterized in that: The matching degree between each interval signal segment and each cluster is obtained according to the distribution of the frequency information of the sound wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster, specifically including: Record any interval signal segment as a selected interval signal segment, and record any cluster as a selected cluster; Calculate the mean value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment to obtain the first eigenvalue of the selected interval signal segment; calculate the entropy value of the frequency data corresponding to all the sound wave signals in the selected interval signal segment to obtain the second eigenvalue of the selected interval signal segment; Calculating the mean value of the frequency data corresponding to all phase differences in the selected cluster to obtain the first eigenvalue of the selected cluster; calculating the entropy value of the frequency data corresponding to all phase differences in the selected cluster to obtain the second eigenvalue of the selected cluster; The absolute value of the difference between the first eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the first difference, and the absolute value of the difference between the second eigenvalue of the selected interval signal segment and the selected cluster cluster is taken as the second difference; the product of the first difference and the second difference is negatively correlated normalized to obtain the matching degree between the selected interval signal segment and the selected cluster cluster.

9. The anti-UAV method based on acoustic detection according to claim 8, characterized in that: The optimizing the ITD decomposition algorithm by using the matching degree specifically includes: For any cluster, the average value of the amplitude data corresponding to all phase differences in the cluster is calculated to obtain the amplitude characteristic value of the cluster; Arrange all clusters in order from small to large according to the amplitude characteristic value of each cluster, decompose and process the acoustic wave signal using the ITD decomposition algorithm, and in the decomposition process of each interval signal segment, adjust the original baseline function in turn according to the arrangement order using the matching degree between the interval signal segment and each cluster, and obtain the modified baseline function of each decomposition of the interval signal segment, wherein the adjustment is a product operation; Based on the modified baseline function and the original sound wave signal, combined with the ITD algorithm principle, an ITD decomposition result is obtained. 10.An anti-UAV system based on acoustic detection, characterized in that: The system is used to perform the steps of an anti-UAV method based on acoustic detection as described in any one of claims 1 to 9, and the anti-UAV system based on acoustic detection includes: A data acquisition module is used to collect the acoustic wave signal of the drone when the drone flies over the monitoring area, and obtain the frequency data and phase data of the acoustic wave signal in each short-time window; A distribution characteristic analysis module, used to obtain distribution characteristic indicators corresponding to different types of phase differences according to the distribution of phase data differences between each short-time window and the same frequency data in adjacent short-time windows; The distortion feature analysis module is used to analyze the distribution of the difference change trend according to the difference in phase data and amplitude between each short-time window and the same frequency data in the adjacent short-time window, and obtain the distortion feature index corresponding to different types of phase differences; A feature matching module is used to cluster the phase differences according to the distribution feature index and the distortion feature index to obtain clusters, and to divide the acoustic wave signal by using the ITD decomposition algorithm to obtain interval signal segments; and to obtain the matching degree between each interval signal segment and each cluster according to the distribution of the frequency information of the acoustic wave signal in each interval signal segment and the distribution of all frequency data corresponding to the phase difference in each cluster; The signal optimization module is used to optimize the ITD decomposition algorithm using the matching degree, and to reconstruct the signal according to the signal decomposition result to obtain the optimal sound wave signal, and the optimal sound wave signal is used to obtain the UAV positioning.

Images