Rotor unmanned aerial vehicle detection method and device, electronic equipment and storage medium

Abstract

The invention provides a rotor unmanned aerial vehicle detection method and device, electronic equipment and a storage medium, and belongs to the technical field of unmanned aerial vehicles, and the method comprises the steps: carrying out the Mel frequency cepstrum coefficient MFCC feature extraction of a to-be-processed sound signal, and obtaining an MFCC feature matrix; the MFCC feature matrix is input into an SVM classifier based on a cubic kernel, and a sound source classification result output by the SVM classifier is obtained; determining an unmanned aerial vehicle sound source based on the sound source classification result; and taking the unmanned aerial vehicle sound source as input data, and determining positioning information of the rotor unmanned aerial vehicle based on a three-dimensional SRP-PHAT algorithm. According to the rotor unmanned aerial vehicle detection method and device, the electronic equipment and the storage medium provided by the invention, the unmanned aerial vehicle can be accurately positioned in a complex environment in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicles, and in particular to a method, device, electronic equipment and storage medium for detecting a rotary-wing unmanned aerial vehicle. Background Art

[0002] With the widespread application of drones in military and civilian fields, how to effectively and accurately monitor and locate drones has become a key issue.

[0003] Currently, optical and radar methods are commonly used to locate drones. However, optical positioning methods rely on the direct transmission of light. When the light path is blocked by physical obstacles (such as buildings, trees or other objects), the receiver cannot receive sufficient light intensity, resulting in positioning failure. Radar signals are also affected by the terrain during transmission. For example, obstacles such as mountains and buildings will block the radar beam and reduce the effective detection range of the radar. Especially under complex terrain conditions, the propagation path of the radar beam will be blocked and reflected many times, seriously affecting the positioning accuracy of the radar.

[0004] Therefore, how to accurately locate the UAV has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The present invention provides a rotor UAV detection method, device, electronic equipment and storage medium, which are used to solve the defect that the UAV position detection technology in the prior art is prone to failure in an obstructed environment.

[0006] The present invention provides a rotary wing UAV detection method, comprising the following steps: Perform Mel-frequency cepstral coefficient (MFCC) feature extraction on the processed sound signal to obtain the MFCC feature matrix; Inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; Based on the sound source classification result, determining the drone sound source; The drone sound source is used as input data, and the positioning information of the rotorcraft drone is determined based on the three-dimensional SRP-PHAT algorithm.

[0007] According to a rotorcraft UAV detection method provided by the present invention, the Mel-frequency cepstral coefficient MFCC feature extraction is performed on the acoustic signal to be processed to obtain the MFCC feature matrix, including: The acoustic signal to be processed is subjected to preprocessing, spectrum analysis processing and parameter transformation processing in sequence to obtain the MFCC feature matrix; The preprocessing is implemented in the following way: Based on a Butterworth high-pass filter, filtering the sound signal to be processed to obtain a noise-reduced signal; Performing frame processing on the noise reduction signal to obtain multiple frames of sound signals; Windowing is performed on each frame of the multiple frames of acoustic signals to obtain multiple frames of preprocessed acoustic signals.

[0008] According to a rotorcraft UAV detection method provided by the present invention, the spectrum analysis processing is implemented based on the following method: Performing Fourier transform on each frame of the multi-frame pre-processed acoustic signal to obtain a multi-frame frequency domain acoustic signal; After converting the multiple frames of frequency domain sound signals into multiple frames of Mel domain sound signals based on a preset conversion formula, filtering each frame of the multiple frames of Mel domain sound signals using a Mel filter bank to obtain a spectral analysis sound signal; Wherein, the preset conversion formula is as follows: ; in, represents the Mel domain acoustic signal obtained after conversion, Represents the conversion relationship between linear frequency and Mel frequency, represents a user-defined slope smoothing function, which is used to smooth the conversion slope of the high-frequency band sound signal.

[0009] According to a rotor UAV detection method provided by the present invention, the expression of the custom slope smoothing function is as follows: ; Among them, s represents the slope change factor. When s>1, the curve becomes steeper and the slope becomes larger. When s<1, the curve becomes smoother and the slope becomes smaller. k represents the sharpness coefficient, which is used to adjust the steepness of the smoothing.

[0010] According to a rotorcraft UAV detection method provided by the present invention, after extracting Mel-frequency cepstral coefficients MFCC features of the acoustic signal to be processed to obtain an MFCC feature matrix, the method further includes: After flattening the MFCC feature matrix into a one-dimensional array, the one-dimensional array is normalized based on the following normalization formula: ; in, Represents the MFCC features after normalization. Represents the original data. For each feature, is the value of the feature in all samples, represents the mean of the original data, represents the standard deviation of the original data.

[0011] According to a rotary-wing UAV detection method provided by the present invention, the UAV sound source is used as input data, and the positioning information of the rotary-wing UAV is determined based on a three-dimensional SRP-PHAT algorithm, including: After dividing the horizontal angle and the pitch angle based on the first angle value, a space grid is established using the QuickHull algorithm to obtain a first airspace grid; Based on the three-dimensional SRP-PHAT algorithm, determining a grid containing the target positioning point in the first spatial domain grid; After dividing the horizontal angle and the pitch angle of the grid based on the second angle value, a spatial grid is established using the QuickHull algorithm to obtain a second spatial grid, wherein the second angle value is smaller than the first angle value; Based on the three-dimensional SRP-PHAT algorithm, the grid containing the target positioning point in the second airspace grid is determined to obtain the positioning information of the rotary-wing UAV.

[0012] The present invention also provides a rotary wing UAV detection device, comprising the following modules: The feature extraction module is used to extract the Mel-frequency cepstral coefficient (MFCC) feature of the processed sound signal to obtain the MFCC feature matrix; A sound source classification module is used to: input the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; A sound source determination module, used to: determine the drone sound source based on the sound source classification result; The positioning determination module is used to: use the drone sound source as input data and determine the positioning information of the rotor drone based on the three-dimensional SRP-PHAT algorithm.

[0013] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, a rotor UAV detection method as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the rotor UAV detection method as described in any one of the above is implemented.

[0015] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned rotorcraft UAV detection methods.

[0016] The rotary-wing UAV detection method, device, electronic device and storage medium provided by the present invention perform Mel-frequency cepstrum coefficient MFCC feature extraction on the sound signal to be processed to obtain an MFCC feature matrix; the MFCC feature matrix is ​​input into an SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; based on the sound source classification result, the UAV sound source is determined; the UAV sound source is used as input data, and the positioning information of the rotary-wing UAV is determined based on the three-dimensional SRP-PHAT algorithm. This solution uses the sound wave signal generated by the UAV during flight for detection. The sound wave signal has strong penetrating power and can penetrate certain obstacles, so it can accurately locate the UAV in a complex environment, and the sound wave signal has strong anti-interference ability, so it can still effectively detect UAVs in areas with complex electromagnetic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 This is one of the flow charts of the rotary wing UAV detection method provided by the present invention; Figure 2 This is the second flow chart of the rotary wing UAV detection method provided by the present invention; Figure 3 It is a schematic diagram of the process of MFCC feature extraction provided by the present invention; Figure 4 It is a frequency domain waveform diagram of low frequency noise provided by the present invention; Figure 5 It is a frequency domain waveform diagram of low frequency noise after being processed by the Butterworth high pass filter provided by the present invention; Figure 6 It is a schematic diagram for comparing frequency response curves provided by the present invention; Figure 7 is a structural schematic diagram of a first airspace grid provided by the present invention; Figure 8 is a structural schematic diagram of a second airspace grid provided by the present invention; Fig. 9 It is a structural schematic diagram of the rotary-wing UAV detection device provided by the present invention.

[0019] Fig.10 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be noted that in the description of the embodiments of the present invention, 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 sentence "include one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements. The orientation or position relationship indicated by the terms "upper", "lower" and the like is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. Unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or it can be a connection between two elements. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0022] The terms "first", "second", etc. in this application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are generally of one type, and the number of objects is not limited. For example, the first object can be one or more. In addition, "and / or" represents at least one of the connected objects, and the character " / " generally indicates that the objects associated with each other are in an "or" relationship.

[0023] Combine the following Figure 1-Figure 10 The present invention describes a rotary-wing UAV detection method, device, electronic device, and storage medium provided by an embodiment of the present invention.

[0024] Figure 1 This is one of the flow charts of the rotary wing UAV detection method provided by the present invention; Figure 2 This is the second flow chart of the rotary wing UAV detection method provided by the present invention; like Figure 1-Figure 2 As shown, the method includes the following: S110, extracting Mel-frequency cepstral coefficients (MFCC) features of the processed sound signal to obtain an MFCC feature matrix; S120, inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; S130, determining the drone sound source based on the sound source classification result; S140, using the drone sound source as input data, and determining the positioning information of the rotary-wing drone based on a three-dimensional SRP-PHAT algorithm.

[0025] It should be noted that the executor of the task construction method provided in the embodiment of the present application can be a server, a computer device, such as a mobile phone, a tablet computer, a laptop computer, a PDA, a vehicle-mounted electronic device, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc.

[0026] like Figure 2 As shown, in S110, the sound signal to be processed is a segmented audio signal collected in real time. It should be noted that the length of the segmented audio signal is not limited to 100ms and can be set according to actual use requirements.

[0027] It is understandable that in the task of identifying drones, the sound sources are only divided into drones and other sound sources, which is a two-classification problem.

[0028] In the S120, the SVM (Support Vector Machine) classifier based on the cubic kernel is used to identify the drone sound source. The cubic kernel function has the ability to process nonlinear features and can better identify the complex acoustic features of drones. One-second audio clips are used during training, and short-term clips of 100 milliseconds are used in real-time detection to ensure the real-time and accuracy of detection.

[0029] In the embodiment of the present invention, the cubic kernel function is expressed as: ; in, and are two sets of MFCC feature vectors, represents their inner product, and c is the offset.

[0030] like Figure 2 As shown, MFCC feature extraction and drone sound source recognition are performed on the audio signal in real time. When the drone sound source is not recognized for a period of time, it is determined that the target sound source disappears and the detection is stopped; when the drone sound source is recognized, the next step of processing is carried out.

[0031] It should be noted that here, the cumulative duration is not limited to 1 second and can be adaptively set according to actual usage requirements.

[0032] In S140, the three-dimensional SRP-PHAT (Steered Response Power - Phase Transform) algorithm is used. This method is suitable for sound source positioning in noisy environments. Through phase weighted processing, it can suppress the influence of environmental noise and interference signals, thereby improving positioning accuracy and robustness, and can realize real-time sound source positioning of drones in complex environments.

[0033] The core of the algorithm is to calculate the time delay and the generalized cross-correlation function. The generalized cross-correlation function calculation of the signal collected by microphones i and j is expressed as: ; in, are the frequency domain signals of microphones i and j respectively, yes The conjugate of It is the time delay. Through PHAT weighting (normalizing the signal amplitude), the influence of noise and reflection can be suppressed, making the sound source positioning more accurate. By traversing all possible positions in the entire three-dimensional space, calculating the cumulative energy of each point, finding the point with the largest accumulated energy, and determining the three-dimensional position of the sound source.

[0034] The rotary-wing UAV detection method provided by the embodiment of the present invention performs Mel-frequency cepstral coefficient MFCC feature extraction on the processed sound signal to obtain an MFCC feature matrix; the MFCC feature matrix is ​​input into an SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; based on the sound source classification result, the UAV sound source is determined; the UAV sound source is used as input data, and the positioning information of the rotary-wing UAV is determined based on the three-dimensional SRP-PHAT algorithm. This solution uses the sound wave signal generated by the UAV during flight for detection. The sound wave signal has strong penetrating power and can penetrate certain obstacles, so it can accurately locate the UAV in a complex environment, and the sound wave signal has strong anti-interference ability, so it can still effectively detect UAVs in areas with complex electromagnetic environments.

[0035] In an optional embodiment, the Mel-frequency cepstral coefficient MFCC feature extraction is performed on the sound signal to be processed to obtain an MFCC feature matrix, including: The acoustic signal to be processed is subjected to preprocessing, spectrum analysis processing and parameter transformation processing in sequence to obtain the MFCC feature matrix; The preprocessing is implemented in the following way: Based on a Butterworth high-pass filter, filtering the sound signal to be processed to obtain a noise-reduced signal; Performing frame processing on the noise reduction signal to obtain multiple frames of sound signals; Windowing is performed on each frame of the multiple frames of acoustic signals to obtain multiple frames of preprocessed acoustic signals.

[0036] It is understandable that the acoustic signals generated by the rotor UAV during flight mainly include rotor rotation noise, mechanical vibration noise and aerodynamic noise. Through analysis, the acoustic signals of the drone have obvious characteristic peaks in the low frequency band (0-1000Hz) and high frequency band (around 6000-8000Hz). However, the high frequency band noise attenuates quickly during air propagation, and the low frequency band noise is usually mixed with wind noise, environmental noise and microphone self-noise. Therefore, it is necessary to enhance the high frequency band characteristics and suppress the low frequency band noise.

[0037] In the embodiment of the present invention, the MFCC feature extraction method is used for the acoustic characteristics of the rotary wing UAV. Optionally, the MFCC feature extraction technology of 128 filters is used and improved for the acoustic characteristics of the UAV.

[0038] Figure 3 It is a flow chart of MFCC feature extraction provided by the present invention, such as Figure 3 As shown, MFCC feature extraction includes preprocessing, spectral analysis and parameter transformation.

[0039] In the embodiment of the present invention, in the preprocessing stage, a Butterworth high-pass filter is added to process the noise below 50 Hz. The gain of the Butterworth filter in the passband is very flat and has no fluctuation, providing a smooth frequency response. Its transfer function expression is: ; in, is the cutoff frequency and n is the order of the filter.

[0040] Figure 4 It is a frequency domain waveform diagram of low frequency noise provided by the present invention; Figure 5 It is a frequency domain waveform diagram of low frequency noise after being processed by the Butterworth high pass filter provided by the present invention; like Figure 4-Figure 5As shown in the figure, after processing with the Butterworth high-pass filter, an obvious drone sound signal envelope can be seen.

[0041] In the embodiment of the present invention, during the frame processing, the characteristics of drone noise are different from those of speech noise, and a sound tending to be steady-state is displayed. The frame length is reasonably increased to avoid the influence of some instantaneous signals on the recognition.

[0042] In an optional embodiment, a Hamming window is used to smooth the two ends of the frame, making the boundary transition more natural and reducing spectrum leakage. The N-point Hamming window can be expressed as: .

[0043] In an optional embodiment, the spectrum analysis processing is implemented based on the following method: Performing Fourier transform on each frame of the multi-frame pre-processed acoustic signal to obtain a multi-frame frequency domain acoustic signal; After converting the multiple frames of frequency domain sound signals into multiple frames of Mel domain sound signals based on a preset conversion formula, filtering each frame of the multiple frames of Mel domain sound signals using a Mel filter bank to obtain a spectral analysis sound signal; Wherein, the preset conversion formula is as follows: ; in, represents the Mel domain acoustic signal obtained after conversion, Represents the conversion relationship between linear frequency and Mel frequency, represents a user-defined slope smoothing function, which is used to smooth the conversion slope of the high-frequency band sound signal.

[0044] Here, in the conversion of the signal from the frequency domain to the Mel domain, each frame of the signal needs to be filtered through the Mel filter group. The conversion relationship between the linear frequency and the Mel frequency is expressed as: ; On the Mel frequency scale, the conversion slope reflects the differences in processing and perception of different frequency bands, especially the differences in sensitivity between high and low frequencies. The human ear's perception of different frequencies is nonlinear, being more sensitive in low frequencies and less sensitive in high frequencies. In the embodiment of the present invention, the acoustic features of the drone are better extracted by using a custom slope smoothing function.

[0045] In an optional embodiment, the expression of the custom slope smoothing function is as follows: ; Among them, s represents the slope change factor. When s>1, the curve becomes steeper and the slope becomes larger. When s<1, the curve becomes smoother and the slope becomes smaller. k represents the sharpness coefficient, which is used to adjust the steepness of the smoothing.

[0046] In the embodiment of the present invention, a custom slope smoothing function based on a sigmoid function is introduced in the high frequency band (6-8kHz) to improve the sensitivity of the Mel filter to this frequency band, and the curve is smoothed at the edge.

[0047] Figure 6 : is a schematic diagram of the comparison of frequency response curves provided by the present invention, such as Figure 6 As shown, the dotted line is the traditional MFCC frequency response curve, and the solid line is the improved frequency response curve. The improved frequency response curve has a higher Mel frequency in the range of 6kHz-8kHz.

[0048] In an optional embodiment, after extracting Mel-frequency cepstral coefficients (MFCC) features of the acoustic signal to be processed to obtain an MFCC feature matrix, the method further includes: After flattening the MFCC feature matrix into a one-dimensional array, the one-dimensional array is normalized based on the following normalization formula: ; in, Represents the MFCC features after normalization. Represents the original data. For each feature, is the value of the feature in all samples, represents the mean of the original data, represents the standard deviation of the original data.

[0049] The rotor UAV detection method provided by the embodiment of the present invention flattens the MFCC feature matrix into a one-dimensional array and then performs standardization processing, so that the processed data is easier to be processed by a machine learning algorithm.

[0050] It can be understood that the MFCC feature matrix input to the SVM classifier is the MFCC features after the standardization process.

[0051] In an optional embodiment, the method of using the drone sound source as input data and determining the positioning information of the rotary-wing drone based on a three-dimensional SRP-PHAT algorithm includes: After dividing the horizontal angle and the pitch angle based on the first angle value, a space grid is established using the QuickHull algorithm to obtain a first airspace grid; Based on the three-dimensional SRP-PHAT algorithm, determining a grid containing the target positioning point in the first spatial domain grid; After dividing the horizontal angle and the pitch angle of the grid based on the second angle value, a spatial grid is established using the QuickHull algorithm to obtain a second spatial grid, wherein the second angle value is smaller than the first angle value; Based on the three-dimensional SRP-PHAT algorithm, the grid containing the target positioning point in the second airspace grid is determined to obtain the positioning information of the rotary-wing UAV.

[0052] It can be understood that the SRP-PHAT method realizes the location of the sound source through a global search of the space, which means that the SRP-PHAT function must be calculated once for each divided subspace. Therefore, the amount of calculation of this method is proportional to the amount of division of the subspace. The finer the division of the subspace, the higher the positioning accuracy, but the greater the amount of calculation, and the more global scanning time required.

[0053] In order to reduce the amount of calculation while ensuring positioning accuracy and improve the real-time performance of positioning, the embodiment of the present invention adopts a secondary search method. Specifically: Figure 7 is a structural schematic diagram of a first airspace grid provided by the present invention; Figure 8 is a structural schematic diagram of a second airspace grid provided by the present invention; like Figure 7-Figure 8 As shown in the figure, the center of the microphone array is used as the coordinate origin. The search is a coarse search. The horizontal angle of 0-360 degrees and the pitch angle of 0-90 degrees are divided into a point every 30 degrees (the first angle value can be set according to actual use requirements). The QuickHull algorithm is used to establish a spatial grid. At this time, the spatial domain grid is as follows Figure 7 After finding the target in the coarse grid, the block is subdivided, and the horizontal angle and pitch angle are divided into one point every 3 degrees (the second angle value can be set according to actual use requirements) for fine search. At this time, the airspace grid is as follows Figure 8 The subdivided airspace grid is shown. By using binary search to optimize the calculation process of SRP-PHAT, the real-time performance of the algorithm can be greatly improved while ensuring the positioning accuracy, making it more suitable for the application of real-time drone sound source detection.

[0054] It is understandable that, in the specific implementation process, the QuickHull algorithm may be used to establish the spatial grid, or other algorithms for calculating the convex hull may be used, and the specific algorithm used is not limited here.

[0055] In summary, the rotor UAV detection method provided by the present invention introduces a Butterworth filter and a custom slope smoothing function based on a sigmoid function to improve MFCC perception of UAV noise signals. The traditional SRP-PHAT spatial segmentation method is optimized, and a secondary search method is adopted to improve the real-time positioning while ensuring positioning accuracy.

[0056] The following is a description of the rotary-wing UAV detection device provided in an embodiment of the present application. The rotary-wing UAV detection device described below and the rotary-wing UAV detection method described above can be referenced to each other.

[0057] Fig. 9 : is a schematic diagram of the structure of the rotary wing UAV detection device provided by the present invention, such as Fig. 9 As shown, the rotary wing UAV detection device may include but is not limited to; The feature extraction module 910 is used to extract Mel-frequency cepstral coefficients (MFCC) features of the processed sound signal to obtain an MFCC feature matrix; The sound source classification module 920 is used to: input the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; The sound source determination module 930 is used to: determine the drone sound source based on the sound source classification result; The positioning determination module 940 is used to: use the drone sound source as input data and determine the positioning information of the rotary-wing drone based on the three-dimensional SRP-PHAT algorithm.

[0058] It should be noted that the rotary-wing UAV detection device provided in the embodiment of the present invention can execute the rotary-wing UAV detection method described in any of the above embodiments during specific operation, which will not be elaborated in this embodiment.

[0059] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.10 As shown, the electronic device may include: a processor 1010, a communication interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020 and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call the logic instructions in the memory 1030 to execute the rotor UAV detection method, which includes: performing Mel frequency cepstral coefficient MFCC feature extraction on the sound signal to be processed to obtain the MFCC feature matrix; Inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; Based on the sound source classification result, determining the drone sound source; The drone sound source is used as input data, and the positioning information of the rotorcraft drone is determined based on the three-dimensional SRP-PHAT algorithm.

[0060] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0061] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, when the computer program is executed by a processor, the computer can execute the rotor UAV detection method provided by the above methods, the method comprising: extracting Mel frequency cepstral coefficient MFCC features of the processed sound signal to obtain an MFCC feature matrix; Inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; Based on the sound source classification result, determining the drone sound source; The drone sound source is used as input data, and the positioning information of the rotorcraft drone is determined based on the three-dimensional SRP-PHAT algorithm.

[0062] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute the rotor UAV detection method provided by the above methods, the method comprising: performing Mel frequency cepstral coefficient MFCC feature extraction on the acoustic signal to be processed to obtain an MFCC feature matrix; Inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; Based on the sound source classification result, determining the drone sound source; The drone sound source is used as input data, and the positioning information of the rotorcraft drone is determined based on the three-dimensional SRP-PHAT algorithm.

[0063] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0064] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0065] 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 method for detecting a rotary-wing UAV, characterized in that: include: Perform Mel-frequency cepstral coefficient (MFCC) feature extraction on the processed sound signal to obtain the MFCC feature matrix; Inputting the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; Based on the sound source classification result, determining the drone sound source; The drone sound source is used as input data, and the positioning information of the rotorcraft drone is determined based on the three-dimensional SRP-PHAT algorithm.

2. The method for detecting a rotary-wing UAV according to claim 1, characterized in that: The Mel-frequency cepstral coefficient MFCC feature extraction is performed on the processed sound signal to obtain the MFCC feature matrix, including: The acoustic signal to be processed is subjected to preprocessing, spectrum analysis processing and parameter transformation processing in sequence to obtain the MFCC feature matrix; The preprocessing is implemented in the following way: Based on a Butterworth high-pass filter, filtering the sound signal to be processed to obtain a noise-reduced signal; Performing frame processing on the noise reduction signal to obtain multiple frames of sound signals; Windowing is performed on each frame of the multiple frames of acoustic signals to obtain multiple frames of preprocessed acoustic signals.

3. The method for detecting a rotary-wing UAV according to claim 2, characterized in that: The spectrum analysis process is implemented based on the following method: Performing Fourier transform on each frame of the multi-frame pre-processed acoustic signal to obtain a multi-frame frequency domain acoustic signal; After converting the multiple frames of frequency domain sound signals into multiple frames of Mel domain sound signals based on a preset conversion formula, filtering each frame of the multiple frames of Mel domain sound signals using a Mel filter bank to obtain a spectral analysis sound signal; Wherein, the preset conversion formula is as follows: ； in, represents the Mel domain acoustic signal obtained after conversion, Represents the conversion relationship between linear frequency and Mel frequency, represents a user-defined slope smoothing function, which is used to smooth the conversion slope of the high-frequency band sound signal.

4. The method for detecting a rotary wing UAV according to claim 3, characterized in that: The expression of the custom slope smoothing function is as follows: ； Among them, s represents the slope change factor. When s>1, the curve becomes steeper and the slope becomes larger. When s<1, the curve becomes smoother and the slope becomes smaller. k represents the sharpness coefficient, which is used to adjust the steepness of the smoothing.

5. The method for detecting a rotary-wing UAV according to any one of claims 1 to 4, characterized in that: After extracting Mel-frequency cepstral coefficients (MFCC) features of the processed sound signal to obtain an MFCC feature matrix, the method further includes: After flattening the MFCC feature matrix into a one-dimensional array, the one-dimensional array is normalized based on the following normalization formula: ； in, Represents the MFCC features after normalization. Represents the original data. For each feature, is the value of the feature in all samples, represents the mean of the original data, represents the standard deviation of the original data.

6. The method for detecting a rotary-wing UAV according to any one of claims 1 to 4, characterized in that: The method of using the drone sound source as input data and determining the positioning information of the rotary-wing drone based on the three-dimensional SRP-PHAT algorithm includes: After dividing the horizontal angle and the pitch angle based on the first angle value, a space grid is established using the QuickHull algorithm to obtain a first airspace grid; Based on the three-dimensional SRP-PHAT algorithm, determining a grid containing the target positioning point in the first spatial domain grid; After dividing the horizontal angle and the pitch angle of the grid based on the second angle value, a spatial grid is established using the QuickHull algorithm to obtain a second spatial domain grid, wherein the second angle value is smaller than the first angle value; Based on the three-dimensional SRP-PHAT algorithm, the grid containing the target positioning point in the second airspace grid is determined to obtain the positioning information of the rotary-wing UAV.

7. A rotary wing UAV detection device, characterized in that: include: The feature extraction module is used to extract the Mel-frequency cepstral coefficient (MFCC) feature of the processed sound signal to obtain the MFCC feature matrix; A sound source classification module is used to: input the MFCC feature matrix into a SVM classifier based on a cubic kernel to obtain a sound source classification result output by the SVM classifier; A sound source determination module, used to: determine the drone sound source based on the sound source classification result; The positioning determination module is used to: use the drone sound source as input data and determine the positioning information of the rotor drone based on the three-dimensional SRP-PHAT algorithm.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the rotorcraft UAV detection method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the rotor UAV detection method as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the rotor UAV detection method as described in any one of claims 1 to 6 is implemented.

Images