110 results about "Fourier spectrum" patented technology

The invention relates to the technical field of bolt length detection data processing, in particular to a bolt length detection system, which is characterized in that a three-dimensional data acquisition module generates a high-density point cloud and extracts a thread curvature extreme point, and a dynamic model matching module screens an optimal lead angle parameter based on a Hausdorff distance; and constructing an axial expansion model carrying phase offset gradient features. The projection mapping correction module adopts a genetic algorithm to optimize a compensation coefficient matrix, and generates a dynamic weight correction parameter to suppress geometric distortion; and the phase error correction module combines Fourier spectrum decomposition and an exponential decay weight coefficient to construct a frequency domain transfer function to eliminate an axial accumulative error. According to the method, the problems of projection distortion and phase deviation caused by the lead angle are effectively solved through geometric-phase double-domain correction, dynamic parameter iteration and an error collaborative suppression mechanism, and the spatial positioning precision of the feature points at the tail end of the thread is improved.

A method for processing a radar signal based on a photonic fractional Fourier transformer comprises: transmitting a linear frequency modulation signal to targets to be detected, receiving echo signals of the targets to be measured, and loading the linear frequency modulation signal and the echo signals onto a single-frequency optical wave by an electro-optical modulator (S1); respectively biasing a sub-modulator and a parent modulator of the electro-optical modulator at different bias points, modulating the single-frequency optical wave by the electro-optical modulator based on the linear frequency modulation signal and the echo signals, and outputting a modulated optical signal (S2); converting the modulated optical signal by a photoelectric detector to a photocurrent (S3); and performing Fourier transform on the photocurrent to obtain a fractional Fourier spectrum, and obtaining distance information of the targets to be measured according to peak positions of each pulse signal in the fractional Fourier spectrum (S4).

The invention relates to a method for measuring the coherence length of a Shell model correlation structure light beam, which belongs to the field of laser space coherence measurement, and comprises the following steps: S1, calculating the coherence lengths # imgabs0 # and # imgabs1 # in x and y directions by using the analytic relationship between the coherence length of the Shell model correlation structure light beam and the spatial coherence modulus; s2, a double-aperture mask is arranged on a light beam shaping transmission path of the laser processing system, the aperture positions are symmetrical, the distance d is adjustable, and after light beams are focused, intensity signals # imgabs2 are collected at a focal plane detector; and S3, extracting an intensity signal # imgabs3 # to obtain a double-aperture interferogram, performing Fourier transform on the interferogram, establishing a corresponding relation between a low-frequency component in a frequency spectrum and a characteristic frequency of an aperture spacing d, and calculating to obtain a spatial coherence modulus # imgabs4 #. According to the invention, based on partially coherent light dual-aperture interference and Fourier spectrum analysis, a direct analytical relationship between the spatial coherence degree and the coherence length is established, and a new way is provided for beam quality control in laser processing.

The invention discloses a self-adaptive underwater single-pixel imaging technology based on Fourier modulation, belongs to the technical field of computational optical imaging, is used for ocean detection and improvement of ocean surveillance and defense efficiency, and adopts a self-designed LED array to replace an expensive DMD and a four-quadrant detector composed of common photodiodes to replace an expensive barrel detector. And high-speed illumination and parallel acceptance are realized while the cost is greatly reduced. The technical route comprises a high-speed underwater single-pixel imaging device, a high-speed Fourier single-pixel imaging algorithm and an adaptive sampling strategy based on Fourier single-pixel imaging. The high-speed single-pixel imaging device comprises a self-designed LED array light modulation module, a converging lens module and a receiving module. The high-speed Fourier single-pixel imaging algorithm comprises a binarization algorithm based on error diffusion and an algorithm based on adaptive compressed sensing. According to the innovative self-adaptive sampling strategy based on Fourier single-pixel imaging, a specific sampling path is generated by fusing circular sampling and random sampling according to the characteristics of an image in a Fourier spectrum, so that Fourier coefficients which have great influence on image quality are acquired as much as possible under water with low sampling rate and high turbidity, and the image quality is improved. Therefore, the high-efficiency imaging of the underwater target object is realized, and the imaging speed and quality are both considered.

The invention provides an automatic substance identification method based on an optimized Raman spectrum analysis system. The method comprises the following steps: performing parameterization and noise reduction on Raman spectrum data through a Fourier spectrum method, and converting high-dimensional data into limited frequency components; then, local nonlinear changes and noise characteristics in data are captured through Gaussian process modeling, and adaptive updating of model parameters is achieved in combination with Bayesian reasoning and multiple times of sequential optimization. In order to further improve the recognition precision, the system adopts characteristic wavelength screening (such as CARS, BOSS and iVISSA) and a spectrum preprocessing technology in data preprocessing to remove redundant information; and finally, integrating the results of a plurality of optimization models through a model fusion technology, and improving the robustness of overall prediction. The whole process forms a closed-loop automatic system with data preprocessing, modeling optimization and real-time feedback, the problems of high calculation complexity and insufficient prediction precision when a traditional method is used for processing complex and high-noise Raman spectrum data are effectively solved, and the method is suitable for multiple fields such as chemistry, environmental monitoring, pharmacy and food detection.

The invention provides a radar echo noise reduction method, device and equipment and a computer readable storage medium, and relates to the technical field of radar noise reduction, and the method comprises the steps: carrying out the subtraction of a radar echo signal and a fitting signal corresponding to the radar echo signal, and obtaining a radar difference signal; determining the Fourier spectrum and the S spectrum of the radar difference signal, and performing band-pass filtering on the S spectrum to obtain a filtered signal; wherein the center frequency of the band-pass filtering is the average value of the frequency spectrum peak frequency of the Fourier frequency spectrum and the S frequency spectrum; decomposing the filtering signal to obtain a modal component of the filtering signal; and according to the fuzzy entropy and the autocorrelation coefficient of each modal component, determining a plurality of target modal components from each modal component, and synthesizing each target modal component into a radar echo noise reduction signal. According to the technical scheme provided by the invention, the radar echo signal can be subjected to noise reduction.

The invention discloses a ship sheet welding part deformation monitoring method and system. According to the method, controllable elastic disturbance is introduced into a welding area through a phased array sound wave excitation module, and tiny response of the surface of a welding part under the action of thermal stress is excited; a multi-frequency interference light intensity image of a welding area is obtained through a laser interference measurement module, phase responses under all sound wave frequencies are extracted through Fourier spectrum analysis, and a multi-frequency comprehensive disturbance amplitude graph is obtained; and in combination with temperature gradient information acquired by the thermal infrared imager, a disturbance direction vector field is constructed, and then a two-dimensional disturbance vector field of the welding area is generated. An abnormal deformation area possibly generated in the welding process is recognized in real time by constructing a judgment early warning mechanism, and visual early warning information is output. The method has the technical characteristics of non-contact, high precision, real-time performance and multi-dimensional cooperation, and is suitable for intelligent deformation monitoring and early risk early warning in the ship thin plate component welding process.

The invention discloses a high-altitude hydropower engineering multi-source heterogeneous data fusion sharing method and system. The method comprises the steps of collecting and storing multi-source data, wherein the multi-source data comprises structured data, BIM data, GIS data, video picture data and text data; performing data pre-check on the multi-source data to ensure that the data format and quality meet related requirements; preprocessing the structured monitoring data by adopting a moving average and Fourier spectrum analysis method; constructing a knowledge graph to model a semantic relationship among the multi-source heterogeneous data; and in combination with a large language model LLM and a retrieval enhancement generation RAG technology, constructing a data sharing interactive question-answering system based on the knowledge graph. According to the invention, the problems of efficient integration, management and sharing of various data types in the whole life cycle of hydropower engineering can be solved.

The invention relates to the technical field of fault early warning, in particular to a vibrating screen fault early warning method and system.The vibrating screen fault early warning method comprises the steps that operating parameters of a vibrating screen during ore screening are obtained, the operating parameters comprise sound characteristics and vibration characteristics, and the operating condition of the current ore is determined; and according to the difference value between the real-time operation parameter of the current ore under the operation condition and the standard operation parameter under the same condition. According to the method, historical screening data of different ores are clustered, operation conditions are recognized, Fourier spectrum analysis of sound and vibration signals is combined, the sensitivity degree of the operation conditions to fault features is calculated, and differential weights are given. In the real-time operation process, according to the current ore type and working condition, weighted comparison is carried out on the collected signals and standard parameters, the fault degree is calculated, if the fault degree exceeds a set threshold value, sound-light alarm is triggered, and through the method, the early warning accuracy can be improved, and the sudden fault risk can be reduced.

Embodiments of the present application relate to the technical field of computer, and particularly relate to a training method and device of a silent image retake detection model and related equipment. The method comprises: obtaining a predicted depth map and a predicted Fourier spectrum map of a sample face image through a pre-training network model; obtaining a pseudo depth map and a real Fourier spectrum map of the sample face image; processing the predicted depth map and the pseudo depth map according to a first loss function of the pre-training network model to obtain a first loss value; processing the predicted Fourier spectrum map and the real Fourier spectrum map according to the first loss function to obtain a second loss value; obtaining a third loss value corresponding to a second loss function of the pre-training network model; and training the model through the first loss value, the second loss value and the third loss value to obtain a silent image retake detection model. By using the above method, the loss of depth and high frequency information of single RGB channel image information in the model training process can be compensated, and the model prediction effect can be improved.

The application discloses a signal adversarial sample detection method based on Fourier spectrum, and belongs to the technical field of signal security of machine learning, and comprises the following steps: preparing a benign sample data set and an adversarial sample data set, respectively extracting Fourier spectrum features of the benign sample data set and the adversarial sample data set and respectively performing feature fusion, composing a feature data set and dividing a feature training set and a feature test set; constructing a detection network, performing binary classification training on the detection network by using the feature training set, and obtaining an optimal detection network model; and inputting the feature test set into the optimal detection network model and outputting a detection result. The data is processed by using discrete Fourier transform, the features of the benign sample and the adversarial sample are respectively extracted and fused, the method process is effectively shortened, and a complex full-connection network is used for classification, so that the features of the electromagnetic signal are more fitted, and the detection precision can be effectively improved.

The invention relates to the technical field of artificial intelligence, and provides a cross-domain flow field prediction method based on a neural operator and multi-target transfer learning, and the method comprises the steps: building a multi-input and multi-output general neural operator converter, combining the multi-input and multi-output general neural operator converter with the multi-target transfer learning, carrying out the multi-modal input coding and Fourier spectrum embedding through the steps S1 to S5, and carrying out the coding and embedding, thereby obtaining a multi-target flow field prediction result. The inputs of the physical domains are mapped to a shared potential feature space by their corresponding parameterized encoders. Coding features of all the domains are input into a shared computing layer, and features processed by the shared computing layer are mapped to physical quantity prediction of the domains through independent parameterization decoders of all the domains. The invention further discloses a system which achieves the technical effects. In the reasoning stage, the efficiency is improved, the flow field simulation time is greatly shortened, and rapid design iteration and parameter optimization are assisted. The method is not only suitable for airfoil flow field prediction, but also can be expanded to multi-scene engineering problems such as hydrodynamics and thermal management.

The invention relates to the field of aggregate morphology detection and image processing, and discloses an aggregate particle morphology quantification method and system based on Fourier spectrum decoupling, and the system comprises a three-dimensional point cloud collection module which is used for obtaining the three-dimensional point cloud data of aggregate particles; the point cloud data processing module is used for processing the three-dimensional point cloud data to obtain the two-dimensional contour data; the morphological feature analysis module is used for processing the two-dimensional contour data to obtain aggregate decoupling feature data; and the detection and analysis module is used for associating the modules. According to the method, the RGB-D depth point cloud system is used for collecting multi-view point cloud, and the complete three-dimensional morphology point cloud is obtained through noise reduction and registration. Through Monte Carlo simulation sectioning, two-dimensional equal circumference mapping, contour Fourier series fitting, gradient operator definition and the like, a method for performing decoupling calculation on the shape, the corner angle and the texture of the aggregate through aggregate form parameters is established, and the accuracy and the accuracy of aggregate three-dimensional form test analysis are improved.

The present invention discloses an image edge detection method, device and complex amplitude modulation module based on complex amplitude filtering topological space differential operation, relating to the technical field of optical information processing. The technical points of the present invention include: placing a two-dimensional image on the object plane of a 4f system composed of two Fourier transform lenses, and performing Fourier transform on the signal light beam carrying the two-dimensional image information through the first Fourier transform lens to generate the Fourier spectrum of the two-dimensional image on the spectrum plane of the 4f system; at the same time, placing a complex amplitude filtering device for realizing complex amplitude filtering topological space differential operation on the spectrum plane of the 4f system to generate a corresponding transfer function, so as to modulate the Fourier spectrum of the two-dimensional image; performing inverse Fourier transform on the Fourier spectrum modulated by the two-dimensional image through the second Fourier transform lens, and obtaining the result of image differential operation on the image plane of the 4f system to realize image edge detection. The present invention can be applied to aspects such as image edge detection and biological microscopic imaging.

This invention discloses a method, system, medium, and device for zero-speed detection of a wheeled robot. The method includes the following steps: acquiring inertial sensor data of the wheeled robot; selecting inertial sensor data based on a preset length sliding window, and calculating the modulus of the selected inertial sensor data to obtain one-dimensional inertial data; standardizing the one-dimensional inertial data; performing a discrete Fourier transform on the standardized data to obtain a Fourier spectrum; extracting time-domain feature data and frequency-domain feature data from the standardized data and the Fourier spectrum based on a preset one-dimensional convolutional neural network; and performing zero-speed detection on the time-domain feature data and frequency-domain feature data based on a preset fully connected neural network. Therefore, a high-performance zero-speed detection method that does not rely on a preset threshold is provided, improving the zero-speed detection capability of the wheeled robot in different motion modes and reducing the impact of engine and other component vibrations on the zero-speed detection effect when the robot is parked.

The invention discloses a periodic variable star light variable curve classification method based on a neural network, and relates to the field of astronomical photometric data classification and identification, and the method comprises the steps: employing an intelligent neural network composition principle, training a plurality of small networks, enabling each small network to be responsible for the identification of a variable star type, enabling all small networks to form a large intelligent neural network, and enabling all small networks to be responsible for the identification of a variable star type; and all variable star types are identified. According to the method, the amplitude of the harmonic wave in the Fourier spectrum of the optical variation curve is used as a characteristic value, supervised learning is performed on the small network, and the learning and identification efficiency of the whole network is improved by combining the symbol logic of the intelligent neural network. In addition, the method has good expandability, for example, when a to-be-recognized variable star type needs to be newly added in the future, only a corresponding small net needs to be added, the small net is added into a large net, and then a small amount of training is iteratively carried out.

The invention discloses a local wave number estimation damage optimal imaging method based on strong constraint Bayesian optimization, and relates to the technical field of material structure health monitoring, and the technical scheme is characterized in that the method comprises the following steps: obtaining a three-dimensional guided wave signal of the surface of a thin-wall structure, converting the three-dimensional guided wave signal to a frequency-wave number domain through multi-dimensional Fourier transform, and obtaining a three-dimensional guided wave signal of the surface of the thin-wall structure; a filtered Fourier spectrum is obtained; adaptively extracting a target modal filtering local wave number from the filtered Fourier spectrum to obtain a local wave number estimation value; and obtaining an imaging quality evaluation index and a constraint frequency band, introducing a strong constraint Bayesian optimization method to carry out optimal frequency search on a local wave number estimation value, and outputting a stable damage imaging result. According to the frequency selection and imaging optimization method disclosed by the invention, an optimization problem of changing depending on manual frequency selection into data driving, strong constraint, noise resistance and efficient sample is solved, the imaging quality and robustness of local wave number estimation method imaging in complex structure damage detection are improved, and an efficient frequency selection and imaging optimization method is constructed.

The application provides a bearing fault detection method and device and electronic equipment. The method comprises: obtaining a vibration signal of a bearing to be detected; transforming the vibration signal to a support domain of a normalized Fourier spectrum; segmenting the Fourier spectrum of the support domain according to a preset iterative forward-backward search algorithm to obtain a plurality of continuous frequency segment ranges; determining a mode component corresponding to a maximum fault characteristic energy index from the plurality of continuous frequency segment ranges as an optimal mode component sensitive to a fault characteristic according to a preset maximum fault characteristic energy index strategy; performing envelope demodulation analysis on the optimal mode component to obtain a detection result of the bearing to be detected, wherein when an envelope demodulation spectrum obtained by envelope demodulation exists, a detection result indicating that the bearing to be detected has a fault is obtained. The application is beneficial to improving the accuracy and reliability of bearing fault detection.

The invention discloses an omni-directional quantitative schlieren measurement device and method based on a 4f system and a radial gradual change filter, and belongs to the field of fluid experiment mechanics and optical precision measurement. The device is provided with a first lens, a linear radial gradual change filter, a second lens and a photosensitive element which are sequentially arranged along an optical path; parallel light emitted from the transparent medium field enters the first lens, and forms a Fourier spectrum surface on a rear focal plane after Fourier transform of the first lens; the filter plate is accurately arranged at the Fourier spectrum surface, and the transmissivity of the filter plate is linearly changed along a set direction; and the frequency spectrum light after intensity modulation is subjected to inverse Fourier transform by the second lens, and finally the light intensity distribution of the image surface is acquired by the photosensitive element. According to the invention, the refractive index, density and pressure distribution of the transparent medium flow field can be rapidly, accurately and quantitatively measured.

This invention provides a seismic intensity estimation system that can improve the accuracy of estimation when estimating the seismic intensity class based on the acceleration waveform obtained by the acceleration sensor of a smart device, while the smart device is placed on a stand on the floor. [Solution] The seismic intensity estimation system 1 includes: an acceleration waveform collection unit 23 that collects acceleration waveforms from each of a plurality of smart devices 10 during an earthquake; an acceleration waveform adjustment unit 24 that rotates the acceleration waveforms of the plurality of smart devices 10 so that their orientations in the horizontal plane coincide with each other to generate orientation-adjusted acceleration waveforms; a frame vibration characteristic derivation unit 25 that derives the vibration characteristics of at least one frame based on the orientation-adjusted acceleration waveforms of at least one pair of smart devices 10; and a floor seismic intensity class estimation unit 26 that derives the Fourier spectrum of the floor based on the acceleration waveform on the frame and the vibration characteristics of the frame to estimate the seismic intensity class of the floor.

This invention relates to a structural damage diagnosis method based on iterative calibration image measurement, belonging to the field of structural damage diagnosis. The method specifically includes the following steps: dividing the engineering structure into multiple substructures, each substructure containing at least three measurement points, with at least 2 / 3 of the measurement points in two adjacent substructures overlapping. Based on the characteristic relationship between the upper and lower structures, the homography matrix H is used to correct the upper structure to the same plane as the lower structure. The overall modal parameters of the structure are extracted using the relative displacements obtained from the tests, and damage identification is performed by combining this with the mode shape curvature. For each layer of the iteratively corrected structure, the natural frequencies and mode shapes of the overall structure are calculated based on the derived relationship between the Fourier spectrum of the local structural relative displacements and the overall structural characteristic quantities. The location of the damage is then determined based on the change in mode shape curvature, with the curvature calculated using the central difference method. This invention provides a convenient and economical method for identifying damage in engineering structures.

The invention discloses a multi-frequency seismic response analysis method based on a Fourier enhanced physical information KAN network, and relates to the technical field of seismic response analysis, and the method comprises the steps: carrying out the Fourier expansion of each feature of an input spatial domain parameter at an input layer, and obtaining a Fourier basis combination; a Fourier spectrum is extracted from a summation result of all Fourier basis combinations in a hidden layer, and the Fourier spectrum is dynamically activated based on a linear combination activation function of Fourier basis; by introducing a KAN structure and decomposing high-dimensional mapping into a plurality of visual single-variable functions, the interpretability of the model is improved, so that the influence of each input variable on output has clear physical significance; for the problems of low parameter efficiency and insufficient high-frequency characterization capability of KAN, Fourier basis function explicit coding input signal frequency domain features are introduced, and expression of neural tangent kernel multi-spectrum is realized.

The invention discloses a weak signal detection method based on z-domain autoregression and analytic continuation, and belongs to the technical field of signal processing and geophysical exploration, and the method comprises the steps: carrying out the multi-group random noise addition of an observation time sequence comprising a target weak signal, and obtaining a multi-group noise addition sequence; calculating an AR-z spectrum for each group of noise adding sequences; performing complex frequency estimation on the noise adding sequences by adopting frequency domain autoregression estimation to obtain complex frequency information of Fourier spectrums corresponding to the noise adding sequences; in combination with prior complex frequency information of a target weak signal, an AR-z spectrum of a noise adding sequence is constructed through analysis and continuation; and geometrically averaging amplitudes of AR-z spectrums obtained by calculation of the multiple groups of noise adding sequences to obtain a robust AR-z spectrum for weak signal detection. Through z-domain autoregression and random noise addition, the frequency resolution limit of traditional Fourier transform is broken through, the signal-to-noise ratio threshold value of the AR-z spectrum is reduced to be smaller than 1, and a universal solution is provided for signal detection in an extreme noise environment.

The present invention relates to the field of fault warning technology, and more specifically, the present invention relates to a vibration screen fault warning method and system, comprising: obtaining the operating parameters of the vibration screen when screening ore, wherein the operating parameters include sound characteristics and vibration characteristics, and determining the current operating condition of the ore; according to the difference between the real-time operating parameters of the current ore at the current moment under the corresponding operating condition and the standard operating parameters under the same conditions. The present invention clusters the historical screening data of different ores to identify the operating conditions, and combines the Fourier spectrum analysis of the sound and vibration signals to calculate its sensitivity to the fault characteristics and assign differentiated weights. During real-time operation, a weighted comparison is performed between the collected signal and the standard parameters based on the current ore type and operating condition, and the degree of fault is calculated. If the set threshold is exceeded, an audible and visual alarm is triggered. This method can improve the accuracy of the warning and reduce the risk of sudden failures.

The present invention belongs to the technical field of micro-nano structure processing, and relates to a polarization interference lithography method and system for three-dimensional micro-nano structures. An amplitude spatial light modulation element is used to modulate the amplitude of the incident light field to obtain an amplitude-modulated light field; a crystal holographic optical element is used to modulate the amplitude-modulated light field to obtain a first transmitted light field; a holographic optical element is used to modulate the first transmitted light field to obtain a second transmitted light field on the Fourier spectrum plane; a first Fourier transform lens is used to perform Fourier transforms on the amplitude-modulated light field, the first transmitted light field, and the second transmitted light field to obtain a Fourier-transformed light field; a second Fourier transform lens is used to perform a Fourier transform on the Fourier-transformed light field to obtain an interference light field with adjustable light intensity and phase on the imaging plane; the light intensity and phase of the interference light field are adjusted so that the distribution of the interference light field matches a preset three-dimensional micro-nano structure, thereby preparing the preset three-dimensional micro-nano structure. This solution can prepare three-dimensional micro-nano structures with high precision and small line widths.

The invention relates to the field of computational imaging, and particularly discloses a Fourier single-pixel imaging method based on space-time three-dimensional joint priori, which comprises the following steps: S1, Fourier spectrum measurement: projecting a sine structured light pattern to a dynamic target scene through a spatial light modulator, collecting a corresponding light intensity signal by using a single-pixel detector, and measuring the Fourier spectrum; fourier spectrum sampling data of the scene is obtained; wherein the sine structured light pattern generates a measurement value of a complex Fourier coefficient through phase shift technology combination; s2, constructing a reconstruction objective function: taking Fourier spectrum consistency of measurement data and a reconstruction result as a data fidelity constraint, and introducing space-time three-dimensional joint priori as a regularization item; the space-time three-dimensional joint priori comprises a 3D Hessian regularization item and a local low-rank time priori item; and S3, an optimization solution step. By adopting the technical scheme of the invention, high-fidelity reconstruction of a dynamic scene can be realized by simultaneously utilizing spatial structure priori and time redundancy information at an extremely low sampling rate.

The application provides a gravity acceleration measurement method and system based on YOLOv8, which comprises the following steps: labeling the pendulum position data in each picture in the pendulum picture data set to construct a corresponding pendulum labeling data set; taking the pendulum picture data set as the model input and synchronously taking the pendulum labeling data set as the label to train and fine-tune a preset YOLOv8 model to generate a corresponding pendulum detection model; based on the video frame rate and the video length, detecting the pendulum in the pendulum swing video through the pendulum detection model to output the center coordinates of the pendulum; normalizing the center coordinates of the pendulum and synchronously combining Fourier spectrum analysis to obtain the corresponding swing period, and calculating the corresponding gravity acceleration according to a preset formula and the swing period. The application can effectively improve the measurement accuracy.

The application provides a video micro-vibration signal correction method and relates to the technical field of data processing.The method comprises the following steps: arranging a vibration sensor on a structure to be measured, capturing a video of the structure to be measured; based on the video capture result, extracting time domain features and frequency domain features of a vibration signal; using the time domain features and the frequency domain features to train and construct a multi-modal fusion calibration model, wherein the multi-modal fusion calibration model is a hybrid CNN-LSTM network architecture; introducing a spectral loss term based on a Fourier spectrum logarithmic amplitude difference, and using the constructed multi-modal fusion calibration model to obtain a dual correction target result of time domain waveform reconstruction and frequency domain key modal alignment, thereby completing the correction of the video micro-vibration signal.The application solves the problem of low precision caused by environmental factors, noise, on-site operation and the like when micro-vibration information is directly extracted from a monitoring video.

The present invention belongs to the field of image processing technology and proposes a method for image edge detection using adjustable directional Fourier single-pixel imaging, comprising: S1, constructing an adjustable directional filter; S2, using the adjustable directional filter as an adjustable directional operator to modulate Fourier speckle to obtain edge modulated speckle; S3, selecting a low-frequency sampling region radius and filtering the image spatial domain using the edge modulated speckle corresponding to the low-frequency region to obtain a corresponding fan-shaped Fourier spectrum; S4, performing adaptive sampling; S5, using the adaptive sampling method in step S4, obtaining higher-frequency Fourier spectrum coefficients based on the Fourier spectrum coefficients of the target low-frequency semicircular region; and S6, obtaining image edge information through inverse Fourier transform. The present invention can achieve fast and high-quality edge extraction of target edges in different directions at low sampling rates, thereby improving the quality of edge detection.

An optical measurement system according to the present invention includes: a light source that generates coherent light; a beam splitter that splits the coherent light into illumination light and reference light; a synthesis unit that optically superimposes the reference light and object light generated by illuminating a sample with the illumination light; an objective lens of an infinity-corrected optical system, the objective lens being disposed between the sample and the synthesis unit; an image sensor that records a hologram generated in the synthesis unit; and a processing device that processes the hologram recorded by the image sensor. The processing device calculates a first complex amplitude distribution on the recording surface of the image sensor on the basis of the hologram, calculates a second complex amplitude distribution on the Fourier surface of the objective lens by calculating diffraction of the first complex amplitude distribution for a first distance between the recording surface of the image sensor and the Fourier surface of the objective lens, and performs coordinate transformation to transform the second complex amplitude distribution into a Fourier spectrum on spatial frequency spectrum coordinates.