77 results about "Self correlation" patented technology

The invention discloses an HRF signal synchronization method, device and system of a dual-mode carrier communication system, and belongs to the technical field of power line carrier communication. The method comprises the following steps: dividing a frequency offset range into a plurality of frequency offset hypothesis intervals based on an allowable maximum residual frequency offset range; performing frequency offset pre-compensation on the received signal according to each frequency offset hypothesis interval to obtain multiple paths of frequency offset compensation signals, and performing cross-correlation processing on the multiple paths of frequency offset compensation signals and a local reference sequence to obtain multiple paths of cross-correlation value sequences; in any cross-correlation value sequence, performing self-correlation processing on cross-correlation values of adjacent symbol intervals to obtain a corresponding self-correlation sequence; correlation peak values corresponding to the frequency offset hypotheses are extracted from the multiple paths of self-correlation sequences respectively, the maximum correlation peak value is determined from the correlation peak values, and if the maximum correlation peak value is larger than a synchronization judgment threshold value, the corresponding position is determined to be the synchronization position. According to the invention, high-precision wireless signal synchronization is realized, and high-quality signal transmission is ensured.

The invention relates to the technical field of civil aircraft maintenance and repair, in particular to a flap health state prediction method based on transferable trend grouping autocorrelation comparative learning, which comprises the following steps of: firstly, mining a typical degradation trend from total data by utilizing a clustering algorithm; dividing the individual samples with differentiation into a plurality of trend data groups with high consistency; secondly, introducing an autocorrelation comparison mechanism into each group, capturing common time sequence dependence crossing a long period among samples, and constructing a plurality of pre-training models capable of representing a specific decline mode; then, dynamic matching is carried out according to the characteristics of the flap to be detected, and the most adaptive trend category and the pre-training model are selected; and finally, on the basis of a migration fine tuning strategy, a small amount of data is used for carrying out directional fine tuning on the pre-trained model, so that an exclusive prediction model of the flap is efficiently obtained, compared with the prior art, a model does not need to be built for each aircraft from zero, migration can be carried out on the basis of a limited model library, and effective management of the health state of the civil aircraft is achieved.

The invention provides a heterogeneous data intelligent fusion method, system and device based on a multi-modal sensor and a medium, and belongs to the technical field of sensor induction data fusion. Self-correlation and cross-correlation functions are calculated through recursive average, and then the variance of each sensor is estimated; calculating a trend correction factor based on the variance sliding average value, and determining the optimal weight of each sensor; calculating an average value of measured values of the sensor at each moment to form a sequence; a polynomial fitting order is adaptively selected through an AIC criterion, and a fitting function is established; and finally, outputting a final fusion result in combination with the weighted fusion value and the fitting value. According to the method, the accuracy, the stability and the adaptive capacity of data fusion of the multi-source heterogeneous sensor are effectively improved, and the anti-interference performance and the robustness of the system are enhanced.

The invention belongs to the technical field of industrial artificial intelligence, and particularly relates to an unsupervised multi-mode industrial anomaly detection method based on self-correlation learning. The method comprises the following steps: acquiring a to-be-detected RGB image and depth map data, and generating a multi-modal pseudo sample based on Berlin noise and texture filling so as to enhance the discrimination capability of a model for abnormity; rGB and deep modal features are fused by using comparative learning, and interaction information between modals is extracted; calculating local association and global association of the features by using a learnable Gaussian kernel function and a self-attention mechanism respectively; and finally, the association difference between the local association and the global association is used as a discrimination standard, and the reconstruction error is combined to calculate an anomaly score, so that the tiny anomaly on the surface of the industrial product is accurately detected and positioned. According to the method, the self-correlation learning framework which fuses the multi-modal features and uses the global-local correlation difference as the anomaly discrimination basis is constructed, so that high-precision anomaly positioning is realized under the condition of lack of anomaly sample labeling.

The embodiment of the invention provides a frame synchronization method and device based on a WI FI system, and the method comprises the steps: carrying out the delay self-correlation of a short training sequence of a main channel and a secondary channel, carrying out the bandwidth detection of a current WI FI frame according to the delay self-correlation result of the main channel and the secondary channel, and when the bandwidth detection result of the current WI FI frame is greater than a specified bandwidth, carrying out the bandwidth synchronization of the current WI FI frame. And coarse synchronization is carried out according to a delay self-correlation result of the main channel and the secondary channel to obtain long training sequences of the main channel and the secondary channel, and a correlation peak is searched for a cross-correlation result of the long training sequences of the main channel and the secondary channel and a long training sequence of a local sequence to carry out fine synchronization so as to determine a symbol boundary of the current WIFI frame. According to the method and the device, the problems that only a main 20M channel is used, other channels are not used, and timing errors are deteriorated due to the length of the PPDU are solved, and the effects of improving the frame synchronization precision of the WI FI system and reducing the error code probability caused by the overlong length of the PPDU are achieved.

A method of processing an electrocardiosignal capable of accurately obtaining heartbeat information, a device for processing an electrocardiosignal, a computer-readable storage medium, and a computer program product. The method of processing an electrocardiosignal includes: a detection signal acquisition step of acquiring a detection signal by detecting an electrocardiosignal of a living body by a sensor; a signal preprocessing step of dividing the detection signal into a plurality of segmented signals with a prescribed time length, calculating an average value of a plurality of signals included in each of the segmented signals, and obtaining a preprocessed signal by subtracting the average value of the segmented signal from each signal sample point included in the segmented signal; and a self-correlation calculation step of calculating a self-correlation function of the preprocessed signal.

The application discloses a water supply network sensor optimization arrangement method based on graph signal processing, and comprises the following steps: acquiring topological structure information of a water supply network, and establishing a user node weighted graph without direction and connection according to the topological structure information to determine a self-correlation localized graph information value of each user node and a cross-correlation graph information redundancy value between different user nodes; wherein the self-correlation localized graph information value of the user node is: the user node v i The cross-correlation graph information redundancy value between the user nodes v j According to the self-correlation localized graph information value of the user node and the cross-correlation graph information redundancy value between different user nodes, a target function is determined: user nodes meeting the target function are screened out as sensor nodes one by one until the number of sensors reaches the requirement.

The invention relates to an obstacle recognition method and device, electronic equipment and a storage medium. The obstacle recognition method comprises the steps that the aftershock value of a current detection period and the aftershock value of a historical detection period of an ultrasonic sensor are acquired; according to the aftershock value of the current detection period and the aftershock value of the historical detection period, calculating an aftershock average value of the whole operation period and an aftershock moving average value of the latest m detection periods; wherein the whole operation period is the operation time after the ultrasonic sensor is powered on; calculating an autocorrelation coefficient between the aftershock value of the current detection period and the aftershock value of the whole operation period according to the aftershock value of the current detection period, the aftershock average value and the aftershock moving average value; the self-correlation coefficient is compared with a preset self-correlation coefficient threshold value, whether the obstacle exists in the detection blind area or not is judged according to the comparison result, and the obstacle in the detection blind area can be effectively recognized when the obstacle and the probe are both static.

The application discloses a method for self-interference multipath channel estimation based on zero correlation zone sequence. The method comprises the following steps: in the channel estimation stage of the simultaneous full duplex communication, a zero correlation zone sequence is sent; by using the good autocorrelation and cross-correlation characteristics of the zero correlation zone sequence, the receiver uses the same zero correlation zone sequence to slide correlate with the received signal, and the self-interference multipath channel can be accurately decomposed into several single-path channels; and the single-path channel parameters are estimated by using the correlation peak and the known zero correlation zone sequence. The application can more accurately estimate the self-interference multipath channel parameters, and ensures the self-interference cancellation ability in the multipath environment.

The invention relates to the technical field of communication, and discloses a communication method, device and system. The method comprises the following steps: a first communication device generates a synchronization signal according to a first sequence and sends the synchronization signal; wherein the first sequence is ideal autocorrelation, the value set of elements of the first sequence is {-A, 0, A}, and A is a constant. By adopting the method, the first sequence comprises at least one element with the value of 0, and the first sequence is ideal in self-correlation, so that the detection complexity of the synchronization signal can be reduced while the detection performance of the synchronization signal is ensured.

According to the tunnel operation safety monitoring method and device, the storage medium and the computer equipment provided by the invention, when an operator performs tunnel detection operation, the original signal of the operator is collected in real time, and signal optimization is performed on the original signal through an anti-interference algorithm, so that tunnel environment noise is inhibited, and high-quality PPG signals and ECG signals are obtained; then heart rates of the two signals are determined through a self-correlation function and Fourier transform, heart rate calculation in the aspects of time domain and frequency domain is achieved, meanwhile, parameter characteristics of the two signals are predicted through a heart rate prediction model, the heart rates are obtained, and therefore the nonlinear relation between the signal characteristics and the heart rates is learned; and finally, performing weighted average on each heart rate based on a self-adaptive weight distribution strategy to obtain a final heart rate, so that the final heart rate has relatively high robustness and accuracy, and when the final heart rate exceeds a normal physiological threshold, an alarm can be given and the real-time position of the operator can be shared, thereby ensuring the life safety of the operator.

The invention belongs to the technical field of material characteristic analysis, and particularly relates to a prefabricated part in-mold compactness monitoring method and system. The method comprises the following steps: applying impact to a measuring point outside a mold, collecting elastic wave time domain response signals of adjacent sensors, and constructing self-correlation attenuation-Gaussian composite window cut-off signals; fourier transform is carried out on the truncated signal, a dimension-entropy weight composite spectrum is generated after frequency sub-bands are divided, and an inherent frequency peak is positioned; calculating a quality factor of the inherent frequency peak, extracting harmonic frequency peak energy to form a multi-scale spectrum peak energy distribution vector, and calculating Shannon entropy; matching a temperature-quality factor coupling model from a preset model library by using Shannon entropy, and correcting a quality factor in combination with an in-mold environment temperature; and inputting the inherent frequency peak, the temperature compensation quality factor and the energy distribution vector into a nonlinear mapping model, and outputting the in-mold compactness of the prefabricated part. According to the method, high-precision and reliable quantitative evaluation of the in-mold compactness of the prefabricated part is realized.

The invention provides a time sequence prediction method based on deep learning, and relates to the technical field of time sequence prediction, and the method reduces the length of a long sequence through multi-scale decomposition, replaces the traditional full-connection attention through combining with an autocorrelation attention mechanism, prevents the calculation overhead from increasing along with the square of the length of the sequence, remarkably improves the processing efficiency of the long sequence, and improves the accuracy of time sequence prediction. A multi-scale decomposition and bidirectional fusion strategy can simultaneously capture fine-grained fluctuation (the finest scale) and a macroscopic trend (the coarsest scale), the non-stationary data processing capability is superior to that of an existing model, and sequential decomposition (trend-seasonal decoupling) and self-correlation attention (long-range dependence modeling) have a synergistic effect, so that loss of key time sequence information is reduced; the method does not need to depend on manually designed prior hypotheses (such as a fixed period and coarse scale initialization), can adaptively process periodic and non-periodic data, and is wider in application scene.

The application discloses a multispectral image registration method based on multiscale deep feature map fusion, which can be used for registering spectral images between different wave bands. The method first acquires consistent feature maps of a target image and a source image by using a multiscale consistent feature map mapper, performs self-correlation feature modification on the multiscale consistent feature maps by using a transformer structure to obtain multiscale matching information feature maps, and then processes the multiscale matching information feature maps based on a self-attention mechanism to output registration parameter residuals. Under the supervision of an image registration true value and consistent feature map content consistency loss function, the network is iteratively run until a preset iteration number is reached, and finally, registration parameters are obtained. Experimental results and analysis on a multispectral dataset show that the multispectral image registration method based on multiscale deep feature map fusion effectively improves the registration accuracy of multispectral images.

The embodiment of the invention discloses a data transmission method, device and system. The method comprises: acquiring a data frame comprising a plurality of subframes, the subframes comprising 226 pilot symbols in one polarization direction, the 226 pilot symbols being generated by a target polynomial and a seed, the target polynomial being x11 + x10 + x9 + x7 + 1; and sending the data frame. The super-frame architecture provided by the embodiment of the invention is relatively low in frame redundancy, and relatively good in sequence self-correlation and cross-correlation characteristics of the pilot symbols, so that the recovery signal of a receiving end is improved, and the quality of the recovered signal is improved.

The invention relates to a pulsed eddy current imaging method of a planar flexible fractal difference measurement eddy current sensor, and belongs to the technical field of nondestructive testing of conductive metal. Compared with the traditional sine excitation, the used pulse excitation can realize the detection of different depth defects of the tested block without using a plurality of different frequency excitation, the detection efficiency is high, the signal conditioning circuit is simple, and the cost is low. Under the condition that a single-frequency excitation high-performance phase-locked amplification technology is not used, the same sensitivity is achieved, and the method has the advantages that the system is simple and easy to implement. By adopting the method, the imaging quality of new features extracted by different cross-correlation algorithms can be quantitatively evaluated, and cross-correlation features and self-correlation features with good signal-to-noise ratios can be distinguished.

The invention relates to the technical field of data processing, and particularly provides a method, a system, equipment and a medium for restoring an unknown signal constellation diagram, and the method comprises the steps: carrying out the self-correlation calculation of an in-phase I-path component of a received signal, carrying out the spectrum analysis of a self-correlation result, so as to enhance a spectrum component corresponding to a symbol rate, estimating the code element rate of the signal from the frequency spectrum; based on the code element rate, analyzing energy change characteristics of I-path and orthogonal Q-path components of the received signal, and detecting and identifying a code element boundary position to obtain a code element boundary sequence; and for each code element interval divided by the code element boundary sequence, aggregating I-path and Q-path sampling values in a stable region of each code element interval to obtain constellation point coordinates corresponding to each code element, and generating a constellation diagram by all constellation points. According to the invention, clear and accurate constellation diagram recovery of various modulation signals, especially high-order QAM signals, under the condition of low signal-to-noise ratio is realized.

The invention provides a massive spatio-temporal data retrieval acceleration method and device based on imitation learning, and the method comprises the steps: modeling a cache decision problem in retrieval acceleration into a Markov decision process, and constructing a CorrelaCache model which comprises two parallel sub-networks: on one hand, carrying out the time sequence coding of a short-term access sequence through a long and short-term memory network, and on the other hand, carrying out the time sequence coding of the short-term access sequence through a short-term memory network; capturing a local access mode; and on the other hand, an autocorrelation module is constructed based on fast Fourier transform, periodic dependency features in the access sequence are identified, and periodic phases are aligned. And the joint prediction layer fuses the features from the LSTM and the self-correlation module, and outputs the expelling probability distribution of each cache data unit through full connection layer mapping. In the training process, a DAgger algorithm is adopted to iteratively update the strategy so as to avoid distribution deviation between strategy training and actual operation; and meanwhile, a composite loss function containing sorting loss and reuse distance prediction loss is designed. The retrieval response efficiency and the resource utilization rate of the method are obviously superior to those of an existing method.

The invention discloses a method for eliminating autocorrelation artifacts of a polarization sensitive optical coherence tomography image. Comprising the following steps: carrying out sample detection by using a PS-OCT (Polarization Sensitive Optical Coherence Tomography) system to obtain interference signals of two polarization channels; performing inverse Fourier transform on the interference signal of one channel to obtain an airspace signal of the interference signal; finding a self-correlation horizontal fringe artifact signal with the maximum intensity in the interference signals of the airspace, and setting the intensities of other positions to be zero; calculating an unwrapping phase of the self-correlation artifact in each A-scan, and taking a minimum value of the unwrapping phase as a reference phase; interpolating the spectral phases of the original interference signals of the two polarization channels according to a reference phase; and performing inverse Fourier transform and polarization parameter calculation to obtain an OCT intensity image and a polarization image with autocorrelation artifacts eliminated. According to the image phase information-based self-correlation artifact removal method, the self-correlation artifacts in the PS-OCT image can be effectively eliminated, additional hardware facilities do not need to be introduced, and the method is simple and effective.

The invention discloses a period-phase optimized robot tactile signal segmentation and alignment method, which can quickly extract effective signal segments when a robot contacts an object, and solves the problems of unstable tactile signal form, baseline drift, uncertain period length and the like. The method provides a joint optimization normal form: firstly, estimating the average period of a tactile signal through self-correlation analysis; then, unified modeling is carried out on the segmentation tangency points and the fragment duration under structural constraints to form variables to be optimized, and an aggregation similarity objective function with normalized mutual measurement standards is constructed; and finally, iteratively solving the optimal segmentation point and phase of the time sequence signal in the contact process by maximizing the objective function. Compared with a traditional non-global optimization method, the method can be suitable for non-fixed-length quasi-periodic signals with unstable forms, has high robustness for amplitude drift and periodic jitter, is higher in segmentation precision and universality, and can provide accurate contact segment effective signals for robot touch recognition.

The application provides a time-frequency learning estimation method for mixed noise sinusoidal signal frequency, and belongs to the technical field of signal processing, comprising the following steps: obtaining and preprocessing a mixed noise sinusoidal signal; filtering to obtain a first signal; calculating a self-correlation function; calculating a preliminary period estimation value of the first signal to obtain a first frequency estimation value; inputting a frequency spectrum feature and a self-correlation peak value feature of the first signal and the first frequency estimation value into a pre-trained frequency fine estimation model to obtain a fine frequency estimation value, which is recorded as a second frequency estimation value; optimizing the last M layers of the model to obtain a first model; continuously optimizing to obtain a second model; continuously obtaining the frequency spectrum feature, the self-correlation peak value feature and the first frequency estimation value of the first signal and inputting them into the second model, and taking the output result of the second model as a final frequency estimation value of the mixed noise sinusoidal signal and outputting the final frequency estimation value.

The invention discloses a power distribution network high-resistance grounding fault feeder identification method based on self-correlation entropy difference characteristics. The method comprises the following steps: S1, collecting zero-sequence current waveforms of all feeders of a power distribution network; s2, performing filtering processing on the zero-sequence current waveform by using the zero-sequence current waveform obtained in the step S1 to obtain a zero-sequence current transient component after the power frequency component is filtered; s3, calculating an autocorrelation entropy difference characteristic value for the zero-sequence current transient component by applying the zero-sequence current transient component obtained in the step S2; and S4, determining the feeder line corresponding to the maximum autocorrelation entropy difference characteristic value as a fault feeder line by using the autocorrelation entropy difference characteristic value obtained in the step S3. Aiming at the high-resistance grounding fault, the invention provides a new fault feature self-correlation entropy difference, and the feeder line identification accuracy of the high-resistance grounding fault is improved.

The invention relates to the technical field of data processing, and discloses an artificial intelligence public service management platform based on a block chain, and the platform comprises the steps: obtaining a bookshelf region gray image, and carrying out the preprocessing of the gray image, and obtaining a standard image; determining a main period of the standard image by self-correlation; carrying out alignment according to the main period, and calculating a phase slip index, doubled intensity and a valley-to-peak ratio; forming an abnormal scalar by using the phase slip index, the magnification intensity and the valley-to-peak ratio, and performing positioning segmentation according to the abnormal scalar, including: if the abnormal scalar belongs to a first range, performing positioning segmentation on the grayscale image by using a convolutional neural network; if the abnormal scalar belongs to a second range, micro yaw angle re-acquisition is executed to obtain a target frame, and the convolutional neural network is controlled to carry out positioning segmentation on the target frame; if the abnormal scalar belongs to a third range, calling manual positioning segmentation; the block chain node is in data connection with the data acquisition module, the first processing module, the second processing module and the anomaly measurement module for data communication connection.

A method for designing a frequency domain ZC sequence synchronization head resistant to ultra-large frequency difference belongs to the technical field of communication, and comprises the following steps: avoiding correlation peak ambiguity caused by overlarge frequency offset through piecewise iterative estimation and compensation of decimal frequency offset; a ZC sequence sideband protection design is adopted, so that a sideband part and a base sequence form a continuous structure, and the time synchronization robustness is improved; in combination with pilot frequency cross-correlation and self-correlation processing, a cross-correlation signal is divided into subsequences for conjugate multiplication and summation, and accurate estimation of integer frequency offset is realized. According to the invention, the synchronization reliability and the system spectrum efficiency can be remarkably improved, high-precision synchronization in an ultra-large frequency difference range is supported, and the synchronization problem in a high-speed mobile communication scene is effectively solved.

The invention discloses a monitoring method and device for quality data of a self-correlation mixing process, and relates to the technical field of disease monitoring. The method comprises the following steps: obtaining a group of multivariate self-correlation data in a historical controlled state and a group of multivariate self-correlation data in a historical out-of-control state; the method comprises the following steps: respectively calculating a sample mean value and a covariance matrix, carrying out decorrelation on samples through a Cholesky decomposition method, constructing an MEWMA sequence by utilizing standard independent data after decorrelation, training a classifier model of a random forest, designing a control chart method based on the random forest, and monitoring a linear autocorrelation process; on the basis of data decorrelation, according to the random forest theory, a novel control chart is generated, the modeling and monitoring problems of quality data in the self-correlation mixing process are effectively solved, the trained random forest model automatically learns and captures complex and nonlinear modes of data in controlled and out-of-control states, and the quality of the self-correlation mixing process is improved. And effective modeling and monitoring can be carried out no matter the data is in skewed, multi-peak or completely unknown distribution.

The invention relates to the technical field of geophysical exploration, and particularly discloses a while-drilling seismic drill string reference signal feature extraction method, which comprises the following steps: obtaining a drill string reference signal; performing self-correlation processing on the drill string reference signal to obtain a correlation domain drill string reference signal; obtaining a correlation domain drill string reference signal frequency spectrum; according to the correlation domain drill string reference signal frequency spectrum, the correlation domain drill string reference signal is decomposed into a low and medium frequency mode and a high frequency mode, and a low and medium frequency signal and a high frequency signal are obtained; performing sparse decomposition on the low and medium frequency signals, further introducing a regularization parameter optimization strategy based on spectrum energy physical constraint, and extracting low frequency signals; and subtracting the low-frequency signal from the low and medium-frequency signal to obtain an intermediate-frequency signal. According to the method, the random noise is suppressed by using self-correlation, and then frequency band separation and sparse reconstruction are performed, so that the target frequency band characteristics of the drill stem reference signal are effectively extracted.

The invention discloses an ultrafast laser time domain shaping and calibrating device, and belongs to the field of ultrafast laser processing. Comprising a time domain shaping module and a self-correlation calibration module, the time domain shaping module splits linearly polarized ultrafast laser into a fixed light arm and a delay light arm, regulates and controls double-pulse time delay through an optical delay line, independently controls laser parameters of the two light arms, and ensures mutual orthogonality of double-pulse polarization states; the self-correlation calibration module is connected with the time domain shaping module through a coaxial optical path, converts double pulses into second harmonic (SHG) signals through an II-type phase matching nonlinear crystal, collects SHG signal intensities under different time sequences through a scanning delay line, generates an intensity self-correlation curve, calibrates a delay zero point and calculates pulse width, and the time domain shaping module is connected with the time domain shaping module through a coaxial optical path. Measurement parameters are fed back to the time domain shaping module for correcting delay parameters. The device can be applied to an existing laser processing platform, a time domain shaping double-pulse laser processing function is provided, the heat accumulation phenomenon is effectively reduced, and the ultrafast laser processing quality is improved.

The invention discloses a main period and amplitude detection method for a horizontal attitude angle of a marine dynamic inclinometer, and the method comprises the following steps: sampling and caching a horizontal attitude angle measurement value output by the marine dynamic inclinometer at high frequency, the caching number being related to the sampling frequency and the maximum and minimum values of the main period; after the cache is filled, subtracting a mean value from the cache sequence, calculating a positive lagging self-correlation coefficient sequence, detecting a peak value position of the self-correlation coefficient sequence, eliminating interference data, and taking the time corresponding to the processed highest peak point as an estimated value of the main period of the attitude angle; establishing an attitude angle amplitude fitting model by taking the estimated value of the main period as a parameter, and fitting the cache sequence from which the mean value is subtracted by adopting a least square fitting method to obtain an amplitude corresponding to the main period; according to the method provided by the invention, the detection precision can be ensured, and the cache requirement on hardware can be reduced.

The present application relates to the technical field of aviation safety and weather prediction, and particularly relates to a method, system and device for predicting aircraft jolt based on decoupled physical constraints; the pre-collected multi-time sequence weather forecast data and aircraft EDR data are aligned on the same space-time grid and converted into probabilistic features containing mean and standard deviation; for sparse aircraft EDR data, an atmospheric physical model correction term based on the relationship between wind shear and turbulent energy dissipation rate is introduced to fill in, and a probabilistic input tensor is obtained; the probabilistic input tensor is input into a self-correlation attention model with fused physical constraints which is pre-constructed and trained; a contribution matrix is extracted from the attention weight of the trained model, the contribution proportion of each prediction time point to the prediction result is quantified, and the probabilistic prediction result of the aircraft jolt and the corresponding contribution matrix are output as the final prediction result; a reliable basis is provided for flight decision, and the accuracy and robustness of aircraft jolt prediction are improved.

The invention discloses an audio and video depth forgery detection method based on audio-visual asynchronous modeling guided by phonemes. The method comprises the following steps: extracting features by using an audio encoder and a video encoder respectively; phoneme embedding features are generated through a time-perceived phoneme extraction module and an embedding module; extracting mouth movement features through a feature alignment technology; designing a global time sequence asynchronous modeling method, and modeling frame-level global asynchronism between full face motion features and audio features; designing a local time sequence asynchronous modeling method, and modeling frame-level local asynchronism between the mouth motion feature and the phoneme embedding feature; designing a self-correlation calculation module of visual and audio modalities, and capturing forgery clues in the modalities; correlation matrixes output by the global time sequence asynchronous modeling method and the local time sequence asynchronous modeling method are fused and input into a classifier, and a detection result is output. The method gets rid of limitation of an existing method, explicit multi-mode joint modeling is achieved, and the method is suitable for audio-visual deep forgery detection in a complex scene.