61results about How to "Exact decomposition" patented technology

A proximity detection system for determining when a single channel receiver becomes proximate to any point on a wireless closed boundary continuously generated by a transmitter. The transmitter includes a magnetic field generator broadcasting a composite, modulated, time-varying magnetic field signal of a particular carrier frequency. The receiver forms a measure of the broadcast magnetic field intensity incident to the location of the receiver or a measure of the power or energy of incident field.

The invention provides a regional division method for satellite observation. The method comprises the follow steps of: step 1, determining the effective region of a target region; step 2, determining a maximum observation lateral swinging angle and a minimum observation lateral swinging angle corresponding to the effective region; step 3, making a current lateral swinging angle equal to the minimum observation lateral swinging angle, and emptying a result sub region set; step 4, adding the current lateral swinging angle into the result sub region set; step 5, adding a set granularity angle to the current lateral swinging angle; step 6, turning to step 4 if the sum of the current lateral swinging angle and the granularity angle is smaller than the maximum observation lateral swinging angle, and otherwise, turning to step 7; and step 7, adding the current lateral swinging angle into the result sub region set so that the effective region is divided into observation strip belts corresponding to each lateral swinging angle in the result sub region set. The regional division method overcomes the limitation of the Gaussian projection on task longitude differences, the accuracy is higher, and in addition, the regional division method is suitable for the target region observation by a plurality of satellites.

The invention discloses a model training method. The method comprises the steps that training batch data is acquired, the training batch data comprises a plurality of training samples, for each training sample, identity features and age features corresponding to sample images are acquired through a feature extraction module, and typical correlation mapping variables corresponding to the features are acquired through a typical correlation mapping module; based on the typical correlation mapping variables corresponding to the training samples, counting correlation coefficients corresponding to the training batch data through a batch correlation coefficient measurement module; performing joint supervised learning through an identity discriminator and an age discriminator, performing decorrelation training of identity features and age features through correlation coefficients, and performing training to obtain a feature extraction module meeting a training target. And based on the decorrelation-based adversarial learning, the correlation between the identity features identified by the feature extraction module and the age features is extremely low, so that the identification precisionis improved. The invention further discloses a cross-age face recognition method, a corresponding device, equipment and a medium.

The invention discloses a multilingual word segmentation method based on dictionaries and grammar analysis. Efficient and accurate word segmentation of mixed texts of Chinese, Japanese, Korean, Cantonese and the like can be realized, flexible lexicon expansion of words for different time periods and different professionals can be realized, lexicon information is updated effectively, and efficient and accurate multilingual language text word segmentation is realized; a word segmentation sub-device of Chinese, Japanese, Korean, Cantonese and other language families, a Chinese quantum word segmentation device and a western language word segmentation device are embedded to realize the accurate word segmentation of each language text; a text segment to be performed with word segmentation is segmented by a built-in language segment coded identification mechanism, each segmented text segment corresponds to a language family, and the word segmentation is carried out by using a corresponding word segmentation sub-device; the word segmentation of western inflectional languages and the smart mode word segmentation of the Chinese, Japanese, Korean, Cantonese can be realized by grammar analysis, and texts containing Arabic numeral information can be processed; and meanwhile, the word segmentation of texts with a plurality of mixed languages can also be realized by the multilingual word segmentation method provided by the invention, thereby getting rid of the limitation that a word segmentation tool can only realize the word segmentation of single language and some individual languages and ensuring the security, accuracy, efficiency, flexibility and universality of word segmentation of texts. The multilingual word segmentation method provided by the invention has a wide application prospect in the text word segmentation fields such as enhancement of mass data text classification, text information extraction, autoabstract, etc.

The invention belongs to the field of seismic signal processing and is a seismic signal noise suppression method based on CEEMDAN and Savitzky-Golay filtering. The method comprises following steps: providing a noisy seismic signal to be processed as an original noisy seismic signal s(t) (t=1,...,L) input to the entire system, wherein L is the sampling point of the seismic signal; performing CEEMDAN algorithmic decomposition on the input noisy seismic signal s(t) to obtain multiple IMF components as shown in the description and residual components rN(t); arranging the IMF components as shown inthe description from high frequency to low frequency sequentially, and determining a boundary M between the noise and the signal by calculating the correlation coefficient between the IMF component and the original noisy seismic signal s(t); finding the IMF component as shown in the description which mainly contains noise signal in the IMF component, the IMF component which contains noise signalas shown in the description serving as an input signal and sending to the SG filter for further smoothing to obtain the filtered signal as shown in the description; and reconstructing the data to obtain a denoised seismic signal as shown in the description. According to the seismic signal noise suppression method based on CEEMDAN and Savitzky-Golay filtering, the mode mixing problem in the mode decomposition method based on traditional experience can be solved; the quality of the seismic signal is improved; and seismic data with a higher signal-to-noise ratio is obtained.

A method of measuring the state of flow above an airfoil using an estimation of the velocity field based on a combination of Particle Image Velocimetry PIV and multiple surface pressure measurements processed through a POD / mLSE algorithm. Integral to the POD / mLSE algorithm is the estimation of the global POD coefficients. The utility of these time dependent coefficients, which are estimated from surface pressure only, are demonstrated in a simple proportional feedback loop (as the time series to drive the actuators) to keep the flow attached. This method requires realistic feedback flow control since surface measurements and not inflow measurements are required for practical applications. The estimation method works well with dynamic strain on flexible bodies and is not limited to estimation from pressure only.

The invention discloses a tensor decomposition model-based interest point type prediction method. The method includes the following steps that: a) the access point coordinate information and stay timeof users in T time periods in each day are obtained from new energy vehicle trajectory data; b) the ranges of the stay areas of the users are defined as uncertain areas according to time periods whenthe users are located at access points, the durations of the stay of the users, the IDs of the types of interest points in the uncertain areas are obtained, and tensors are constructed; c) clusteringprocessing is performed on trajectory data corresponding to all the users; d) the types of the interest points of the users are predicted by using a tensor decomposition model; e) after the information of work, places of residence and places of departure is combined, the similarity of the types of interest points between the places of departure and destinations are adopted as the regularization item constraint of tensor decomposition, and the low-rank approximate solutions of the tensors are calculated; and f) an objective function is iteratively optimized by using an alternating least squares method, and the probabilities of the access of different kinds of interest points by the users are obtained and are adopted as a prediction results. With the method of the invention adopted, the types of the interest points accessed by the users can be accurately predicted.

The embodiment of the invention discloses a method for predicting output power of a photovoltaic power generation system, variation mode decomposition is performed on historical output power data of the photovoltaic power generation system in a preset time period, an extreme learning machine prediction model is built according to a plurality of decomposition components obtained by decomposition and corresponding meteorological data, a prediction result of each decomposition component is calculated according to the extreme learning machine prediction model, and a sum of the prediction results is used as a prediction result of the output power of the photovoltaic power generation system. A variation mode decomposition algorithm has good noise robustness and non-recursiveness, and selection of reasonable parameters can effectively avoid a mode aliasing phenomenon, thereby obtaining a high-accuracy decomposition signal, and facilitating improvement of prediction accuracy; and the characteristics of good generalization performance and fast learning speed of an extreme learning machine can further improve prediction precision and prediction efficiency. In addition, the embodiment of the invention also provides a corresponding realization device, which further enables the method to have practicability, and the device has corresponding advantages.

The invention relates to a nondestructive testing method for debonding defects of a solid rocket engine interface, and the method combines an electromechanical impedance method and a frequency response function method and is used for testing a solid rocket engine combustion chamber interface. The method comprises the following steps of forming a detection system by using a universal equipment NI multi-channel high-speed data acquisition device, a PCB force hammer, a sensor and the like; knocking the surface of a structure to be detected by using the force hammer to provide a pulse-shaped force; and measuring a time domain waveform of a response signal through a signal acquisition program to obtain a frequency response function curve of an engine heat insulation layer / lining layer / propellant interface so as to judge whether the interface structure is intact. The method is used for the flexible and rapid nondestructive detection of the debonding defects of the interface of the solid rocket engine propellant / lining layer / heat insulation layer of the composite fiber material with the diameter of about 2000 mm, particularly for an integrally assembled product, the detection can be realized on a storage site without decomposition, the method has the characteristics of convenience and high efficiency, and the nondestructive testing method is suitable for the rapid inspection of the debonding defects of the interface of the integrally assembled product.

The invention discloses a decomposition method of image texture and structure based on a total variational model of a guide map, belonging to the technical field of image processing. According to theinvention, firstly, the local main structure of the image is reconstructed based on the guide image filter, then the texture descriptor is calculated according to the reconstructed local structure image, finally, the accuracy and computational efficiency of texture and structure decomposition are improved by combining the multi-scale total variational model and block translation method. The technical scheme of the invention can obtain better texture / structure decomposition effect for noisy images, the textures and structures of different scales can also be decomposed accurately, and the decomposed structure layers can keep the original shading of the image, avoiding the blurring of the structure caused by local smoothing or the color block effect caused by the global optimization method. In addition, the technical proposal of the invention needs to extract simple features, and does not depend on learning a large number of image samples.

The invention discloses a method for decomposing and reconstructing a current instantaneous sequence component of a triangular connection current transformer. The method comprises the steps that S1, an instantaneous sequence component decomposition algorithm model is built, and the positive sequence component and the negative sequence component of a fundamental wave and the positive sequence components and the negative sequence components of all subharmonic waves are made to be completely decoupled and decomposed under a synchronous coordinate system; S2, the component correlation coefficients am<(1)> and am<(2)> of the m subharmonic waves is obtained; S3, the compensating current reference values of all triangular link current controllers are reconstructed to obtain the reference currents of all triangular links, and the reference currents are used for current tracking and control of the triangular current transformer. The method has the advantages that the positive sequence component and the negative sequence component of the current of the fundamental wave and the positive sequence components and the negative sequence components of the currents of all the subharmonic waves can be effectively extracted, and load currents and all triangular link currents can be decomposed and reconstructed accurately and fast. The accurate compensating current reference values are provided for the triangular connection current transformer, the number of times of harmonic compensation can be selected flexibly, and the defects that a traditional sequence decomposition algorithm is large in calculated amount and only fundamental components are considered are overcome.

In a sound field control apparatus including multiple speakers, multiple microphones gathering sound radiated from the multiple speakers, a mode decomposition filter that performs mode decomposition on a sound pressure distribution, and a control filter that controls the input signals to be input to the multiple speakers such that the mode amplitudes of the modes decomposed by the mode decomposition filter can have a predetermined value, a sound pressure distribution in the acoustic space is measured, and the sound pressure distribution in the acoustic space is expressed by using a sinusoidal function and cosine function of a space frequency of the mode to be controlled in amplitude. The mode space frequency is corrected such that the expressed sound pressure distribution can be equal to the measured sound pressure distribution, and the filter coefficient for the mode decomposition filter is determined based on the mode space frequency obtained by the correction (corrected mode space frequency).

The invention discloses a multi-material decomposition method for a single-energy-spectrum CT image. The multi-material decomposition method comprises the following steps: (1) acquiring the single-energy-spectrum CT image; (2) aiming at a CT image, constructing a decomposition target function comprising a data fidelity term and three penalty terms according to a CT image domain multi-material decomposition theory, wherein the data fidelity term ensures that the error between the measured value and the true value is as small as possible, in the three penalty terms, the first term uses a total variation term to ensure the piecewise constant characteristic of a CT material image, the second term uses a 0 norm term to ensure the sparsity of materials in the CT image, and the third term uses afeature function term to ensure that a multi-material decomposition result meets the constraint that the volume fraction is 0-1 and the sum of the volume fractions of all the materials is 1; and (3) solving an initial value of the target function by adopting a matrix inversion method based on dual-material hypothesis, and solving the target function by adopting an alternating direction multipliermethod, so that accurate decomposition of various materials under single-energy common CT is realized, and the decomposition precision is equivalent to that of dual-energy CT.

The invention discloses a mechanical fault diagnosis method based on multi-sensor multivariate data fusion. The method comprises the following steps: acquiring multivariate vibration signals x1(t), x2(t),..., xn(t) of a bearing acquired by multiple sensors; preprocessing the multivariate signal by using a non-local mean denoising method; adopting noise-assisted adaptive projection eigentransformation multivariate empirical mode decomposition (NAAPITMEMD) method to obtain a plurality of IMF (intrinsic mode function) groups which are the same in length and have the same appearance sequence of same-frequency components of the signals in each group; calculating a fault correlation factor (FCF) by using correlation analysis, and selecting effective IMFs; and finally, extracting a fault characteristic frequency through spectrum analysis to realize mechanical fault diagnosis. The common problems of power imbalance and mode aliasing in multi-element signal processing can be effectively relieved, and the accurate mechanical fault characteristic frequency can be obtained.

The invention discloses a distortion signal electric quantity metering method based on empirical wavelets. Firstly, original voltage signals and current signals in a power grid are collected as to-be-observed signals; empirical wavelet transform is carried out on the to-be-observed signals, specifically, spectrum division is carried out adaptively on the to-be-observed signals to divide the signals into different modes, and then, empirical wavelet transform is carried out on each mode to decompose the to-be-observed signals; and fundamental wave signals and distortion signals of the to-be-observed signals are extracted from an empirical wavelet transform decomposition result to obtain active power P. The problems that the wavelet basis and decomposition layer number need to be selected in wavelet transformation of an existing electric quantity metering technology, self-adaptive capacity is poor, self-adaptive detection cannot be achieved, wavelet decomposition has the frequency aliasing phenomenon, and the existing technology is limited by the Heisenberg uncertainty principle are solved.

The invention discloses an image fusion method based on cosine modulation filter group. The original inputted image is decomposed into sub-band signals with cosine modulation analyzing filter group, then calculates the fusion measuring index of each sub-band image with measuring index based on local contrast, carries on fusion operation to these sub-band signals according to fusion rule that selecting the maximal fusion measuring index and acquires a group of new fused sub-band signals, and carries on consistency check to the sub-band image being fused, finally, acquires the fused image with synthesized filter group. The invention upgrades the quality of the fused image greatly.

The invention relates to a simulation method for variation modal decomposition parameter optimization. The method comprises the following steps: constructing an analog signal; using a genetic algorithm to improve a particle swarm algorithm to optimize variational mode decomposition, and calling as PSOGA-VMD; using wavelet transform analog signal components to perform time-frequency analysis, and comparing the time-frequency analysis with analog signal set frequency characteristics to detect decomposition precision.

The invention discloses a step-by-step decomposition method for a shallow sea complex terrain. On a shallow sea land frame, various scales of seabed terrains are superposed to form a complex geomorphic system, and consequently, it is difficult to carry out quantitative studies. According to the method, a Fourier analysis method is combined with a Butterworth filter to realize conversion of terrainsignals from a spatial domain to a frequency domain, and superimposed terrains are decomposed into a plurality of groups of single-type terrains with different frequencies through frequency division.An example test result shows that the method can quickly and accurately decompose the complex seabed superimposed terrain into terrains with different spatial scales. The method has important practical value in the aspects of marine surveying and mapping, marine engineering safety evaluation, marine science research and the like.

The invention, which belongs to the technical field of spectrum detection, particularly relates to a plasma emission spectrum interference correction method. Calculation is performed based on other characteristic spectrum line information of a corresponding element of an interference line B in a corrected spectral segment 20 to obtain a spectrum line intensity of the interference line B; and thusa to-be-analyzed spectrum peak 11 is decomposed conveniently to obtain a spectrum line intensity of an analysis line A, thereby completing e correction of the spectral interference. The method has thefollowing technical effects: the spectrum line can be decomposed accurately based on the intensity relationship of the same type of emission spectrum lines in a same ion state in the plasma emissionspectrum theory and the detection and calculation of the emission intensity of the spectrum line being easy to affect are realized, so that the accuracy of the component detection result is improved and the detection element range is extended. Therefore, the application range of LIBS technology is further expanded.

The invention relates to the field of machining, in particular to a trial cutting device for the analysis of workpiece surface quality problems and an analysis method. The trial cutting device comprises a base, wherein the base is rectangular, and an X-axis test component, a Y-axis test component, a Z-axis test component, an A-axis test component, a B-axis-C-axis test component and an origin hole are arranged on the base; the X-axis test component, the Y-axis test component and the A-axis test component are all parallel with the edges of the base; the X-axis test component is arranged on the edge of the base, and the Y-axis test component and the A-axis test component are separately arranged on the edges, adjacent to the X-axis test component, of the base; the Z-axis test component is a bulge with a trapezoidal longitudinal section and is arranged in the middle of the base; the origin hole is formed in the middle of the top plane of the Z-axis test component; and the B-axis-C-axis test component is a ring and is arranged between the Z-axis test component and the X-axis test component and between the Y-axis test component and the A-axis test component. Through the independent processing of coordinates, workpiece surface quality problems caused by complicated multi-axis simultaneous processing are solved, and the accurate and rapid positioning of failure reasons is realized.

The invention discloses an image fusion method based on cosine modulation filter group. The original inputted image is decomposed into sub-band signals with cosine modulation analyzing filter group, then calculates the fusion measuring index of each sub-band image with measuring index based on local contrast, carries on fusion operation to these sub-band signals according to fusion rule that selecting the maximal fusion measuring index and acquires a group of new fused sub-band signals, and carries on consistency check to the sub-band image being fused, finally, acquires the fused image with synthesized filter group. The invention upgrades the quality of the fused image greatly.

In a method of filtering an image with bar-shaped structures by means of Gabor filters, which are formed in the spatial domain by a two-dimensional Gaussian bell-shaped curve on which a cosine function is superimposed in a main direction, the image is divided into tiles, a predominant direction of the bar-shaped structures is determined for each tile and the filtration is undertaken in such a way that one tile at a time is rotated until the predominant direction lies at right angles to the main direction of the Gabor filter, one filtration takes place in the main direction and another filtration takes place at right angles to this, and the filtered tile is rotated back again.

The arc-shaped pipeline construction method comprises the following steps that pipe sections with different arc lengths are subjected to lofting and splitting in a BIM drawing with tee joints, reducing joints and connecting points with different radians as breakpoints; after pipeline lofting is completed, the arc length and the radius of each split arc-shaped pipeline are counted and then delivered to a processing factory to be processed; after arrangement of all the arc-shaped pipelines in the BIM drawing is completed, in combination with the interval requirement of the anti-seismic supports and hangers, point distribution is deeply conducted on the shared supports, and the supports on the arc-shaped pipelines are required to be perpendicular to the arc-shaped pipelines; after support point distribution is completed, the positions of the supports are accurately positioned through accurate paying-off of the BIM paying-off robot; and the arc-shaped pipeline subjected to cold bending machining in the factory is conveyed to the site and conveyed to the corresponding position to be assembled according to the marked number, and the combined pipe sections are hoisted. The construction period of the arc-shaped pipeline can be greatly shortened, and the arc-shaped pipeline is good in forming, neat and attractive.

The invention discloses a pontoon bridge lower transverse connecting device. The pontoon bridge lower transverse connecting device comprises a lower transverse plate pin and a lower transverse plate pin seat. The lower transverse plate pin and the lower transverse plate pin seat are arranged on two adjacent pontoon bridge units; and the lower transverse plate pin can rotate in a horizontal plane,is inserted into the lower transverse plate pin seat after rotating by a set angle to achieve transverse connection between lower portions of two pontoon bridge units, and is screwed out of the lowertransverse plate pin seat after rotating by a set angle reversely to achieve unlocking of the two pontoon bridge units.

The invention discloses a knowledge question and answer intelligent search method based on block decomposition, and the method comprises the following steps: preprocessing knowledge data imported into a database, and converting the knowledge data into a triple data set of structured data; coding the triple data set of the converted structured data by adopting One-hot coding; segmenting the data set into a training set and a test set, and constructing a knowledge question and answer intelligent search model by adopting a BTD tensor decomposition algorithm. According to the method, the three-order binary tensor problem is converted into a three-order binary tensor problem in mathematics, the missing subject or object entity can be predicted, all real triple subsets are used for learning embedding vectors of entities and relations, the method is more generalized decomposition based on mathematical tensors, and formed factors can be decomposed more accurately. According to the method, under the characteristics of low search accuracy and poor user search experience, search complementation can be performed in a more accurate manner after a large amount of data training, so that an efficient and accurate search system is realized.

The present invention provides a pulse noise filtering method for load switch event detection which is used to solve the problem that a common noise elimination device cannot effectively filter noisewith non-stationary and non-Gaussian features in the prior art. The power signal noise filtering method comprises the steps of: inputting actually measured power signals; performing segmentation on the actually measured power signals and rearranging data to a matrix; transforming the power signal matrix; performing matrix initialization prior to iteration; and performing iteration calculation on the matrix after the initialization is completed to obtain a matrix after pulse noise is filtered, and further converting the matrix to the power signals. Through transformation for the power signal matrix, the sparsity of the power signals is processed to effectively filter the pulse noise in the power signals and achieve high fidelity for the useful power signals, the phase step features of the power signals cannot be damaged, the method is simple and practicable, the computing speed is fast, and therefore, the accurate switch event detection can be performed to perform accurate decompositionfor domestic electricity load.

The invention discloses a material decomposition and identification method based on phase CT, and the method comprises the steps: 1, enabling m to be equal to 0, and initializing a first substrate image f (m) and a second substrate image g (m); 2, according to a formula (6), calculating an estimated value and a corresponding residual error of the discretized absorption projection data in the mth iteration and an estimated value and a corresponding residual error of the discretized differential phase projection data in the mth iteration; 3, according to a formula (5), iteratively solving a value f (m + 1) of a first base material image f (m) in the (m + 1) th iteration and a value g (m + 1) of a second base material image g (m) in the (m + 1) th iteration; 4, performing regularization operation on f (m + 1) and g (m + 1); 5, if the stop condition is not met, enabling m to be equal to m + 1, and executing the step 2 again. The absorption projection information and the differential phase projection information are combined for base material decomposition, the noise of the obtained base material image is smaller, and the noise of the equivalent atomic number image calculated subsequently is smaller than that of a conventional method.