107 results about "Cross-spectrum" patented technology

The invention discloses a high-precision frequency synchronization method based on cross-spectrum phase slope. The method comprises the following steps: a reference node sends a frequency information transmission signal; the common node receives the frequency information transmission signal; the common node performs mutual correlation with a reference signal generated by a local clock through matched filtering to obtain time delay information; the common node calculates a phase difference according to the time delay information and calibrates a local frequency according to the phase difference; wherein the time delay information is measured through a cross-spectrum phase slope method, and the cross-spectrum phase slope method calculates time delay according to the slope of a phase-frequency response curve of a matched filtering result. According to the high-precision frequency synchronization method based on the cross-spectrum phase slope, the problems of low phase discrimination resolution and low frequency offset resolution during frequency synchronization caused by low sampling frequency of a distributed system are solved.

The invention belongs to the technical field of acoustic measurement, and particularly relates to an acoustic flow tube test platform and a test method thereof, transmission loss measurement of an acoustic metasurface under a flow field working condition is carried out through the constructed acoustic flow tube test platform, and a test piece mounting tube of the acoustic flow tube test platform can be used for mounting most acoustic metasurface test pieces. The related technical research of the acoustic metasurface is facilitated; meanwhile, the test platform creates a flow acoustic measurement environment, weakens the turbulence pressure disturbance influence by means of a reference signal, self-spectrum and cross-spectrum signal processing method based on a scattering matrix theory, and adopts a double-load method to perform parameter solution on a scattering matrix and calculate a class of acoustic characteristics of the acoustic metasurface to be tested, thereby solving the problem of low Mach number (llt; the method can be used for measuring the transmission loss of the acoustic metasurface in the low and medium frequency range (300-2000Hz) under the tangential flow (0.3 Mach), and the technology can be widely applied to the fields of unmanned aerial vehicles, eVTOLs, household appliances, automobiles and the like with noise reduction requirements.

The invention discloses a direct positioning method based on cross-spectral domain subspace data fusion and Taylor expansion, which comprises the following steps of: firstly, integrating time domain receiving signals of all sensors, and segmentally calculating cross correlation of the receiving signals of every two sensors; a cross-spectrum model is obtained through cross-correlation, forward space smoothing is carried out on the cross-spectrum model, a cost function is constructed through subspace data fusion, and then the extreme value of the cost function is searched to obtain rough estimation of the position of the radiation source; and finally, obtaining a position compensation value by using Taylor expansion to obtain the precise estimation of the position of the radiation source. According to the method, the complexity of calculating the subspace can be reduced, the hardware and calculation burden of the base station is reduced, and the real-time requirement of the time difference positioning system is met; meanwhile, the positioning precision of the algorithm is effectively improved. In computer simulation, compared with other time difference positioning methods, the method provided by the invention shows better positioning precision and better anti-multipath capability.

The invention discloses a sound source localization method based on multi-frequency separation and Newton optimization deconvolution, relates to the technical field of array acoustic signal processing, and is used for solving the problem that weak sound sources and multiple sound sources are difficult to identify. According to the method, dominant frequency is extracted through multichannel frequency domain analysis, a cross-spectrum matrix is constructed in combination with a near-field propagation model and a guide vector, delay summation beam forming is executed to obtain sound source preliminary distribution, then the distribution is regarded as a convolution result, a maximum likelihood model is introduced, and a two-stage deconvolution strategy of coarse estimation and Newton method fine optimization is adopted to obtain a high-resolution sound source. According to the multi-sound-source positioning method, subgrid-level analysis of sound source positions and amplitudes is achieved, finally, all frequency results are fused, continuous sound source images are smoothly output through a two-dimensional Gaussian kernel, the resolution and real-time performance of multi-sound-source positioning are remarkably improved, and the multi-sound-source positioning method is suitable for high-precision acoustic imaging in a complex sound field.

The invention discloses an indoor multi-sound-source positioning method based on DOA estimation and DOA association, and the method comprises the steps: judging single-source time-frequency points through the correlation of time delay inequality vectors of adjacent time-frequency points between microphones, and constructing a variable-size elastic single-source region; applying the FSSZ distribution condition of the sound source in a reference array to all arrays by using the high correlation characteristic of the FSSZ time-frequency distribution of the same sound source at different microphone arrays in a frame; performing middle DOA estimation on the FSSZs of all the arrays in the whole observation time period by using a circular integral cross spectrum method, and constructing a sound source position histogram in combination with FSSZ aggregation degree weighting; 2D positions of the plurality of sound sources are estimated from the sound source position histogram using 2D-CFAR. According to the invention, under the conditions of indoor reverberation and noise, position estimation can be carried out on a plurality of sound sources with an unknown number, the detection performance under the conditions of positioning precision and missed detection is obviously superior to that of an existing method, and the defects of the prior art are overcome.

The invention discloses an electroencephalogram multi-index network standardized modeling method which comprises the following steps: S1, adopting a multi-country electroencephalogram cross-spectrum data set based on a large-scale full life cycle as original electroencephalogram data, calculating a cross-spectrum matrix of the original electroencephalogram data, and preprocessing the original electroencephalogram cross-spectrum matrix; s2, constructing a weighted undirected network by using the preprocessed original electroencephalogram cross spectrum matrix to serve as a brain function network; s3, calculating network features of three dimensions of brain network function integration, function separation and centrality of the brain function network, and establishing an evolution trajectory of the network features; and S4, designing an interpretable lightweight neural network regression model, and reconstructing upper triangular elements of the brain function network to obtain a complete brain function network, namely a normalized brain function network corresponding to the age. The invention further discloses an electroencephalogram multi-index network standardized modeling system. The principle of the brain function network can be reflected more accurately.

The invention relates to the technical field of wind tunnel flow sound separation, in particular to a method for separating acoustics and related pneumatic wall pressure fluctuation in a closed wind tunnel, which comprises the following steps: acquiring sound pressure of a measurement signal in the closed wind tunnel, and storing the sound pressure in a cross-spectrum matrix; constructing a related pneumatic wall pressure fluctuation interference model of the closed wind tunnel according to physical factors such as convection velocity in the closed wind tunnel, a related transverse attenuation degree and a related longitudinal attenuation degree; constructing a related pneumatic wall pressure fluctuation interference model and a related pneumatic wall pressure fluctuation interference-acoustic separation model of probability factor analysis according to the low-rank property and the sparsity; according to the sparse prior information, constructing a prior model to constrain a related pneumatic wall pressure fluctuation interference-acoustic separation model; and obtaining maximum likelihood estimation of the parameters through an expectation maximization algorithm, and determining contribution degrees of acoustics and related pneumatic wall pressure fluctuation pressure of the to-be-tested component in the closed wind tunnel. The problem that an existing method does not consider interference physical information and is difficult to be directly applied to closed wind tunnel flow-sound separation is solved.

The invention discloses a passive target speed estimation algorithm based on a sparse Bayesian learning cross spectrum, and belongs to the technical field of array signal processing. The method comprises the following steps: constructing a target cross-correlation field model, carrying out cross-spectrum on sound pressure and a delayed sound pressure channel to obtain cross-spectrum sound intensity flow, constructing a dictionary matrix according to target cross-spectrum frequency point features, establishing a joint probability density distribution function by virtue of a sparse Bayesian learning model, and finally estimating signal power according to maximum logarithm Bayesian evidence. And carrying out spectrum peak search on the signal power estimation value to realize speed estimation of the target. According to the method, the cross-spectrum sound intensity flow is used as input, all cross-spectrum frequency points are compensated, alignment of cross-spectrum results of different frequency points is achieved, the background noise suppression effect is achieved, and multi-frequency-point processing results are fused. Meanwhile, the method has good speed estimation precision, can maintain good performance under the condition of low signal-to-noise ratio, and has high engineering application value.

The invention discloses a user health data monitoring method and system based on a smart watch, and relates to the technical field of smart wearable health monitoring, and the method comprises the steps: carrying out the optimization through Hilbert transform and cross-spectral density, carrying out the iteration of a sparse coefficient through employing MCMC, carrying out the dimension reduction construction of a track matrix through employing Ortho-SSA, carrying out the screening through ReliefF and VC-SSA, and initializing a GBTN model. NAS is used for adjustment, integration is carried out through federated average and incremental learning, and the health state is monitored. According to the method, through the VIB model, cross-spectral density optimization, GMM noise modeling and MCMC sparse iteration, the dynamic coupling relation between modes is accurately captured, the robustness of data representation is improved, and through fusion of causal reasoning and federated learning, the accuracy of the health monitoring state is improved.

The invention discloses a modularized hydraulic system testing system and method, and relates to the technical field of hydraulic testing, and the method comprises the following steps: S1, collecting outlet pulsating pressure signals of a high-pressure plunger pump and a low-pressure vane pump by adopting a nanosecond-level synchronous time mark system, generating a phase residual error map, and calculating a phase residual error map of the outlet pulsating pressure signals of the high-pressure plunger pump and the low-pressure vane pump; the hidden resonance risk is preliminarily judged through frequency drift and amplitude gradient calculation; and S2, on the basis of the synchronous acquisition base line, carrying out cross spectrum coherence degree calculation and phase difference time sequence analysis on the pulsating pressure sequence, identifying a frequency dislocation anchor point, generating a pump source pulsating interference fingerprint, and determining a time window of recessive resonance. According to the method, real-time identification, dynamic intervention and closed-loop control of recessive resonance in the hydraulic system are realized through nanosecond-level synchronous acquisition, phase conjugate inversion and frequency dislocation traction strategies, energy coupling caused by frequency dislocation is effectively avoided, the operation stability and safety of the system are improved, and the service life of equipment is prolonged.

Doppler correction of phase-encoded LIDAR includes a code indicating a sequence of phases for a phase-encoded signal, and determining a first Fourier transform of the signal. A laser optical signal is used as a reference and modulated based on the code to produce a transmitted phase-encoded optical signal. A returned optical signal is received in response. The returned optical signal is mixed with the reference. The mixed optical signals are detected to produce an electrical signal. A cross spectrum is determined between in-phase and quadrature components of the electrical signal. A Doppler shift is based on a peak in the cross spectrum. A device is operated based on the Doppler shift. Sometimes a second Fourier transform of the electrical signal and the Doppler frequency shift produce a corrected Fourier transform and then a cross correlation. A range is determined based on a peak in the cross correlation.

The invention discloses a transformer fault sound source localization method and system based on orthogonal sharp beams, and belongs to the technical field of power equipment sound source localization and state detection.The method comprises the steps that N sound sensors are arranged in a three-dimensional space around a transformer in an array mode, a three-dimensional sound field measuring plane is formed, and sound pressure signals are collected and preprocessed; constructing a cross-spectrum matrix between the acoustic sensors based on the preprocessed acoustic pressure signals, and performing eigenvalue decomposition and principal component extraction on the cross-spectrum matrix; orthogonal wave beam output is calculated by combining the cross spectrum matrix after principal component extraction, a preliminary sound source image is generated, and grid points in the preliminary sound source image are screened by combining a preset sound source intensity threshold value; using a CLEAN-SC algorithm to carry out iteration sharpening on the screened sound source image, and when an iteration termination condition is satisfied, outputting a sound source localization image after iteration sharpening wave beam processing; and an area with the intensity higher than a set threshold value is identified in the sound source positioning map, the position of the fault source is analyzed, and a fault positioning report is generated.

The invention relates to the field of fault analysis, in particular to a low-frequency fault source positioning method integrating a small sound array and asynchronous measurement, which comprises the following steps: determining the range category of an equipment monitoring range according to the number of target monitoring equipment; determining the equipment state of one type of monitoring range according to the proportion of the key monitoring equipment and the inspection reference value so as to optimize the inspection route, and determining whether to adjust the inspection equipment combination or not according to the monitoring frequency coefficient; in response to a signal abnormal condition, determining whether to adjust the signal acquisition parameters according to the signal quality parameters; determining whether to adjust the array working parameters of the key inspection equipment according to the frequency matching coefficient, and determining an array adjustment mode; and under the condition that the array adjustment is completed, the fault sound source position of the abnormal monitoring equipment is positioned based on the to-be-analyzed sound source cross-spectrum matrix, and the timeliness and the accuracy of the fault judgment result of the target monitoring equipment are improved.

The invention relates to a sparse Bayesian learning target orientation estimation method with robust near-field mismatch. The method comprises the following steps: forming an initialized spatial spectrum by using near-field wave velocity; reconstructing a cross-spectral density matrix; correcting the cross-spectral density matrix; constructing a to-be-optimized objective function and a constraint condition by combining a spatial spectrum expression of a sparse Bayesian learning algorithm; obtaining vector estimation under an optimization condition by solving an optimization problem; correcting the guide vector; updating the historical guide vector; substituting into an iterative formula of sparse Bayesian learning to complete spatial spectrum estimation; and iteratively repeating the previous steps until a stopping condition is met, and outputting a spatial spectrum estimation result. The method has good orientation estimation performance for the near-field target, and has good engineering practical prospects.

The invention provides a repositioning method of an ensemble assimilation average field based on radar combined reflectivity, and the method continuously improves the ensemble average field according to the observed radar combined reflectivity, continuously corrects a mode forecast initial field, reduces the error accumulation, and enables the forecast result to maintain higher accuracy in a longer time range. According to the method, power spectrums of combined reflectivity are obtained through transformation based on the similarity degree of mode forecast data and observation data, the relation between the mode forecast data and the observation data is established through cross spectrums, the spatial position closeness degree is judged through the phase difference between the mode forecast data and the observation data, the smaller the phase difference is, the closer the power spectrum positions are, and the more stable the power spectrums are, the closer the power spectrums are; and finding out the ensemble member closest to the observation data in the numerical mode forecasting, namely, taking the ensemble member with the minimum spatial distance from the observation data as the optimal member, replacing the ensemble average with the optimal member, updating the ensemble member, and taking the updated ensemble member as an initial field of the numerical mode forecasting, thereby improving the accuracy of the ensemble forecasting.

The invention discloses a photovoltaic panel vibration monitoring method and system based on AI and heterogeneous convergence network communication, and the method comprises the steps: enabling each sensor terminal to transmit a local relative timestamp generated by a free operation timer to a gateway base station along with an uplink data message after completing the vibration data collection and processing in the same collection period; the gateway base station calculates the relative transmission delay inequality of any node pair according to each message receiving moment and the local relative timestamp, constructs a frequency domain phase correction factor at each discrete frequency point according to the relative transmission delay inequality, and corrects the uncorrected cross-spectral density to obtain the corrected cross-spectral density; and a coherence function value is further calculated by combining the self-spectral density of each node, and a vibration abnormal area is positioned according to the spatial distribution of the coherence function value on the photovoltaic array topology. The method can improve the accuracy of multi-node vibration joint analysis and the reliability of abnormal positioning, and is suitable for online monitoring of a photovoltaic power station.

The invention discloses a large-span roof fluctuating wind pressure spectrum prediction method based on frequency domain spectrum intrinsic orthogonal decomposition. The method comprises the following steps: establishing a fluctuating wind pressure cross spectrum at any two spatial positions based on a fluctuating wind pressure spectrum; forming a fluctuating wind pressure cross spectrum matrix of all measuring points of the roof by the fluctuating wind pressure cross spectrum; solving a spectrum eigenvector matrix and a spectrum eigenvalue matrix of the fluctuating wind pressure cross-spectrum matrix; and reconstructing the fluctuating wind pressure spectrum matrix into a fluctuating wind pressure spectrum matrix with a finite order. According to the large-span roof fluctuating wind pressure spectrum prediction method based on the frequency domain spectrum eigen orthogonal decomposition, the frequency domain spectrum eigen orthogonal decomposition method based on the large-span roof surface fluctuating wind pressure cross-spectral density function can consider turbulence coherent structures in the frequency domain and the space range at the same time; the influence of characteristic turbulence on wind pressure characteristics and a wind pressure forming mechanism can be disclosed, and the method can be successfully applied to reconstruction of a fluctuating wind pressure spectrum on the roof surface and fluctuating wind pressure spectrum prediction of a sparse wind pressure matrix.

Embodiments of the present application disclose a river bed load transport characteristic determination method and device, electronic equipment and a computer readable medium, relating to the field of engineering technology. The method performs fast Fourier transform on the sound pressure signals of the microphone array to obtain a sound pressure matrix in the frequency domain; based on the sound pressure matrix, cross-spectrum operation is performed to obtain a cross-spectrum matrix; based on each scanning grid point j on the surface of the monitoring sensor plate as an imaginary sound source point, a sound propagation model is selected according to the aerodynamics theory to determine the sound signal monitored by any microphone; each grid point is scanned, the auto-spectrum of the simulated sound signal is calculated, and the sound source holographic image of the bed load shock wave is determined; the sound source intensity is simulated, the bed load spatial position is obtained, and the bed load particle size component information is obtained through the calibration function relationship. That is, in the present application, the bed load spatial position and the bed load particle size component information are obtained, this method can accurately determine the transport characteristics of the bed load, and becomes an important support for quantifying the disaster prevention and reduction work of mountainous rivers.

The application discloses a kind of arbitrary pulsating wind field active control wind tunnel physical simulation method and system, comprising: establishing generalized pulsating wind field spectrum model, it has 6 controllable pulsating wind spectrum in the energy distribution model parameter of frequency domain;Based on generalized pulsating wind field spectrum model, inter-spectrum matrix is calculated using coherence function;Based on harmonic superposition method, inter-spectrum matrix is converted into pulsating wind field space-time information;Pulsating wind field space-time information is introduced into the control program of active control wind tunnel, through real-time monitoring and iterative adjustment model parameter, until the target wind field generated reaches expectation.The application can simulate wind field with arbitrary frequency energy distribution by generalized pulsating wind field spectrum model with 6 model parameters, combined with iterative closed-loop control, significantly improve the simulation accuracy and adaptability of complex wind field such as mountain valley wind, typhoon.

The present invention relates to the field of underwater acoustic detection technology, and provides an underwater acoustic target azimuth estimation method for a small array, including: forming an orthogonal structure by using four hydrophones with two pairs perpendicular to each other and the same distance from the intersection point (or reference center), subtracting (or differentiating) the received signals of each pair, then performing spectral estimation on the subtracted (differentiated) signals, and then multiplying their spectra by the conjugate of the reference center signal respectively. Through the imaginary part of the conjugate cross-spectrum, three azimuth estimation models and an averaging method based on the three azimuth estimation models are proposed to achieve azimuth angle estimation of signals from 0° to 360°. In view of the actual requirements of portable aviation-dropped, ship-dropped or suspended small sonar buoys or underwater submersible buoys, as well as underwater mine acoustic fuzes and other devices, the present invention proposes targeted signal detection and azimuth estimation algorithms to achieve robust and reliable signal detection and azimuth estimation.

The invention discloses a multi-channel signal receiving calibration method and device based on a cross-spectrum method, and relates to the field of signal processing, and the method comprises the following steps: S1, configuring system parameters, and generating a test signal; s2, performing frequency domain transformation on the test signal; s3, performing cross spectrum analysis on the test signal; s4, performing time delay extraction on the test signal; s5, performing compensation calibration; and S6, experimental verification. According to the multi-channel signal receiving calibration method and device based on the cross-spectrum method, by optimizing a cross-spectrum calculation process, introducing high-resolution time delay estimation and improving a phase compensation mechanism, the problems existing in the prior art are effectively solved, time delay and amplitude inconsistency between channels are accurately estimated and corrected, and the accuracy of calibration is improved. The problems that the calibration precision is limited due to high calculation complexity and poor real-time performance of a traditional method, the time delay estimation performance is reduced in a low signal-to-noise ratio environment, and the requirements of a modern multi-channel system for high synchronism and low delay cannot be met are solved.

Interference removal device (100), comprising: a Fourier transform unit (3) to divide an input signal into frames which are divided into segments by a fixed time, wherein the input signal contains a component other than a target signal, and to transform the input signal into a frequency spectrum to generate a plurality of pairs of frequency spectra by using a delay amount, wherein the pairs of frequency spectra have a fixed time difference; a cross-spectra unit (4) to generate a plurality of cross-spectra or a plurality of coherences from the plurality of pairs of frequency spectra with the fixed time difference; and a disturbance removal unit (5) to average the multitude of cross-spectra or the multitude of coherences in order to extract a power spectrum of the target signal.

A wave spectrum calculation method based on GNSS wave measuring buoy is provided, which includes: placing GNSS drift buoys in the sea, collecting buoy position data; then obtaining three-dimensional velocity data of ocean waves; performing high pass filtering on velocity data of a sliding window to remove noise signals and trend terms, retaining wave motion data; calculating a cross spectrum and wave direction spectrum. The present disclosure solves limitations of traditional GNSS measurement methods applied to ocean buoy wave measurement, such as limited observation distance, high cost, tedious measurement and average calculation accuracy, and a need for correction services. The measurement accuracy is high and no additional correction services are required, which is more suitable for wave measurement, the wave spectrum algorithm in this method can adapt to more complex observation environments and incorporate results of a velocity measurement algorithm into a matched wave spectrum algorithm.

The present invention provides a method for estimating the depth of a moving target based on a single vector hydrophone, which uses the vector hydrophone to receive sound pressure and particle velocity signals; constructs a sound pressure-particle velocity cross-spectrum; performs time-domain equally-spaced discretization processing on the received sound pressure and particle velocity signals to construct a discrete signal matrix; obtains the preliminary noise-reduced sound pressure and particle velocity signals; decomposes the signals through EWT, estimates the signal frequency according to the LOFAR spectrum analysis, determines the mode function component containing the frequency, and reconstructs the signals; constructs a sound pressure-particle velocity sound field spatial interference spectrum; calculates the horizontal complex sound intensity and the vertical complex sound intensity to estimate the vertical arrival angle of the target direct wave; performs a discrete Fourier transform on the sound pressure-particle velocity sound field spatial interference spectrum along the curve of the vertical arrival angle of the target direct wave, and the sub-maximum value after removing the maximum value at the vertical symmetry axis corresponds to the target depth estimation value. The present invention realizes the estimation of the depth of a moving target by using the received signals under the influence of noise, and has good practical engineering application capabilities.

This invention relates to a method and system for locating coaxial counter-rotating sound sources in rotorcraft, belonging to the technical field of sound source identification methods. It includes: processing the sound pressure signal of the moving sound source using continuous wavelet transform to obtain time-frequency domain variation results; performing Doppler effect correction based on experimental conditions; estimating the spatial location and relative intensity of the noise source using a cross-spectral matrix and weight vector; and performing phase-locked analysis of the rotor noise source intensity using a phase-locked averaging method. This invention differs from traditional time-domain delay superposition algorithms and frequency-domain cross-correlation spectrum methods. It can image unsteady sound sources using time-frequency domain characteristics, solving the acoustic imaging problem of counter-rotating propeller noise and possessing potential applications in the rotorcraft industry.

The invention discloses a weak sound source sound pressure level calibration method based on an improved DAS beam, relates to the technical field of sound source positioning, and is used for solving the problem of repeated weak sound source positioning. According to the method, a distance matrix is constructed based on array geometric parameters, and a cross spectrum matrix and a steering vector are respectively calculated for a plurality of target frequencies in combination with a near-field propagation model to obtain a multi-frequency acoustic image; and performing normalization and relative threshold side lobe suppression on each frequency sound image result, superposing and highlighting a real sound source area on a scanning grid, suppressing a side lobe irrelevant to a position, and utilizing a sound pressure level calibration model obtained by training reference data to convert a sound image output quantity into a real sound pressure level, so as to realize spatial positioning and quantitative sound pressure evaluation of multiple sound sources. The weak sound source can still be clearly distinguished under the background that the sound pressure level difference of the strong and weak sound sources is large and the noise is complex, and the sound source positioning precision and the engineering detection reliability are improved.

The present application belongs to the technical field of acoustic measurement, and particularly relates to an acoustic flow tube test platform and a test method thereof, wherein the transfer loss measurement of an acoustic metasurface under a flow field working condition is performed through the acoustic flow tube test platform, the test piece installation tube of the acoustic flow tube test platform can install most of the test pieces of the acoustic metasurface, and is beneficial to the acoustic metasurface related technology research; meanwhile, the test platform creates a flow acoustic measurement environment, based on the scattering matrix theory, the reference signal, the self-spectrum and the cross-spectrum signal processing method are used to weaken the influence of the turbulent flow pressure disturbance, the double load method is used to solve the parameters of the scattering matrix and calculate the acoustic characteristics of the acoustic metasurface to be measured, and the transfer loss measurement problem of the acoustic metasurface under the low Mach number (<0.3 Mach) tangential flow and in the low-middle frequency range (300-2000 Hz) is solved, and the technology will be widely used in the fields requiring noise reduction such as unmanned aerial vehicles, eVTOL, household appliances, automobiles and the like.

The invention belongs to the technical field of ship and ocean engineering, and discloses a method and a system for inverting a three-dimensional sea wave spectrum along with ship motion based on a deep learning model. According to the method, cross spectrum calculation is carried out among different ship motion time sequences to obtain a ship motion cross spectrum; carrying out data shuffling processing on the ship motion cross spectrums obtained in different time periods, and constructing a training set and a test set; and according to the constructed training set, inputting the training set into a CNN-ED-DE model, and carrying out CNN-ED-DE model training. And inputting the obtained input test data set into the trained CNN-ED-DE model, and solving to obtain a test data set inversion result. The method provided by the invention is completely based on motion sensor data such as a shipborne inertial measurement unit (IMU), does not need external remote sensing equipment or communication support, can estimate sea wave spectrum information in real time in any sailing state, and is suitable for online deployment in scenes such as complex course and unsteady sea conditions.

The invention provides a sea wave direction detection method and system based on an extended maximum likelihood method, and belongs to the technical field of ocean monitoring. The method comprises the following steps: acquiring sea wave height time sequence data of a measurement point location, and performing preprocessing; performing inversion on the preprocessed wave height time sequence data to obtain wave height data of the sea waves; constructing a cross spectrum matrix according to the wave height data; acquiring position information of the measurement point and setting a transfer function of the measurement point; based on the cross spectrum matrix, the position information of the measurement points and the transfer function, utilizing an extended maximum likelihood method to generate a sea wave direction spectrum; and drawing a heat map according to the sea wave direction spectrum, and determining the main wave direction of the sea wave according to the color shade change in the heat map. By means of the method, the precision and robustness of sea wave direction detection are remarkably improved, and powerful technical support is provided for ocean monitoring.

The invention discloses a rotating sound source recognition and positioning method based on a deep convolutional attention neural network, and belongs to the field of rotating sound source recognition, and the method comprises the steps: firstly carrying out the motion compensation and modal expansion of a sound pressure cross-spectrum matrix through a modal decomposition rotating beam forming algorithm, and generating a sound source distribution diagram in a rotating scene; then the sound source distribution map is input into a deep convolution attention neural network model, the model adopts an encoder-decoder structure, a channel and a space attention mechanism are fused, key sound source features are dynamically focused, noise interference is suppressed, and high-precision reconstruction of the sound source distribution map is achieved; and finally, carrying out local maximum value detection on a prediction map output by the network, and extracting a sound source space position. According to the method, the Doppler effect and phase distortion problems in rotating sound source positioning are effectively overcome, and the positioning precision, the anti-noise performance and the calculation efficiency in a complex rotating field are remarkably improved.