96 results about "Frequency encoding" patented technology

The invention discloses a three-dimensional vision calibration and positioning method and system based on an event camera, and belongs to the field of positioning. The method comprises the steps of constructing an array containing a plurality of unique frequency coding light sources, constructing a binocular event camera system, collecting light source array light signal event streams, extracting pixel positions, frequencies and timestamp information, determining pixel coordinate centroids of the light sources through frequency and pixel clustering analysis, and establishing frequency and centroid mapping. Auxiliary light sources with the same coding rule are arranged on the surface of a target object, and after a binocular system collects light signals of the auxiliary light sources, the center-of-mass coordinates of the light sources are matched according to the mapping relation. And based on event camera calibration parameters and binocular external parameters, correcting and stereoscopically matching centroid coordinates, and calculating three-dimensional coordinates of the auxiliary light source through triangulation. And in combination with the known geometric topological relation of the auxiliary light source on the surface of the object, the spatial pose of the target object is solved by using a rigid transformation algorithm, and high-precision positioning is completed. According to the embodiment, the calibration precision can be improved in a complex application environment.

The invention relates to the technical field of geological exploration, and particularly discloses a geological structure automatic identification and interpretation method based on deep learning, and the method comprises the steps: obtaining seismic data, logging data, gravity data and / or magnetic data, and generating a fusion tensor through space-time alignment and dynamic weight distribution; performing space-frequency domain joint coding on the fusion tensor by using an improved Transform encoder, and generating multi-scale geologic feature representation; a generative adversarial network is constructed, a generator of the generative adversarial network outputs a geological structure pseudo label, and a discriminator and a geological expert rule base collaboratively optimize a prediction result; and quantizing a confidence interval of a prediction result based on Bayesian deep learning, and combining geological priori knowledge to generate a dynamic interpretation scheme. Through multi-source data fusion, an innovative algorithm and model collaboration, accurate recognition and scientific interpretation of the geological structure are achieved, the accuracy and efficiency are improved, the cost is reduced, and the method has remarkable advantages in complex geological exploration.

The invention discloses a digital human facial expression synchronization method and device and a medium. A target video is divided to obtain a corresponding video set, and the video set is preprocessed to obtain a standard video set; extracting audio features and video features based on an audio and lip motion synchronization discriminator to calculate cosine similarity, and realizing conversion of the audio features and the video features in the same parameter space by combining a binary cross entropy loss function; performing three-dimensional shape model modeling on the face according to the standard video set, obtaining a three-dimensional space rotation translation matrix, a camera observation direction and an expression coefficient of each frame of face image and a shape coefficient of a face three-dimensional model, and obtaining an average face shape and key point coordinates; and according to the key point coordinates and the camera observation direction, obtaining spatial geometric features of multi-resolution Hash coding, combining audio features to obtain conditional features, and based on the conditional features and high-frequency coding information, constructing and rendering volume density and color to generate an audio-driven digital human facial expression image.

The invention discloses a ground feature classification method based on a frequency-space collaborative learning hierarchical fusion network. The method comprises the following steps: step 1, data preparation; step 2, the HMFE module extracts frequency domain features; step 3, extracting spatial domain features by an MLSA module; 4, the AWF module dynamically integrates the features; 5, executing a classification task; the MLSA module strengthens cross-scale interaction through a tree fusion structure and a cross attention mechanism, realizes efficient fusion of HSI and LiDAR data in a spatial domain, improves classification consistency of complex scenes, and reduces boundary blur phenomena; the HMFE module introduces a learnable frequency coding mechanism into the Mmba module so as to enhance spectrum-frequency components with strong discrimination; the AWF module realizes dynamic integration of spatial domain and frequency domain features through adaptive weighted fusion, fully mines deep complementarity of the spatial domain and the frequency domain, and improves the utilization efficiency of the model for heterogeneous data.

The invention discloses a signal control system for a mining locomotive, and aims to solve the problems of weak interference resistance and low safety of a traditional system. The system comprises a signal acquisition module, a dynamic frequency coding processing module, an intelligent decision-making module, an execution module, a communication module and a power supply module. The signal acquisition module acquires the position, speed and environmental parameters of the locomotive through an infrared position, speed and temperature and humidity sensor; the dynamic frequency coding module monitors the signal-to-noise ratio of a sub-band every 10 ms, when the signal-to-noise ratio of the sub-band is lower than 15 dB, the sub-band is automatically switched to a high-signal-to-noise-ratio sub-band larger than or equal to 20 dB, an encryption mapping table is generated for synchronous updating, and the equipment deals with sudden interference through an emergency band. The intelligent decision-making module generates a control instruction, and speed adjustment and parking are achieved through the execution module. An explosion-proof cable and a low-frequency through-the-earth technology are combined for communication, and a power supply is of an explosion-proof type and has multiple protections. The system improves the anti-interference performance and safety of signal transmission, adapts to the mine environment, and supports the intelligent operation of mine transportation.

The invention provides a method and system for constructing a knowledge representation learning model based on multi-dimensional information interaction and dynamic frequency perception. The method comprises the steps of constructing a multi-dimensional information encoder, constructing a historical query encoder and constructing a multi-dimensional perception decoder. According to the method, firstly, multi-dimension, entity type and time structure information in time knowledge graph reasoning is effectively modeled through a multi-dimensional information encoder; next, a historical query encoder aggregates their neighbouring semantics through constructed query-related historical facts, and additionally obtains duplicated global candidate facts through dynamic historical frequency coding, thereby providing global constraints for scoring in the decoder. Finally, the entity representation is decoded using a multi-dimensional perceptual decoder.

The invention discloses a fault diagnosis method based on a multi-modal cross-sensor Transform. The fault diagnosis method comprises the following steps: carrying out Fourier transform on input time domain data to obtain frequency domain data; calculating an embedded vector of the multi-modal data, and respectively inputting the embedded vector into a time encoder and a frequency encoder to extract features; enhancing the multi-modal features to generate two groups of different enhanced views; calculating time comparison loss, frequency domain comparison loss and time frequency comparison loss, and dynamically adjusting the weight of a loss function; finely adjusting the parameters of the classifier, and keeping the parameters of the encoder unchanged; and integrating outputs of different modes, and calculating a final classification result. According to the fault diagnosis method based on the multi-modal cross-sensor Transform, the problems of data integration and feature extraction in multi-modal cross-sensor fault diagnosis are solved through self-supervised learning and multi-modal data fusion, and the diagnosis accuracy and the model robustness are improved.

The invention relates to the technical field of welding data processing, and discloses a multi-sensor data fusion processing system in the welding process. The system comprises a multi-mode welding data acquisition module, a welding process feature portrait generation module, a feature preprocessing module, an environment coding module and a process parameter mapper. The multi-mode welding data acquisition module is used for acquiring welding current, voltage, a molten pool image, a welding seam shape, environment temperature and humidity and equipment vibration frequency; a feature portrait generation module generates a feature portrait containing welding quality and process parameter feature vectors according to the welding quality and process parameter feature vectors; the feature preprocessing module carries out time domain segmentation and frequency domain transformation on the multi-modal data to generate a standard welding process parameter data set; the environment encoding module encodes the environment temperature and the equipment vibration frequency to generate an environment state encoding sequence; and the process parameter mapper matches and maps the two parameters, and outputs a process parameter mapping characteristic value. According to the system, multi-modal data integration and process parameter dynamic matching are achieved, and the stability of the welding process is guaranteed.

The invention relates to the technical field of electromagnetic wave detection and imaging, and discloses a radar target echo generation method based on a physical information neural radiation field, which effectively improves the physical consistency of generated echo signals, accurately simulates the energy attenuation characteristics at different observation angles, and improves the radar target echo generation efficiency. The problem of signal completion under the sparse view angle condition is solved. The signal generated by the method is highly matched with measured data in the aspects of amplitude distribution, phase continuity and energy envelope form, a reliable data basis is provided for radar target identification and feature analysis, and the perception capability of a network on target geometric details is enhanced through high-dimensional position coding. Therefore, the generated echo signal can accurately reflect the spatial distribution characteristics of the scattering center. Meanwhile, due to the introduction of the multi-frequency coding sequence, the network can adaptively capture electromagnetic scattering characteristics under different scales, and a reliable technical means is provided for echo generation of a complex target.

The invention provides a delay determination method, device and system of a magnetic resonance system and a storage medium, and the method comprises the steps: carrying out the image collection of a target object based on a preset echo interval, and obtaining a first echo phase image, a second echo phase image and a third echo phase image, the first echo phase image and the third echo phase image are acquired based on a frequency coding gradient magnetic field of the first polarity; determining a first phase difference image corresponding to a single echo interval based on the first echo phase image and the third echo phase image; superposing the first echo phase image and the first phase difference image to obtain a reference phase image corresponding to the second echo phase image; and determining delay information of the magnetic resonance system based on the reference phase image and the second echo phase image. Based on the scheme of the invention, the delay information of the magnetic resonance system can be accurately determined.

The invention discloses a strain clamp X-ray image enhancement and defect identification method and system, and belongs to the technical field of image processing. The method comprises the following steps: acquiring an internal image of a wire clamp through X-rays, and performing feature reconstruction and defect modeling in a frequency space and a potential space; a multi-scale frequency encoder is adopted to extract detail features of a tension clamp image, a frequency domain structure selection module is combined to enhance defect edge response, and finally robust identification of an abnormal region is realized in a potential diffusion process. Therefore, good defect detection precision and boundary definition can be kept under the complex conditions of strong background noise interference, low image contrast and the like.

This invention provides a method for correcting the localization errors caused by field inhomogeneity and obtaining position information of a wireless marker in an MRI scanner. In one embodiment, said method comprises the steps of: a) Providing said wireless marker in said workspace; b) Performing a pulse sequence with said MRI scanner, comprising: i) Exciting all spins inside said workspace; ii) Applying a pre-phasing gradient along a frequency encoding; iii) Applying dephasing gradients in two directions orthogonal to said frequency encoding direction; and iv) Applying two readout gradients with same amplitude and opposite polarities along said frequency encoding direction to form a first and a second gradient echo; c) Acquiring a first and a second MR signal using an array of RF receiver coils by sampling said first and second gradient echoes respectively; and d) Reconstructing tracking signals from said first and second MR signals and calculating said position information.

The present invention discloses a method and apparatus for generating frequency-encoded pulse data for use in a spiking neural network. The method comprises converting the time-encoded pulse data of the spiking neural network through one or more conversion functions to obtain frequency values, and generating the frequency-encoded pulse data of the spiking neural network based on the frequency values. The present invention advantageously supports more than one pulse coding format, provides the ability to integrate multiple coding formats, and provides compatibility with multiple coding processes.

The present invention relates to a DAC-based pulse radar multi-waveform signal generation method, which belongs to the field of signal processing. The method can flexibly set various signal parameters, adopts a fully digital method to design and generate pulse radar signals in various relatively complex environmental scenarios, and realizes the generation of various pulse radar signal waveforms in the application environment, including frequency modulation (conventional frequency modulation, frequency diversity, frequency agility), frequency repetition rate (PRI) modulation (conventional, staggered, jitter, sliding, jumping), and intra-pulse modulation (intra-pulse single tone, linear frequency modulation, nonlinear frequency modulation, phase encoding, frequency encoding, and hybrid encoding). The method can simultaneously combine various pulse radar waveforms, dynamically adjust according to the DAC working data bit width, and generate corresponding signals according to the signal waveform output mode. The method has obvious advantages in the simulation generation of pulse radar signals in electromagnetic environments.

The invention discloses a composite modulation signal blind identification method based on time-frequency analysis, which comprises the following steps of: firstly, judging and classifying input pulse signals, and then respectively carrying out identification and parameter estimation on determined signal types. According to the scheme, down-conversion analysis processing, signal time-frequency analysis processing and signal classification identification decision processing are firstly carried out on AD sampling intermediate frequency data sent by a digital signal processing board FPGA to obtain an identification result of a radar signal intra-pulse modulation type, and then corresponding parameter estimation is carried out according to the radar signal intra-pulse modulation type. According to two composite modulation styles of frequency coding and linear frequency modulation combination and frequency coding and two-phase coding combination, on the basis of time-frequency parameter characteristics of various composite modulations, related characteristic parameters are accurately extracted, a modulation type identification process is constructed, and time-frequency analysis is performed on radar signals through fast Fourier transform and phase correlation, so that the identification accuracy of the radar signals is improved. And furthermore, intra-pulse parameter estimation is accurately carried out on the radar signal.

This invention discloses a method for distinguishing and quantifying the enhanced oil recovery (EOR) mechanism of CO2 huff and puff in unconventional oil reservoirs, relating to the field of oil and gas extraction technology. The method is based on nuclear magnetic resonance (NMR) experiments to determine the T2 spectrum and one-dimensional frequency coding of saturated oil cores. By analyzing the T2 spectrum and one-dimensional frequency coding of saturated oil cores, the CO2 huff and puff recovery rate and the effective CO2 reach distance are calculated. Based on the experimental data, a component model is established and calibrated to fit the experimental recovery rate and effective CO2 reach distance, thereby distinguishing and quantifying the EOR mechanism at different stages of CO2 huff and puff. This method can accurately classify the enhanced oil recovery mechanism during CO2 huff and puff into oil expansion, oil molecule diffusion, elastic energy release, and dissolved gas drive, effectively identifying the main controlling mechanism of CO2 huff and puff enhancement, providing a basis for optimizing field operations.

The invention discloses a high-temporal-spatial-resolution magnetic resonance metabolism imaging acquisition and reconstruction method, which comprises the following steps of: constructing a pulse sequence based on prior knowledge, and performing image acquisition, reconstruction and compound concentration correction by using the pulse sequence based on the prior knowledge; the prior knowledge-based pulse sequence comprises a plurality of balanced steady-state free precession bSSFP sequences, a pulse phase difference is added between radio frequency pulses corresponding to each step of phase coding in each balanced steady-state free precession bSSFP sequence, and the values of the pulse phase differences corresponding to different balanced steady-state free precession bSSFP sequences should be different. According to the invention, a magnetic resonance imaging acquisition mode is adopted, and more efficient frequency coding is adopted in the reading direction Gx for space coding, so that the acquisition efficiency is improved, the signal-to-noise ratio of an image is higher, and the image is not easily influenced by a non-uniform magnetic field; the signal intensity value is converted into the concentration of the metabolic compound, and absolute quantification of the concentration of the metabolic compound is achieved.

An optical ranging and detection system enables a reconfigurable very wide field of view, high sampling of spatial points per second, and high optical power handling by effectively combining different wavelength, time and frequency encoding, and spatial selectivity using an architecture. A transmitter is capable of generating multiple narrow beams, encoding different beams, and transmitting in different spatial directions. A receiver can distinguish and extract range and reflectivity information of the reflected beams. Three-dimensional imaging of an environment is achieved by scanning the field of view of the transmitter. Control and signal processing electronics fabricated in a chip are packaged together with a chip containing photonic components of the ranging system. The optical ranging and detection system generates terahertz beams in addition to optical beams, and the combined two beams allow reconfigurable spectral analysis.

Edge computing units that are outfitted with signal receivers and provided at local sites or edge locations are configured to capture RF signals and process the RF signals using transformer models. A spectrum encoder of a transformer model performs attention on frequency-encoded features extracted from spectrograms of the RF signals and features extracted from representations of RF signals associated with discrete RF events, and embeddings generated by the spectrum encoder are decoded to classify the RF signals or a source of the RF signals. Additionally, a spectrum decoder of a transformer model is configured to receive a query for information and to generate a response to the query using a spectrum decoder that performs attention on embeddings received from a spectrum encoder, as well as frequency-encoded features representative of the query, and features representative of semantic information regarding RF signals and sources of the RF signals.

The invention relates to the technical field of mechanical fault detection, and particularly provides a fault diagnosis method and device, a medium and a computer program product, and the method comprises the steps: obtaining an operation state signal of a mechanical device; preprocessing the operation state signal to obtain operation characteristic data of the mechanical equipment, and performing frequency coding on the operation characteristic data to obtain a peak sequence to be processed; inputting the to-be-processed peak sequence into a liquid machine LSM model, processing the to-be-processed peak sequence based on a pulse time sequence dependent plasticity STDP algorithm, updating the weight between neurons in the LSM model, and outputting a to-be-identified peak sequence; and inputting the to-be-identified peak sequence into the convolutional neural network model for feature identification to obtain a fault diagnosis result of the mechanical equipment. According to the invention, the accuracy and efficiency of fault diagnosis can be improved.

The invention discloses a method for distinguishing and quantifying an unconventional oil reservoir CO2 huff and puff recovery ratio improving mechanism, and relates to the technical field of petroleum and natural gas extraction. According to the method, a saturated oil core T2 spectrum and a one-dimensional frequency code are measured on the basis of a nuclear magnetic resonance (NMR) experiment, and the CO2 huff and puff recovery ratio and the CO2 effective action distance are calculated by analyzing the saturated oil core T2 spectrum and the one-dimensional frequency code; a component model is established and calibrated based on experimental data so as to fit the recovery ratio and the CO2 effective action distance in the experiment, and then EOR mechanisms of different CO2 huff and puff stages are distinguished and quantified. According to the method, the yield increasing mechanism in the CO2 huff and puff process can be accurately divided into oil expansion, oil molecule diffusion, elastic energy release and dissolved gas drive, the main control mechanism of CO2 huff and puff yield increasing can be effectively judged, and a basis is provided for optimization of field operation.

A system for RF based frequency encoding utilizing a Bloch-Siegert shift, includes a controller, an RF encoding system, and an injection transformer simultaneous transmit and receive filter. The controller generates RF excitation pulses, RF based frequency encoding pulses, and a cancellation signal. The RF encoding system includes one or more RF coils configured to transmit the RF excitation pulses and RF based frequency encoding pulses, and to receive an MR signal from the subject where the MR signal includes a leakage signal induced by the RF based frequency encoding pulses. The injection transformer simultaneous transmit and receive filter is in signal communication with the controller and the RF encoding system. The injection transformer simultaneous transmit receive filter is configured to receive the cancellation signal, the MR signal including the leakage signal, and to cancel the leakage signal from the received MR signal to generate a filtered MR signal.

The present invention relates to a hybrid brain-computer interface method that combines visual and auditory weak stimulation. Its technical characteristics are: determining the video stimulation interface, the number of visual stimulation targets and the visual stimulation attributes, and realizing the visual weak stimulation function by means of high-frequency flickering of the peripheral visual field; realizing the auditory weak stimulation function by applying binaural frequency-divided low-sound-pressure-level auditory stimulation signals to the left and right ears; performing visual weak stimulation and auditory weak stimulation at the same time, and selectively paying attention to the auditory stimulation of the left or right ear to complete the task of the hybrid paradigm; while performing visual weak stimulation and auditory weak stimulation, extracting the steady-state visual evoked potential characteristics and auditory steady-state response characteristics of the user's electroencephalogram signal and performing identification and classification to generate coding instructions. The present invention combines the visual weak stimulation method with the auditory weak stimulation method, and uses two electroencephalogram signals for encoding at the same time, thereby improving the naturalness and friendliness of human-computer interaction, and can encode more targets using fewer frequencies.

Aspects of the present disclosure include methods for characterizing particles of a sample in a flow stream. Methods according to certain embodiments include generating frequency-encoded fluorescence data from a particle of a sample in a flow stream; and calculating phase-corrected spatial data of the particle by performing a transform of the frequency-encoded fluorescence data with a phase correction component. In certain embodiments, methods include generating an image of the particle in the flow stream based on the phase-corrected spatial data. Systems having a processor with memory operably coupled to the processor having instructions stored thereon, which when executed by the processor, cause the processor to calculate phase-corrected spatial data from frequency-encoded fluorescence data of a particle a flow stream are also described. Integrated circuit devices (e.g., field programmable gate arrays) having programming for practicing the subject methods are also provided.

The invention discloses a lumbar vertebra scanning processing method and device, a storage medium and electronic equipment, and the method comprises the steps: obtaining a sagittal three-dimensional image of a lumbar vertebra, carrying out the three-dimensional segmentation of the lumbar vertebra and an intervertebral disc in the sagittal three-dimensional image, and obtaining a lumbar vertebra segmentation result and an intervertebral disc segmentation result; inputting the lumbar vertebra segmentation result and the intervertebral disc segmentation result into a preset anatomical key point detection network for anatomical key point identification to obtain an anatomical key point set, and constructing a local vector space based on the anatomical key point set, according to a preset magnetic resonance lumbar vertebra scanning constraint rule, target scanning positioning parameters corresponding to multiple target scanning orientations are matched in the local vector space, and the target scanning positioning parameters at least comprise a phase coding direction, a frequency coding direction, a slice selection direction and a slice center position; and controlling magnetic resonance equipment to execute lumbar vertebra scanning based on the target scanning positioning parameter. The lumbar vertebra scanning parameter planning can be automatically completed.

Methods for determining parameters of particles in a flow stream from scattered light (e.g., in a particle analyzer of a flow cytometer) are described. Methods according to certain embodiments include illuminating particles in a flow stream with a frequency-modulated laser beam modulated at a reference frequency, detecting scattered light from the particles with a photodetector, generating a frequency-encoded data signal from the detected scattered light, synchronizing the frequency-encoded data signal to the reference frequency, and determining one or more parameters of the particles based on the synchronized frequency-encoded data signal. Systems for practicing the subject methods and non-transitory computer-readable storage media having instructions for practicing the subject methods are also provided.

The application provides a non-destructive size measurement method for underground targets by combining frequency coding and half-integer phase control, and belongs to the field of non-destructive size measurement of underground targets. The method solves the problem that the prior art cannot effectively overcome the near distance blind area and suppress the frequency domain sidelobe. The method comprises the following steps: constructing an equal-interval frequency grid; generating a transmission frequency sequence by using a step parameter coprime with the total frequency point number for coding sorting; defining the effective transmission duration of each frequency point time slot, so that the near distance blind area range is determined by meeting the half-integer period condition; in each frequency point time slot, a linear sweep signal is generated, and the cumulative phase of the signal at the end of the time slot is controlled to meet the half-integer circle constraint; the signal is transmitted in sequence, and the echo signal is received synchronously; the complex frequency response of each frequency point is extracted, and after frequency domain coherent synthesis, the phase information is differentially approximated to obtain the group time delay; and the target geometric size is calculated according to the medium propagation speed and the group time delay. The method is used in the field of pipe network detection.

The invention discloses a social relation analysis method based on an autocorrelation mechanism, and the method comprises the steps: firstly constructing a continuous dynamic network based on a social network, extracting first-order neighbors of all nodes in the continuous dynamic network, and obtaining the neighborhood information of the nodes; secondly, according to node neighborhood information, obtaining information about nodes in all aspects through node coding, edge coding, time coding and frequency coding, and converting the information into a time sequence; then, correlation between node interaction is obtained through an autocorrelation mechanism for a time sequence, and node representation is obtained through self-adaptive reading; and finally, training through stochastic gradient descent to obtain each node representation in the social network, and analyzing the social network through the obtained node representation to predict the probability of interaction between users. According to the method, the traditional self-attention method is exceeded by identifying the dependency relationship and aggregating the information at the sequence level, and accurate social relationship analysis is carried out.

An optimal pulse power measurement method for multinuclear simultaneous integrated magnetic resonance imaging includes the following steps: successively or simultaneously exciting multiple nuclides within a slice in a same repetition time (TR) of magnetic resonance imaging (MRI); applying, in a slice encoding gradient channel, a frequency encoding gradient in a direction opposite to a slice selection gradient, and acquiring free induction decay (FID) signals of all the nuclides; simultaneously changing shaped radiofrequency (RF) pulse power values of all the nuclides for multiple times, and acquiring multiple corresponding FID signals for each nuclide; and performing Fourier transform on the FID signals of each nuclide, selecting a spectral peak in an absolute spectrum, integrating the spectral peak for comparison, and taking a shaped RF pulse power value corresponding to a maximum signal intensity of each nuclide as an optimal shaped RF pulse power corresponding to a current TR.

The invention relates to the technical field of artificial wetland pollution management, in particular to an artificial wetland pollution prediction method which comprises the following steps: retrieving literatures to obtain related parameters; performing frequency coding and feature scaling on the parameters to obtain a data set; dividing the data set into a training set and a test set; establishing a predictor model based on a stacking model, wherein the predictor model comprises a base learner and a meta learner; respectively training the models of the base learner by adopting the training set; obtaining a prediction result of the base learner as an input of the meta learner; using a genetic algorithm to optimize the prediction weight of the base learner model; performing weighted fusion on prediction results of the three models of the base learner, predicting the test set again, and calculating a final performance index; and measuring the deviation and prediction performance of the predicted value of the constructed wetland. According to the method, the problems of limited prediction precision and poor stability of a traditional single model are solved.