32 results about "Subspace decomposition" patented technology

The invention belongs to the technical field of crossing of information and ocean science and technology, and discloses a UUV broadside parallel array covariance weighted fusion DOA estimation method. The method comprises the following steps: firstly, calculating auto-covariance and cross-covariance matrixes of two sub-arrays, carrying out weighted fusion on covariance matrixes of different sources through differential common-array expansion of a virtual aperture, then reconstructing a complete covariance matrix by respectively adopting a convex optimization method of trace regularization and nuclear norm regularization according to the characteristics of the fused matrix, and finally, carrying out weighted fusion on the covariance matrixes of different sources. The pitch angle and the azimuth angle of the target are jointly estimated by performing subspace decomposition on the constructed extension matrix, and the effectiveness of the method is verified by a simulation result. According to the method, the precision of parameter estimation is improved through data fusion, the degree of freedom is remarkably improved under a small number of physical array elements, high-precision azimuth estimation of the UUV on multiple targets in the underwater environment is achieved, the target detection capacity of the UUV is enhanced, and the method has high engineering application value.

The invention discloses a six-phase motor driver harmonic current suppression method based on multi-subspace control, and belongs to the technical field of motor control. In order to solve the problem that low-order harmonic current of a six-phase motor in a z1-z2 subspace is difficult to effectively suppress, the invention provides a scheme that a mathematical model of the six-phase permanent magnet synchronous motor is established, multi-subspace decomposition is carried out, and stator current is divided into a d-q subspace and the z1-z2 subspace; a double-closed-loop control architecture with a speed loop as an outer loop and a current loop as an inner loop is constructed, a d-q subspace adopts a PID regulator to realize fundamental current control, and z1-z2 subspaces realize suppression of fifth and seventh harmonic currents through a parallel structure of PI and a quasi-proportional resonance controller; and a PWM signal is synthesized through inverse transformation and dual three-phase space vector pulse width modulation, and an inverter is driven to realize stable output of fundamental wave energy and harmonic suppression. The method effectively reduces the harmonic distortion rate, improves the operation efficiency and stability, and is suitable for occasions of new energy vehicles, ship propulsion, rail transit and the like.

The invention provides a digital quantum simulation method, which comprises the following steps of: performing symmetry analysis and subspace decomposition on a target Hamiltonian to obtain an effective Hamiltonian and a symmetry constraint condition; constructing a target unitary operator according to the effective Hamiltonian, and initializing a quantum circuit population of the target unitary operator; calculating entanglement characteristics of each quantum line in the quantum line population, performing structure variation on the quantum line population based on the entanglement characteristics, and performing line fragment crossing in combination with a symmetry constraint condition to obtain a candidate line set; performing line parameter optimization on the candidate line set based on the fidelity gradient to obtain an optimized line set, and extracting gate distribution structure features of the optimized line set; and non-uniform sampling is carried out on the optimized line set based on gate distribution structure characteristics to obtain a sampling line subset, an optimal quantum line is screened out from the sampling line subset based on the deviation between an experimental observation value and a theoretical expected value, and quantum simulation is carried out based on the optimal quantum line.

The invention discloses a data processing method and system based on Lie algebraic dynamics and an entropy modulation mechanism, and relates to the technical field of artificial intelligence and data processing, and the method comprises the steps: constructing a group action operator on a physical manifold, and carrying out the manifold enhancement of input data; the Hilbert subspace decomposition and decoupling of the feature channel are completed in parallel through a learnable decorrelation kernel and tensor contraction operation by using a matrix parallel elastic orbit layer; mapping the features to a symplectic geometric phase space, and introducing an adaptive symplectic respiration operator and a Hamiltonian kinetic equation to carry out physical conservation evolution; constructing a Lie algebra classifier, and determining a data category by calculating a Lie bracket transpose norm of a feature generator and a category base; and guiding model training by using an entropy modulation action quantity loss function containing physical potential energy. The method can solve the problems of poor generalization and weak anti-noise capability of a traditional deep learning model, and can be widely applied to the fields of power equipment monitoring, computer vision, natural language processing and the like.

The invention discloses a self-adaptive subspace projection suppression method based on optical pump magnetometer external noise, and belongs to the technical field of neural magnetic imaging (MEG) signal processing.The method comprises the steps that brain magnetic signals are collected through a wearable optical pump magnetometer array and preprocessed, and a signal and noise subspace decomposition model is established; integrating a model-driven noise modeling method and a data-driven noise modeling method, and constructing a noise subspace base vector; constructing a signal subspace basis vector in combination with a guide field matrix and event correlation analysis; identifying residual noise components based on event correlation analysis, and perfecting noise subspace modeling; signal noise separation is realized by adopting a hierarchical projection strategy; and analyzing the time correlation of internal and external space signals, and constructing a time projection matrix to remove residual interference to obtain a clean MEG signal. According to the invention, various types of external noise interferences can be effectively suppressed, the signal-to-noise ratio and the reliability of brain magnetic signals are remarkably improved, and the method is particularly suitable for practical application of a wearable optical pump magnetometer system.

The invention discloses a subspace joint parameter estimation algorithm for constructing an analytic signal based on Hilbert transform in high-precision multi-surface phase-shift interference, belongs to the technical field of high-precision optical interference measurement, and aims to solve the problems of inaccurate frequency estimation caused by spectrum leakage and a fence effect in a traditional algorithm. And the problem that a modern spectrum estimation algorithm cannot directly process real-value light intensity signals is solved. The method comprises the following steps: firstly, performing Hilbert transform on a real-value light intensity sequence to construct an analysis signal; then, a Hankel matrix is constructed based on the analysis signal, and subspace decomposition is carried out; further, taking the obtained high-precision frequency as prior information, and performing joint correction on the amplitude and the phase by using a double-section phase difference correction method; and finally, reconstructing the surface topography according to the corrected phase. According to the invention, the limitation of spectrum leakage and the like is overcome, the joint fine positioning of frequency, amplitude and phase is realized, and the precision, robustness and measurement repeatability of interference parameter demodulation are obviously improved.

The application discloses a method and system for generating hyperspectral images based on hyperspectral and multispectral image fusion. Y The method comprises the following steps: performing subspace decomposition to extract left low-rank vectors U of the hyperspectral image to be generated; Z The method comprises the following steps: performing subspace decomposition to extract left low-rank vectors U of the hyperspectral image to be generated; U And right low-rank vectors V of the hyperspectral image to be generated are generated through the trained deep generation network; V The left low-rank vectors U and the right low-rank vectors V are fused to generate the hyperspectral image X The system comprises optical elements and imaging sensors for acquiring input hyperspectral images and multispectral images. Y The system comprises optical elements and imaging sensors for acquiring input hyperspectral images and multispectral images. Z The application can generate hyperspectral images based on hyperspectral and multispectral image fusion, and reduce the cost and efficiency of hyperspectral image acquisition.

The application belongs to the technical field of information and ocean science, and discloses a UUV side parallel array covariance weighted fusion DOA estimation method. First, the self-covariance and mutual covariance matrices of two sub-arrays are calculated, then the virtual aperture is extended by difference co-array, the covariance matrices of different sources are weighted and fused, then, according to the characteristics of the fused matrix, the trace regularization and kernel norm regularization convex optimization methods are used respectively to reconstruct the complete covariance matrix, finally, the pitch angle and azimuth angle of the target are jointly estimated by subspace decomposition of the constructed extended matrix, and the simulation results verify the effectiveness of the method. The application improves the parameter estimation accuracy through data fusion, significantly improves the degree of freedom under a small number of physical array elements, realizes high-precision azimuth estimation of multiple targets in the underwater environment of the UUV, enhances the target detection capability of the UUV, and has high engineering application value.

This invention discloses a time-frequency cooperative Doppler frequency offset suppression method and system for underwater acoustic OFDM communication, belonging to the field of underwater acoustic communication. The method includes: designing a novel channel adaptive tracking resampling factor estimation method, which performs cross-correlation processing only between two specific synchronization signal segments of the received signal, enabling it to adaptively track the time-varying channel and significantly improving the robustness of the communication system in underwater acoustic environments; based on the estimated resampling factor, performing time-domain resampling operations to transform non-uniform frequency offset into uniform frequency offset, thereby eliminating the impact of time-domain scale transformation on the underwater acoustic signal; and designing a residual uniform frequency offset suppression method based on subspace decomposition, using frequency domain eigenvalue decomposition to finely suppress the residual uniform frequency offset of the underwater acoustic signal. Combining the advantages of time-domain resampling and frequency-domain eigenvalue decomposition, it performs excellently in suppressing inter-carrier interference of underwater acoustic signals in highly mobile scenarios, significantly improving signal detection accuracy.

This invention discloses a sparse Bayesian target azimuth estimation method based on far-field dictionary reconstruction under strong near-field interference. The method includes subspace decomposition of the observation data to construct a signal subspace projection operator; projecting and normalizing the far-field dictionary to obtain a projected reconstruction dictionary, which replaces the far-field dictionary in sparse Bayesian learning, and jointly performing sparse Bayesian iterative estimation with the near-field dictionary; finally, outputting the far-field target azimuth estimation result. This method effectively partitions the subspace and constructs a projection matrix accordingly to suppress the impact of strong near-field interference on far-field azimuth estimation, significantly improving the far-field target azimuth estimation capability of the sparse Bayesian method under strong near-field interference conditions.

The invention discloses an old and ancient book cover image retrieval method based on dynamic partitioning and multi-modal feature fusion, and relates to the technical field of image recognition, old and ancient book cover images are used as query vectors to be input into a retrieval model, and a matched book list corresponding to old and ancient books is obtained according to the query vectors; the training process of the retrieval model is as follows: performing visual feature extraction and text feature extraction on the cover image of the ancient and old book, and partitioning; inputting the blocked visual features and text features into a cross attention mechanism, and generating an original high-dimensional fusion feature vector including a layout text structure and image layout information; performing subspace decomposition and quantization coding on the original high-dimensional fusion feature vector by adopting a product quantization algorithm to obtain a compressed feature piece vector, and constructing an ancient and old book feature database according to the compressed feature piece vector; taking cover images of the old and ancient books as input query vectors, indexing the old and ancient book database by adopting a three-level progressive query architecture, and returning a matched book list; the trainable parameters of the retrieval model are jointly optimized by constructing a total loss function; according to the retrieval method, multi-level understanding and digital reconstruction of ancient book covers are achieved, and rapid retrieval is achieved through efficient index construction.

The invention discloses a communication fault positioning method based on comparative learning, and the method specifically comprises the steps: obtaining link state data, node operation data and alarm data, and building a unified time index to generate a communication state feature sequence; performing vector coding and normalization to generate an initial feature representation set; executing sample division and sample pair construction to generate a positive and negative sample pair set; subspace decomposition is executed to generate a fault feature subspace and an environment feature subspace, and a fault prototype vector set, a prototype drift sequence and a constraint feature representation set are generated; performing similarity calculation and contrast loss calculation to generate a gradient vector set, and performing projection decomposition on conflict gradients to generate a corrected gradient set; executing parameter updating according to the correction gradient to generate a communication state feature representation set; and calculating the distance between the feature representation and the fault prototype, determining a fault type identifier, and outputting a fault positioning result. According to the invention, multi-source fusion comparison positioning is realized, the precision is high, and the stability is strong.

This invention proposes a beamforming method for ISAC systems based on subspace decomposition and scalar optimization. The invention mainly includes the following steps: (1) deriving the Fisher information matrix for the target angle, time delay, radial velocity, and target reflection coefficient; (2) inverting and diagonalizing the Fisher information matrix to obtain the expression for the joint Cramé-Rao bound; (3) constructing a beamforming optimization problem with the goal of minimizing the joint CRB; (4) decomposing and simplifying the covariance matrix into a scalar optimization problem using the channel subspace characteristics; (5) providing closed-form optimal solutions for beamforming vectors for different antenna configurations; and (6) reconstructing the beamforming vectors based on the closed-form optimal solutions to obtain a transmit waveform that meets both sensing accuracy and communication requirements. This invention can significantly improve the parameter estimation accuracy for moving targets, effectively balance sensing and communication performance, and has low computational complexity, making it convenient for practical engineering deployment.

The invention discloses a decomposition deconvolution beam forming method based on subspace clustering, and belongs to the field of direction of arrival estimation in passive detection. According to the invention, the problem of poor target discovery capability of the existing deconvolution beam forming algorithm is solved. According to the spectral clustering-based subspace decomposition method provided by the invention, the covariance matrix can be adaptively decomposed into the strong signal subspace and the weak signal subspace according to the size of the characteristic value without estimating the number of information sources, so that the problem of array gain reduction of a traditional deconvolution beam forming algorithm in a strong interference scene is solved; and the target discovery capability is improved. According to the DOA estimation method provided by the invention, null constraint deconvolution beam forming and a signal-to-noise ratio-based fusion algorithm are introduced, so that energy leaked to weak signal subspace components by interference can be effectively suppressed under non-ideal conditions, the influence of strong interference on a weak target can be effectively suppressed, and the robustness of the algorithm and the target discovery capability are improved. The method can be applied to direction-of-arrival estimation in passive detection.

The invention relates to a fast open-loop GNSS-R height measurement method and system based on space-time frequency combination, and the method comprises the steps: collecting multi-frequency GNSS signals, carrying out the coarse time delay and coarse Doppler frequency offset estimation of each multi-frequency GNSS signal, and carrying out the compensation of the multi-frequency GNSS signals through coarse Doppler frequency offset; based on the coarse time delay and the compensated multi-frequency GNSS signal, carrying out forming processing on a wave beam in a time domain direction, and constructing a space-time-frequency joint vector; through a spatial domain smoothing technology, rank increasing is carried out on the autocorrelation matrix of the space-time-frequency joint vector, and the problem of rank deficiency of the autocorrelation matrix is solved; performing feature analysis on the autocorrelation matrix according to a subspace decomposition method, and estimating direct view path time delay, specular reflection path time delay and a direct view path incoming angle; and deducing a satellite elevation angle according to the direct-view path incoming angle, and inverting the height from the receiver to the ground surface through the delay inequality of the direct-view path and the mirror reflection path and the satellite elevation angle. According to the technical scheme of the invention, high-precision effective estimation of the height of the receiver can be realized quickly with low complexity.

The invention discloses a multi-modal sentiment analysis method based on progressive comparative learning, and belongs to the technical field of sentiment analysis, and the method comprises the steps: carrying out the feature extraction of text, audio and video modals in a multi-modal data set; separating generality and characteristics between modes, and performing subspace decomposition; optimizing the text by using the feature subspaces of the audio and the video, and carrying out multilayer progressive contrast learning to obtain optimized text features; enhancing subspace features of the audio and the video by using the optimized text features; feature stacking and dimension reduction are carried out, emotion weights are calculated, and emotion perception weighted fusion and prediction are carried out based on the emotion weights. According to the method, the problems of modal imbalance, noise interference, low feature quality, insufficient utilization of comparative learning and lack of emotion pertinence of a fusion strategy in the prior art are solved. According to the method, the feature discrimination is higher, modal imbalance and noise interference can be effectively relieved, the fusion process is more targeted, emotion clues can be accurately captured, and high-performance emotion analysis can be realized.

The invention provides an array signal processing method and device, electronic equipment and a storage medium, and relates to the technical field of radar signal processing, receiving signals of a non-uniform array are mapped into receiving data of a virtual uniform array, and a covariance matrix is reconstructed by using a neural network model based on the data, so that the array signal processing efficiency is improved. And finally, estimating the direction of arrival of the signal through subspace decomposition and polynomial rooting operation, so that the problems of low estimation precision and narrow application range caused by the fact that a non-uniform array structure does not meet application conditions of a traditional algorithm, covariance matrix estimation is greatly influenced by signal conditions and subspace discrimination is insufficient in the prior art can be solved. The technical effects of adapting to a non-uniform array scene, improving the accuracy and robustness of covariance matrix estimation under a wide signal-to-noise ratio, enhancing the distinction degree of a signal and a noise subspace, further improving the direction-of-arrival estimation precision and resolving power and having physical interpretability are achieved.

This invention proposes a fusion subspace DOA estimation method based on sparse Bayesian learning. The implementation steps are as follows: first, an overcomplete steering vector dictionary matrix is ​​constructed, and a sparse representation model is built. Then, the signal is reconstructed using the sparse Bayesian learning algorithm. After obtaining the reconstructed signal, a signal covariance matrix is ​​constructed. Finally, eigenvalue decomposition is performed on the covariance matrix to construct a spatial spectral function, and the peak value of this spectral function is searched across the entire angular range. This invention integrates the SBL algorithm and the MUSIC algorithm. The MUSIC algorithm is used to perform subspace decomposition on the reconstructed covariance matrix, extracting the noise subspace and utilizing its orthogonality with the steering vector, thereby improving the angular resolution of the SBL algorithm. This makes this invention applicable to achieving high-resolution DOA estimation under conditions of few or even single snapshots.

The invention discloses a super-resolution multipath time delay estimation method for Fourier precoding OCDM signals in a dynamic scene, and the method comprises the steps: carrying out the Fourier precoding of a pilot symbol, generating a precoding orthogonal linear frequency modulation wavelength division multiplexing signal through modulation, and eliminating the coupling effect between the time delay and Doppler frequency shift in a time-frequency double-selection channel; doppler frequency shifts of different paths are directly estimated according to the position offset of the received pilot signals in the Fresnel domain, and the number of the multiple paths is determined; and calculating a covariance matrix of the received signal, carrying out decoherence processing, carrying out subspace decomposition on the covariance matrix by using a multiple signal classification algorithm to obtain a spatial spectrum, and determining a multipath time delay estimation value according to a spectrum peak position of the spatial spectrum. The method is suitable for underwater acoustic mobile communication and environment perception with Doppler frequency shift, the relative root-mean-square error of time delay estimation can be remarkably reduced in a high-speed state, and a reliable solution is provided for super-resolution time delay estimation in a dynamic scene.

This invention discloses a guided process construction method based on industrial modeling, comprising the following steps: transforming the target industrial process construction problem into a Markov decision process model; generating intrinsic reward signals based on the novelty of the target industrial process path, historical access confidence, or model prediction error; assigning extrinsic reward signals to each stage or subtask of the target industrial process through a multi-layered extrinsic reward mechanism; inputting the intrinsic and extrinsic reward signals together as the optimization objective into a deep Q-network based on subspace decomposition; and automatically generating an industrial process structure and parameter configuration scheme that satisfies the predetermined optimization objective or process constraints based on the output of the deep Q-network based on subspace decomposition. This invention enables efficient and intelligent optimization and automatic modeling of industrial process schemes in sparse reward and high-dimensional complex process scenarios through a multi-layered reward mechanism and a deep Q-network based on subspace decomposition.

The application relates to a wide-band sensing method based on a MIMO-based MWC sub-Nyquist sampling structure, which comprises the following steps: a MIMO-based MWC wide-band sampling front end is used to model a signal in an mCSL matrix; a static covariance matrix is constructed by using the correlation between column vectors of the matrix; subspace decomposition is performed on the static covariance matrix to obtain eigenvalues corresponding to signals and eigenvalues corresponding to noises, and noise reconstruction is performed; according to the orthogonality of the signals and the noises, a calculation vector based on the mCSL matrix structure is obtained based on the reconstructed noises, and the calculation vector is sorted; support sets corresponding to the first K values of the calculation vector are used as negative frequency points where PU users are located, and all center frequency points of the PU users are obtained according to the symmetry of real signal spectrum. The application can improve the spectrum recovery effect.

This invention discloses an off-grid DOA estimation method and system for polar impulse noise environments, belonging to the field of polar underwater acoustic detection. The method includes: obtaining an array received data matrix based on signals received through a uniform linear array with half-wavelength spacing; initializing the maximum number of iterations, column full-rank random matrices, and row full-rank random matrices; iteratively solving the improved mixed correlation entropy of the residual fitting error matrix as the objective function for subspace decomposition optimization based on the maximum mixed correlation entropy criterion to obtain the optimized array received data matrix; and using an iterative sparse projection algorithm to jointly estimate the sparse signal matrix and grid offset based on a new off-grid model constructed using multi-order Taylor expansion terms, thereby estimating the target's DOA. This invention overcomes the grid mismatch problem to achieve high-precision estimation of target location and exhibits good performance in both Gaussian noise and impulse noise environments.

The application discloses a structure settlement visual measurement method and system based on deep learning compensation, and relates to the technical field of building settlement measurement. The method comprises the following steps: according to a mobile detection platform, driving a detection component to collect a structure region and determining structure data; constructing a settlement measurement component by spatial decomposition and implicit causal analysis of settlement dimensions, motion dimensions and interference dimensions; performing background domain segmentation on the structure data, executing subspace decomposition feature recognition and implicit causal coupling analysis, and determining settlement data pairs; performing directional projection and three-dimensional mapping through a three-dimensional projector, realizing three-dimensional distance measurement, and generating and updating a settlement curve display. The technical problems of the existing structure settlement measurement, such as the difficulty in effectively correcting the spatial displacement measurement error caused by attitude change and environmental interference, and the insufficient distance measurement accuracy, are solved, the technical effect of realizing high-precision non-contact distance measurement and dynamic compensation of structure settlement displacement is achieved, and the stability of spatial distance measurement and the environmental adaptability are improved.

The invention discloses a multi-radiation source target direct tracking method, which is characterized in that a plurality of spatially distributed receiving arrays perform data acquisition on an unknown number of targets, each receiving array is composed of a plurality of array elements, and snapshot data are synchronously acquired and received, so that a receiving signal model under a multi-array collaborative observation condition is established. For the model, a plurality of spatial covariance matrixes are constructed, noise subspace information of the spatial covariance matrixes is extracted, and a MUSIC spatial spectrum is formed in combination with a subspace decomposition theory. According to the method, a spatial spectrum value is used as an observation likelihood function, and a generalized Labeled Multi-Bernoulli (GLMB) filtering framework is embedded, so that joint estimation of the number and the state of multiple targets is realized. And finally, performing propagation and weighted updating on the state particle set by combining a sequential Monte Carlo method to obtain the position of the target.

The invention discloses a multi-angle one-dimensional scattering center three-dimensional reconstruction method, and belongs to the technical field of radar signal processing. According to the method, through a scattering center dimension raising method combining subspace decomposition, RANSAC, spherical projection and visibility screening, a three-dimensional scattering center space structure of a target is recovered on the basis of a multi-angle one-dimensional distance scattering center; and the three-dimensional scattering center space distribution of the target can be deduced only by depending on multi-angle one-dimensional broadband distance domain scattering center estimation. According to the method, the problem that the three-dimensional scattering center cannot be recovered when only small-angle or discrete angular domain measurement is carried out is solved, and the availability of the three-dimensional scattering center in a non-cooperative observation scene is remarkably improved; the method has the advantages of high robustness and strong anti-noise and false scattering center capability. The method does not depend on continuous angle coverage, can only depend on non-missing data to realize three-dimensional scattering center inversion when part of angles are missing, and is especially suitable for non-cooperative targets such as maneuvering aircrafts, ships and satellites.

The invention provides an electrified equipment defect diagnosis method, system and equipment based on sound-heat and visible light data fusion. The method comprises the following steps: acquiring an acoustic signal, infrared thermal image data and a visible light image; performing real-time alignment of a space coordinate system based on visible light image features and infrared hot spot distribution, and establishing a mapping relation between a three-dimensional coordinate system and image pixel coordinates; suppressing environmental acoustic noise in the acoustic signal by using an adaptive beam forming algorithm based on feature subspace decomposition; performing dynamic atmospheric transmissivity compensation and radiation temperature measurement inversion on the infrared thermal image data according to the measurement distance and environmental meteorological parameters obtained in real time to obtain a corrected equipment surface absolute temperature field; generating a three-dimensional fusion temperature field reflecting the state of a medium in the equipment through an iterative algorithm; and inputting the preprocessed acoustic signal, the three-dimensional fusion temperature field and the visible light image into a trained multi-channel deep learning model so as to output a defect category, a three-dimensional coordinate and a risk index.

The invention relates to the technical field of new-generation mobile communication base station equipment and transmission, and discloses an RIS-MIMO signal transmission method based on adaptive beamforming, which comprises the following steps: a base station calculates a spatial covariance matrix according to uplink detection signals accumulated in a first preset time period, and executes subspace decomposition to extract slowly varying statistical characteristics; the base station retrieves a target code word based on a null space constraint criterion and drives a reconfigurable intelligent surface (RIS) to lock a phase so as to construct a static reflection channel; in the RIS locking period, the base station performs digital pre-coding on the data stream to be transmitted and transmits the data stream based on the equivalent channel state information containing the channel, and the method utilizes a dual-time-scale decoupling mechanism, configures the RIS through long-period statistical characteristics to construct a stable physical transmission boundary, and solves the problems of channel aging and pilot frequency overhead in a fast fading scene.

The application provides a subspace-assisted multi-prior photon counting spectral CT reconstruction method and device, and relates to the technical field of CT reconstruction. The method first obtains a noise feature map tensor and an orthogonal basis through subspace decomposition on a spectral CT image; then a reference image block is selected on the noise feature map tensor, and similar image blocks to the reference image block are extracted from all channels of the spectral CT image to generate a non-local similar full-channel tensor group; a global sparse regularization term about the non-local similar full-channel tensor group is constructed by using dictionary learning to obtain a denoised non-local similar full-channel tensor group, and the denoised non-local similar full-channel tensor group is multiplied by the orthogonal basis to obtain a clean tensor image; then a double regularization constraint reconstruction model is constructed according to TV regularization constraints of each channel of the clean tensor image and the spectral CT image; finally, the augmented Lagrange multiplier method is used to update the double regularization constraint reconstruction model until a preset reconstruction quality requirement is reached.

The invention discloses a structure settlement visual measurement method and system based on deep learning compensation, and relates to the technical field of building settlement measurement. The method comprises the following steps: driving a detection assembly to collect a structure area and determine structure data according to a mobile detection platform; a settlement measurement assembly is constructed through spatial decomposition and recessive causal analysis of the settlement dimension, the motion dimension and the interference dimension; performing background domain segmentation on the structure data, executing subspace decomposition feature recognition and recessive causal coupling analysis, and determining settlement data pairs; directional projection and three-dimensional mapping are carried out through a three-dimensional projector, three-dimensional distance measurement is achieved, and settlement curve display is generated and updated. The technical problems that in existing structure settlement measurement, space displacement distance measurement errors are difficult to effectively correct due to posture changes and environmental interference, and the distance measurement precision is insufficient are solved, and the technical effects of achieving high-precision non-contact distance measurement and dynamic compensation of structure settlement displacement and improving space distance measurement stability and environmental adaptability are achieved.

The invention discloses an electroencephalogram data signal monitoring and intelligent classification method based on deep learning, and the method comprises the following steps: collecting multichannel electroencephalogram signals, and carrying out band-pass filtering, notch filtering, re-reference and amplitude normalization to obtain preprocessed electroencephalogram signals; performing sliding window segmentation on the preprocessed electroencephalogram signal to obtain an electroencephalogram window sequence; constructing a stability analysis data set; sSA stationary and non-stationary double sub-space decomposition is executed to obtain a stationary sub-space projection matrix and a non-stationary sub-space projection matrix; obtaining a stable representation sequence, inputting the stable representation sequence into a deep learning classification model, and outputting a classification result and classification confidence; projecting the electroencephalogram window sequence to obtain a non-stationary representation sequence, and judging whether online self-adaptive updating is triggered or not; when online updating is triggered, SSA statistical characteristics are updated, the minimum updating time interval is set, and if yes, rollback is carried out, and alarm information is output. The stability and reliability of electroencephalogram monitoring and classification are improved.