54 results about "Discrete spectrum" patented technology

The application discloses a carrier aggregation method and a communication device, relates to the technical field of communication, and can improve the use rate of discrete spectrum. The method is as follows: first configuration information is sent, including a first CBW and a first BWP of a terminal device on a first carrier, the first CBW being the bandwidth of the first carrier; second configuration information is sent, including a second CBW and a second BWP of the terminal device on a second carrier, the second CBW being the bandwidth of the second carrier; the spectrum of the first BWP and the second BWP is discontinuous, the second carrier and the first carrier support combination into downlink in-band continuous CA; at least one of the first carrier and the second carrier satisfies the following conditions: the first carrier includes an operator spectrum and a non-operator spectrum, and the first BWP is an available BWP on the operator spectrum; the second carrier includes the operator spectrum and the non-operator spectrum, and the second BWP is an available BWP on the operator spectrum. The embodiment of the application is used for carrier aggregation.

Embodiments of this application relate to the communications field, and disclose a parameter configuration method, to resolve a problem of how to determine a parameter of a second-type bandwidth part BWP in a case of a discrete spectrum. A solution includes: for the parameter of the second-type BWP, determining a parameter determining rule based on a parameter type of the second-type BWP, where the second-type BWP includes N frequency domain resources, N is a positive integer, and N is greater than or equal to 2; and determining a parameter value of the second-type BWP according to the parameter determining rule. The embodiments of this application relate to a process of configuring the parameter of the second-type BWP.

The invention relates to the technical field of motor diagnosis, and discloses a motor eccentricity fault diagnosis method and system based on current harmonic analysis, and the method comprises the steps: carrying out the power frequency component extraction of a three-phase stator current signal, obtaining a residual current signal, carrying out the complex analysis envelope transformation of the residual current signal, and obtaining an envelope analysis signal; performing synchronous interpolation resampling on the envelope analysis signal to obtain a complex envelope signal; performing frequency spectrum estimation on the complex envelope signal to obtain a discrete frequency spectrum of a module value sequence; performing equalization processing on the order spectral line amplitude in the discrete spectrum to obtain a whitened spectrum, and performing inverse transformation on the whitened spectrum to reconstruct an angular domain amplitude signal of the target motor; performing cyclostationary analysis on the angular domain amplitude signal to obtain a third-order cyclic cumulant of the angular domain amplitude signal at a specific cyclic frequency; according to the third-order cycle cumulant, the type and severity level of the eccentric fault are judged; the accuracy of motor eccentricity fault diagnosis can be improved.

The invention discloses a mobile microphone array sound source positioning method based on a frequency spectrum correction theory. The method comprises the following steps: driving an N array element microphone array to move, presetting M sampling positions, monitoring the real-time position of the array, and establishing a sound field model; when the array moves to a preset sampling position and pauses to be stable, recording the sampling starting time, and performing synchronous sampling to obtain a sound source signal; sampling at all preset positions is completed, M * N sampling signals are obtained, discrete spectrum analysis and correction are carried out, and frequency domain data and correction frequency are obtained; taking the data of a certain sampling position as a reference, and performing phase compensation on the data of other positions by using the correction frequency to obtain a virtual microphone array; and extracting an azimuth angle and a pitch angle corresponding to a peak value by using a beam forming algorithm, namely, a sound source orientation. According to the method, the microphone array is driven to move to realize the dense microphone array, the sound source positioning cost is reduced, and the positioning accuracy is improved by introducing a frequency spectrum correction theory.

The invention discloses a spectrum super-resolution method and system based on an FP cavity and deep learning, and the method comprises the steps: S1, carrying out the initialization of spectrum collection, and selecting light with a specific wavelength after the light is acted with a target detection object; s2, performing photoelectric conversion and digitization to generate low-resolution original discrete spectrum data of corresponding wavelengths; s3, continuous spectrum scanning is carried out, and original discrete spectrum data collection of the whole target wave band is completed; and S4, performing super-resolution spectrum reconstruction, and judging whether cyclic reconstruction is performed or not based on a reconstruction data quality evaluation result until the quality is qualified. According to the method, an accurate mapping model from a low-resolution spectrum to a high-resolution spectrum is established through a pre-trained deep neural network, and an original spectrum which is collected by an FP cavity and is limited by the theoretical limit optical resolution of the FP cavity can be reconstructed into a super-resolution spectrum of which the resolution remarkably exceeds the theoretical limit. The problem that the resolution of an FP cavity in a long-wave region is sharply reduced is effectively solved, and the full-wave-band uniform high resolution is realized.

The invention discloses a satellite-ground convergence network information age optimization method based on multi-agent layered deep reinforcement learning, and relates to the technical field of communication, and the method comprises the steps: S1, initializing network parameters; s2, initializing # imgabs0 # and # imgabs1 #, and setting T, Ns, F, t = 1; s3, if t% Ns is equal to 0, executing S4; otherwise, executing S5; s4, in the upper-layer DQN, selecting actions for the current state of each agent by utilizing a greedy strategy, obtaining state transition based on discrete spectrum allocation, storing the state transition into # imgabs2 #, and updating theta and theta'by utilizing # imgabs3 # samples; s5, for each agent, inputting a lower layer observation state to PPO, obtaining state transition based on continuous power distribution, storing the state transition into # imgabs4, updating zeta and # imgabs6 by using a # imgabs5 sample, and if t% F is equal to 0, updating an old strategy network; s6, if tlt; if T is greater than T, t = t + 1, and executing S3; otherwise, ending iteration. And finally, finding a spectrum and power joint allocation scheme for minimizing the information age.

The present invention discloses a spectrum semantic communication system for sparse data completion and radiation source positioning, including four basic modules: semantic extraction, channel encoding and decoding, semantic recovery, and task completion. The system aims to cope with the challenges posed by large-scale spectrum data to the transmission performance of traditional communication systems and alleviate the problem of spectrum scarcity. First, the semantic extraction module extracts discrete semantic information from sparse spectrum data to reduce the extracted features as much as possible without affecting the task accuracy. Secondly, in order to combine the semantic communication system with the traditional digital communication system, the discrete spectrum semantics are converted into a bit stream through the channel encoding and decoding module and restored at the receiving end. Then, in order to complete the subsequent tasks, the semantic recovery module designs a fully connected network and uses nonlinear regression to restore continuous spectrum data. Finally, in order to complete the radiation source positioning task, the sparse spectrum map is completed into a complete spectrum map through an automatic encoder, and a convolutional network is designed to output the positioning result. Simulation results show that under the premise of greatly reducing the amount of data transmission, the scheme can achieve similar task completion effects compared with the traditional scheme at a higher signal-to-noise ratio, and is far better than the traditional scheme at a lower signal-to-noise ratio, thereby improving the transmission efficiency of spectrum data.

The application provides a method and device for evaluating the harmonic interference degree of a converter transformer by using a vibration signal. The method comprises the following steps: collecting multiple vibration signals collected by multiple vibration sensors as vibration samples; obtaining the frequency spectrum of each vibration sample; constructing a discrete frequency spectrum sequence of the vibration sample by using the amplitude of a preset reference frequency and its multiple frequencies in the frequency spectrum of the vibration sample; identifying the effective local peak value of the discrete frequency spectrum sequence of each vibration sample; calculating the energy concentration degree feature and the energy distribution entropy feature of the vibration sample based on the effective local peak value; and finally evaluating the harmonic interference degree of the converter transformer at the current moment according to the energy concentration degree feature and the energy distribution entropy feature of each vibration sample. The scheme evaluates the harmonic interference degree of the converter transformer by using the vibration signal, thereby preventing the converter transformer from malfunctioning due to excessive harmonic interference.

The invention belongs to the technical field of distributed optical fiber sensing, and relates to an anti-distortion distributed optical fiber sensing method and system based on array phase spectrum demodulation, an ultra-weak reflection optical fiber Bragg grating array is used as a sensing optical fiber, a series of reflection peaks with the signal-to-noise ratio higher than that of RBS are generated, a reflection peak sequence is formed, and the beat frequency signal spectrum amplitude is increased. A discrete beat frequency signal with frequency domain distortion output by an OFDR sensing system is converted to a frequency domain through Fourier transform of a symmetric window function to obtain a corresponding discrete frequency spectrum, and characteristic parameters such as amplitude and phase of a reflection peak sequence in the frequency spectrum are tracked and extracted; and then frequency correction factors of the reflection peak sequence in all measurement periods are calculated by using an amplitude interpolation correction method to obtain a corrected accurate array phase spectrum, and phase difference and strain calculation are performed to realize multi-point quasi-distributed strain demodulation for suppressing frequency domain distortion such as Doppler frequency shift and the like. The frequency resolution limitation caused by the fence effect is broken, and the fidelity of the frequency domain distortion signal is recovered.

The application discloses a star-ground fusion network information age optimization method based on multi-agent layered deep reinforcement learning, relates to the technical field of communication, and comprises the following steps: S1, initializing network parameters; S2, initializing and setting T, Ns, F, and t=1; S3, if t%Ns=0, executing S4; otherwise, executing S5; S4, in the upper-layer DQN, the current state of each agent is selected by using a greedy strategy to select an action, a state transition based on discrete spectrum allocation is obtained, and the state transition is stored in a sample to update theta and theta'; S5, for each agent, an observation state of the lower layer is input to PPO, a state transition based on continuous power allocation is obtained, and the state transition is stored in a sample to update zeta and if t%F=0, an old policy network is updated; S6, if t

The invention discloses a method for determining structural stress range distribution and related equipment, which output stable and reliable structural stress range distribution by optimizing a frequency domain discrete and time domain reconstruction mode and combining a multi-round statistical optimization and fusion strategy. The method comprises the following steps: acquiring a stress power spectrum of a target structure, randomly dividing frequency intervals in a preset frequency band and distributing phases to obtain a discrete spectrum; reconstructing a time domain stress time calendar and carrying out rain flow counting, and generating multiple groups of dimensionless rain flow stress range probability density distributions; and after average fairing and moment keeping correction, candidate sub-distribution is generated by adopting a multi-fairing strategy, and final distribution is obtained by combining evaluation index weighted fusion. According to the method, the time calendar authenticity is improved through random discretization, distribution convergence is realized through multi-time statistical fusion, and smoothness and statistical characteristics are taken into account through moment maintenance and multi-strategy fusion. Only stress power spectrum input is needed, the method is suitable for complex scenes such as broadband and multi-peak spectrum, the method can be embedded into an existing system for automatic operation, analysis cost is reduced, and fatigue evaluation precision is improved.

The application relates to the technical field of motor diagnosis, and discloses a motor eccentric fault diagnosis method and system based on current harmonic analysis. The method comprises the following steps: extracting a power frequency component of a three-phase stator current signal to obtain a residual current signal, and performing complex analytic envelope transformation on the residual current signal to obtain an envelope analytic signal; performing synchronous interpolation resampling on the envelope analytic signal to obtain a complex envelope signal; performing spectrum estimation on the complex envelope signal to obtain a discrete spectrum of a modulus sequence; performing equalization processing on the order spectrum line amplitude in the discrete spectrum to obtain a whitened spectrum, and inversely transforming and reconstructing the whitened spectrum into an angular domain amplitude signal of a target motor; performing cyclic stationary analysis on the angular domain amplitude signal to obtain a third-order cyclic cumulative quantity of the angular domain amplitude signal at a specific cycle frequency; and determining the type and severity level of the eccentric fault according to the third-order cyclic cumulative quantity. The application can improve the accuracy of motor eccentric fault diagnosis.

This invention discloses a method for noise optimization in discrete spectrum nonlinear frequency division multiplexing (DSP) systems using geometric shaping, belonging to the field of communication technology. The method includes: considering ASE noise and establishing a channel model under noisy conditions; based on the channel model under noisy conditions, designing an olive-shaped 32QAM in the DSP system; and by optimizing the geometric distribution of constellation points, mapping the optical signal to the point distribution of the olive-shaped 32QAM, reducing the probability of high-power points and minimizing the impact of accumulated spontaneous emission noise (ASE) and phase noise, thereby improving the signal's noise immunity and system transmission stability.

The invention relates to a digital transmitter (10) for providing a high-frequency analog signal, the transmitter comprising: conversion means (3) for receiving a digitally modulated baseband signal (12) and providing a discrete spectrum of the digitally modulated baseband signal (12) in polar coordinates or in oblique coordinates; and a synthesis means (90) for generating the high-frequency analog signal on the basis of the discrete spectrum provided in polar coordinates or in oblique coordinates, where the synthesis means (90) is a means for performing a continuous inverse Fourier transform. The invention also relates to a method for providing a high-frequency analog signal, comprising: receiving a digitally modulated baseband signal; and providing a discrete spectrum of the digitally modulated baseband signal in polar coordinates or in oblique coordinates; and generating the high-frequency analog signal based on the discrete spectrum provided in polar coordinates or in oblique coordinates by performing a continuous inverse Fourier transform.

The invention discloses a precise micro-vibration environment simulation method based on an active vibration isolation technology. A passive vibration isolation unit is used for supporting the platform, a table top vibration sensor is used for sensing vibration of the platform, and an actuator is used for outputting control force and serving as an active excitation source at the same time; based on any target micro-vibration standard, quantizing the target micro-vibration standard into a discrete frequency spectrum, and reconstructing a time domain vibration excitation signal conforming to the frequency spectrum characteristic; a vibration signal of the mass center position of the platform is obtained through calculation of sensors at all positions of the platform; designing a feedback controller according to the system model, and calculating to obtain a feedback control signal by using the difference between the platform mass center vibration signal and the vibration excitation signal; and converting the final feedback control signal into a corresponding output acting force by using an actuator. According to the invention, excitation of any target micro-vibration standard can be realized, and micro-vibration environment simulation far lower than ground vibration can be realized by adding an active vibration isolation technology.

The embodiment of the invention relates to the technical field of signal processing, and discloses a signal processing method, a storage medium, an integrated circuit, a device and terminal equipment, and the method comprises the steps: carrying out the discrete spectrum analysis of a to-be-processed signal based on an iterative compressed sensing technology, and determining a target frequency component; amplifying the target frequency component, and judging whether the current iteration is converged or not based on the amplified target frequency component; if the current iteration converges, performing frequency domain inverse transformation on the amplified target frequency component to obtain a target signal; and based on the target signal, performing signal recovery, velocity ambiguity resolution and / or direction of arrival estimation on the to-be-processed signal. According to the signal processing method provided by the embodiment of the invention, the number of iterations can be reduced on the premise that the cost is not increased, the convergence speed of the algorithm is effectively improved, the balance between the hardware cost and the calculation performance is realized, and meanwhile, the method is suitable for the signal with a very large missing sampling point proportion and has very high universality.

The present invention discloses a high-speed target detection method based on a planar interferometric photoelectric imaging system, the method comprising the following steps: step 1, designing parameters of the planar interferometric photoelectric imaging system; step 2, calculating spectrum acquisition data of the planar interferometric photoelectric imaging system and analyzing its imaging performance; step 3, constructing a detection data model of the sampling spectrum of the planar interferometric photoelectric imaging system; step 4, designing a frequency-space domain target detection algorithm; step 5, comparing and analyzing the target detection speed in the frequency-space domain, and then optimizing the system parameters. The present invention uses the sparse discrete spectrum of the observation scene acquired by the planar interferometric photoelectric imaging system to construct a data model for high-speed target detection, which has the advantages of being able to directly perform target detection in the frequency domain, having a small amount of detection model data, and having a fast target detection speed.

The present application relates to the field of optical fiber sensing technology, and more particularly to a kind of FP-FBG fusion spectrum demodulation method based on discrete spectrum analysis, comprising: driving tunable laser quasi-continuous scanning FP-FBG sensor and obtaining discrete reflection spectrum sequence through photoelectric acquisition circuit, determining the index of FBG peak in the sequence and obtaining FBG center wavelength by using centroid method fitting, selecting target neighborhood with the index as center and replacing the neighborhood sequence value with the intensity mean of spectrum sequence to obtain processed FP spectrum sequence, performing autocorrelation transformation on the processed FP spectrum sequence to obtain autocorrelation sequence, and finding the peak-to-peak distance between main peak and adjacent peak in the autocorrelation sequence, and finally calculating the sensing demodulation result according to the calibration relationship.The present application can effectively remove the shielding of narrow-band FBG to FP spectrum period information, suppress random noise and significantly improve the demodulation accuracy of low-quality spectrum.

The invention discloses an infrared spectrum detection method, device and system for wearable equipment. The device can be implemented as an integrated circuit chip, a module or a chip-module combination. The device comprises a multi-discrete SWIR light source driving and interface unit, an InGaAs detection reading unit, a time division scanning control unit and a synchronous demodulation processing unit. The control unit sequentially drives the multiple discrete SWIR light sources to emit according to a preset time division scanning sequence, and modulates emitted light in each time slot; the InGaAs detector collects skin reflection or back scattering signals, the synchronous demodulation unit executes related demodulation or phase-locked demodulation, amplitude features of all discrete wavelengths are extracted, and discrete spectrum feature vectors are formed and output. The scheme is suitable for wearable equipment, and has the advantages of resisting ambient light interference, being capable of expanding multi-emission position or multi-distance measurement and facilitating chip and modular integration.

The application belongs to the field of deep learning and remote sensing image processing, and specifically discloses a hyperspectral few-shot classification network construction method and a hyperspectral feature classification method, which comprises the following steps: acquiring hyperspectral remote sensing images and sample label data thereof; performing cross-domain reconstruction and spectrum curve extraction, and filtering common information between categories through eigenvalue decomposition to obtain discrete spectrum curves; extracting physical invariance features and taking them as constraint rules to generate virtual samples; inputting the reconstructed hyperspectral data and spectrum self-supervised auxiliary information into a double-branch variational autoencoder network to perform multi-loss constraint of cross reconstruction, so that the hyperspectral data and the spectrum self-supervised auxiliary information of the same category show greater similarity in the latent space; and outputting final feature prediction results based on a multinomial logistic regression classifier to complete the construction of the hyperspectral few-shot classification network. The application can realize high-precision and high-robustness hyperspectral feature classification under the condition of sample scarcity.

The application belongs to the field of fiber lasers, and particularly relates to a soliton pulse interval regulation method based on nonlinear Fourier transform, which comprises sequentially performing intensity normalization and nonlinear Fourier transform on full-field information of multiple soliton pulses to obtain nonlinear spectrum, and extracting each eigenvalue and corresponding discrete spectrum therefrom; identifying the eigenvalue set according to the size of the imaginary part of the eigenvalue to obtain a discrete spectrum set corresponding to the soliton pulse; determining the discrete spectrum corresponding to each soliton pulse according to the intensity size relationship of the multiple soliton pulses; regulating the modulus of the discrete spectrum corresponding to the soliton pulse according to the current interval and the required interval of any two soliton pulses in the time domain to realize the movement of the soliton pulse in the time domain and achieve the required interval; and performing inverse nonlinear Fourier transform on the nonlinear spectrum obtained after regulation to obtain the soliton pulse after interval regulation. The application guarantees the quality of the signal while regulating the interval of the soliton pulse.

This application belongs to the field of satellite communication technology and discloses a frequency usage situation reconstruction method based on a spectrum knowledge graph. The method includes: Step 1, deploying a low-Earth orbit satellite constellation and collecting spatiotemporally non-uniformly discrete spectrum data; Step 2, constructing a spectrum knowledge graph based on the collected data, and representing the multidimensional associations of the spectrum situation through "entity-relationship-attribute" triples; Step 3, extracting graph structure features from the knowledge graph; Step 4, fusing the graph structure features with runaway attributes and inputting the result into an improved machine learning model for situation reconstruction; Step 5, outputting the potential spectrum situation at unobserved times. This application significantly improves the accuracy and robustness of situation reconstruction by mining deep associations in spectrum data through knowledge graphs, and reduces the dependence on continuous observation.

The invention provides a self-convolution window design method for electric energy quality monitoring. The self-convolution window design method is used for improving the detection precision of electric power harmonic signals. According to the method, firstly, an objective function is designed, the objective function is determined by a sidelobe asymptotic attenuation rate and a sidelobe peak level, and the characteristics of a convolution window are flexibly controlled by adjusting the weight between the sidelobe asymptotic attenuation rate and the sidelobe peak level. Then, a solution space meeting constraint conditions is quickly searched by utilizing a genetic algorithm, and then, generated candidate solutions are finely optimized by utilizing an interior point method, so that the superiority of the solutions is improved; based on an optimal solution obtained by a hybrid optimization algorithm, a discrete frequency spectrum mode sequence of a convolution window is obtained through derivation, a target frequency band is corrected by applying a double-spectral-line interpolation algorithm, and accurate detection of a power harmonic signal is realized. According to the invention, a novel convolution window function can be designed autonomously, the problem that a traditional window function is insufficient in capability of detecting different signals is overcome, and the precision and accuracy of signal detection are improved.

The application belongs to the field of optical measurement, and specifically discloses a spatial spectrum acquisition method and system of a spectral confocal sensor illumination system, which comprises determining the initial spatial position of a fiber acquisition point, obtaining discrete spectrum distribution data of an illumination space; obtaining the light intensity maximum value of a target wavelength in different height directions, combining to generate the chief ray of the target wavelength and determining the focusing position of the target wavelength; obtaining the spectrum focal plane position of each wavelength in space; multiplying the spectrum focal plane position and the step number of each direction, and combining with the initial spatial position to obtain the actual coordinate value of the focal plane in the sensor coordinate system, determine the first mapping relationship of the wavelength and the height, determine the three-dimensional mapping relationship of the wavelength, the pixel and the height at the focal plane position according to the first mapping relationship and the second mapping relationship, and perform performance analysis on the spectral confocal sensor based on the three-dimensional mapping relationship. The application can improve the performance analysis efficiency and accuracy of the line spectrum confocal sensor.

The invention relates to a Bayesian filtering-based passive synthetic aperture algorithm and a storage medium. The Bayesian filtering-based passive synthetic aperture algorithm comprises the following steps of: 1, preprocessing data, reading array element data, framing N pieces of array element data to obtain L frames of data, and performing windowing FFT (Fast Fourier Transform) on each frame of data of each array element to obtain a discrete frequency spectrum of a broadband signal of each array element; step 2, broadband phase normalization: due to the fact that each discrete frequency point of the discrete frequency spectrum of each array element broadband signal has one phase and the phases are related to corresponding frequency point serial numbers, the average phase psi n of the time period is calculated; according to the method, Doppler information caused by motion and space phase difference information between array elements can be utilized at the same time, so that the estimation performance is improved.

This invention discloses a time-difference localization method for frequency-hopping radiation sources based on the maximum discrete spectral value, relating to the field of frequency-hopping radiation source localization technology. The method includes the following steps: constructing a model of the frequency-hopping signal received by the observation station based on the geometric positions of the observation station and the target radiation source; establishing an approximate expression for the frequency-hopping signal based on the discrete spectral characteristics of the signal; generating an objective function related only to the radiation source position using the target radiation source position information contained in the received signal; and achieving direct localization of the frequency-hopping radiation source through an exhaustive search algorithm. This method is advantageous for direct localization, provides accurate localization, and possesses strong versatility and scalability.

The embodiment of the invention relates to the technical field of signal processing, and discloses a signal processing method, a storage medium, an integrated circuit, a device and terminal equipment, the method is applied to signal recovery of a to-be-processed signal, and the method comprises the following steps: carrying out interference detection on the to-be-processed signal, and removing an interfered part in the to-be-processed signal; carrying out discrete spectrum analysis on the to-be-processed signal of which the interfered part is removed, and determining the to-be-processed signal of which the frequency component is greater than a first threshold value as a first frequency component; and on the basis of a signal obtained by performing frequency domain inverse transformation on the first frequency component, replacing an interfered part in the signal to be processed to obtain a recovered signal. According to the signal processing method provided by the embodiment of the invention, the interfered part in the to-be-processed signal can be subjected to signal recovery with low cost, high efficiency and high quality, and the recovered signal does not have a false alarm phenomenon.

This application discloses a method and related equipment for determining the stress range distribution of a structure. By optimizing the frequency domain discretization and time domain reconstruction methods, combined with multi-round statistical optimization and fusion strategies, a stable and reliable structural stress range distribution is output. The method includes: acquiring the stress power spectrum of the target structure, randomly dividing frequency intervals within a preset frequency band and assigning phases to obtain a discrete spectrum; reconstructing the time domain stress history and counting rainflow to generate multiple sets of dimensionless rainflow stress range probability density distributions; after average smoothing and moment preservation correction, using a multi-smoothing strategy to generate candidate sub-distributions, and combining them with evaluation indicators for weighted fusion to obtain the final distribution. This application improves the realism of the time history through random discretization, achieves distribution convergence through multiple statistical fusions, and balances smoothness and statistical characteristics through moment preservation and multi-strategy fusion. Only the stress power spectrum is required as input, making it suitable for complex scenarios such as broadband and multi-peak spectra. It can be embedded into existing systems for automated operation, reducing analysis costs and improving fatigue assessment accuracy.

The invention discloses a self-adaptive convolution wavelength demodulation method and system based on a reference filter, and the method comprises the steps: obtaining original waveform data collected by an optical fiber temperature sensor, and carrying out the secondary Gaussian filtering of the original waveform data, and obtaining a smooth discrete spectrum; obtaining the noise level of the original waveform data, and dynamically adjusting the convolution kernel of the convolution operation according to the noise level; acquiring a reference spectrum under the same acquisition condition as the original waveform data based on a long-wave-pass filter, and performing convolution operation on the reference spectrum and the smooth discrete spectrum in combination with the adjusted convolution kernel to obtain a drift sequence and a convolution sequence set; and performing curve fitting through the drift sequence and the convolution sequence set to obtain a temperature-wavelength drift corresponding curve, and obtaining a corresponding temperature value according to a wavelength detection value. By introducing the adaptive filtering method, the parameters of the convolution kernel are dynamically adjusted, different noise environments are effectively adapted, the wavelength demodulation precision is improved, and the robustness and accuracy of the demodulation method are effectively improved.