216 results about "Separation result" patented technology

The application discloses a role separation method based on voice, comprising the following steps: extracting feature vectors frame by frame from a voice signal to get a feature vector sequence; assigning role tags to the feature vectors; using the feature vectors with role tags to train a deep neural network DNN model; and judging a role sequence corresponding to the feature vector sequence according to the DNN model and a hidden Markov model HMM trained using the feature vectors, and outputting a role separation result, wherein the DNN model is used to output the probability of each role according to an input feature vector, and HMM is used to describe the jump relationship between roles. The invention further provides a role separation device based on voice. According to the method provided by the invention, as the DNN model with strong ability of feature extraction is used to model speaker roles, the method has stronger characterization ability than a traditional GMM (Gaussian Mixture Model), role characterization is more detailed and accurate, and therefore, a more accurate role separation result can be obtained.

The invention discloses a satellite attitude control system failure diagnosis device based on a state observer and an equivalent space and a satellite attitude control system failure diagnosis method based on the state observer and the equivalent space, which belong to the field of aerospace and aim to solve the problems of high hardware complexity, low control accuracy and low failure diagnosis algorithm effectiveness of the conventional failure diagnosis method. The method provided by the invention comprises the following steps that: 1, a failure diagnosis observer outputs a satellite triaxial angular rate residual according to output signals of an actuator and a gyro sensor; 2, an equivalent vector space description module constructs equivalent space descriptions of the gyro sensor according to the output signal of the gyro sensor, wherein an output equivalent vector p is used for judging whether the gyro sensor fails or not; and 3, a failure diagnosis and isolation module outputs a failure separation result indicating that the actuator or the gyro sensor fails according to the satellite triaxial angular rate residual obtained by the step 1 and the equivalent vector p obtained by the step 2, and further judges which axis of the failing part fails.

It relates to the after potential exploration of an oil filed, with the ability of analyzing the oil distribution, left resource quantity, and so on. It comprises the data inspection device, data preprocessing device, KH value separation device, oil well water absorption preprocessing device, water absorption section separation device, dynamic personnel analysis separation device, small layer weighting separation device, single sand separation result accumulation device, data base, water filling effect, strata HK, and weighting of the oil reserve weighting by the dynamic personnel, determining the strata output through the water absorption section, water content of the strata through well fluid section, with the whole process being computerized and accurate.

A multicell cyclone with a plurality of cyclone cells and a method for the production thereof, in which the cyclone cells of the multicell cyclone are combined into groups from which deposited particles are cleaned out by vacuum suction through channel structures. This permits better control of the discharge of separated particles. Vacuum suction of the separated particles can be effected, for instance, via openings (19) which are jointly provided for the group (21) of cyclone cells. The group of cyclone cells can furthermore be produced as a module, which can, for instance, be assembled from two parts (12, 13) in which like structural elements of a group of cyclones are combined into one respective part. This makes it possible to optimize the separation result with little structural complexity and at low production cost.

A voice recognition system (10) for improving the toughness of voice recognition for a voice input for which a deteriorated feature amount cannot be completely identified. The system comprises at least two sound detecting means (16a, 16b) for detecting a sound signal, a sound source localizing unit (21) for determining the direction of a sound source based on the sound signal, a sound source separating unit (23) for separating a sound by the sound source from the sound signal based on the sound source direction, a mask producing unit (25) for producing a mask value according to the reliability of the separation results, a feature extracting unit (27) for extracting the feature amount of the sound signal, and a voice recognizing unit (29) for applying the mask to the feature amount to recognize a voice from the sound signal.

The invention discloses a matrix-transformation-based method for underdetermined blind source separation. The method comprises the following steps of: gradually transforming a mixed matrix to obtain a transformed matrix, then acting the transformed matrix on an observation signal to gradually remove each source signal, and constructing multi-stage binary shielding templates by zero points generated in each observation signal to gradually separate the source signal. By the method, the requirement on sparsity of the source signal is reduced; aliasing of at most M-1 paths of source signals is prevented (wherein M is the number of sensors); source signals for forming each time frequency point can be clearly known; the underdetermined separation problem of music signals and noise signals is solved; the statistical property requirement on the source signal is low; the underdetermined separation problem of Gaussian signals and related signals is solved; processing stages are adjusted according to the requirement of separation precision; when the processing level is more, the separation result is better; the separation process is realized by aid of the matrix transformation; and the complexity is relatively low.

The present application relates to separation processes and systems and more specifically to hybrid carbon dioxide separation processes. In one embodiment, the system for the separation or removal of carbon dioxide comprises an apparatus for a selective separation of carbon dioxide (CO₂) from flue gas—typically exhaust gases, syngas or natural gas streams—using one or more so-called CO₂ reverse selective membrane(s) in the first separation unit to enrich a feed gas stream which contains carbon dioxide with CO₂ and by separating other constituents of the gas stream. Thus, the feed gas stream is separated in the first separation unit by CO₂-reverse-selective separation into a CO₂-lean gas stream and a CO₂-enriched gas stream. The CO₂-enriched gas stream is fed to a second separation unit which is a CO₂-selective separation unit. The second separation results in a purified CO₂-rich gas stream and a remaining CO₂-lean gas stream.

The invention relates to a mechanical vibration fault characteristic time domain blind extraction method, and belongs to the technical field of mechanical equipment status monitor and fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method includes: firstly, expanding a vibration observation signal into a high dimension signal subspace; then, obtaining a low dimension signal; afterwards, performing FastICA independent component analysis, calculating normalization kurtosis of all independent components, figuring out a component signal corresponding to the minimum normalization kurtosis, and using an orthogonal matching pursuit algorithm to reconstitute periodic signals; subsequently, removing the reconstituted periodic signal from each independent component, and then using an improved KL distance algorithm to calculate a distance matrix among the independent components after the periodic signals are removed from the independent components, and performing dynamic particle swarm clustering so as to obtain an estimation signal; finally, analyzing an envelope demodulation spectrum of the estimation signal, and performing fault diagnosis. The mechanical vibration fault characteristic time domain blind extraction method is suitable for processing a long convolution data problem, can effectively reduce influences from periodic ingredients on a blind separation result, and simultaneously can solve blind separation result order uncertainty problems, and finally achieves bearing fault characteristic extraction.

A signal processing device includes a signal transform unit which generates observation signals in the time frequency domain, and an audio source separation unit which generates an audio source separation result, and the audio source separation unit includes a first-stage separation section which calculates separation matrices for separating mixtures included in the first frequency bin data set by a learning process in which Independent Component Analysis is applied to the first frequency bin data set, and acquires a first separation result for the first frequency bin data set, a second-stage separation section which acquires a second separation result for a second frequency bin data set by using a score function in which an envelope is used as a fixed one, and executing a learning process for calculating separation matrices for separating mixtures, and a synthesis section which generates the final separation results by integrating the first and the second separation results.

The invention discloses an algorithm for separating vibration signal blind sources under a strong noise environment. The method comprises the following steps: 1, de-noising a group of given mixed signals containing noise through a time delay autocorrelation method to acquire de-noised mixed signals; 2, performing mean value removal and steady whitening pretreatment on the mixed signals acquired in the first step to further reduce the influence of the noise signals on the separation result; and 3, calculating second-order and fourth-order cumulated amount of the initial separate signals, using the sum of diagonal elements of second-order and fourth-order cumulated amount matrixes as a cost function, and maximizing the cost function to similarly diagonalize the joints of the cumulated amount matrixes so as to realize the separation of the signals of the independent sources and acquire orthogonal separation matrixes. The method combines the conventional de-noising method and the blind separation algorithm to realize the separation of the mixed signals under the strong noise environment, and has the advantages of good separation effect, high convergence rate and de-noising effect free from the limitation of the set threshold value compared with the conventional algorithm.

An image processing apparatus comprises: a conversion unit that converts an original into image data by scanning the original; a detection unit that detects a blank page in the image data converted by the conversion unit; a separation unit that deletes image data of the blank page detected by the separation unit from the converted image data; and a transmission unit that transmits to an external apparatus (a) the converted image data from which the image data of the blank page is separated by the separation unit, (b) the separated image data of the blank page, and (c) notification information formed based on a separation result by the separation unit.

The invention relates to an unmanned aerial vehicle longitudinal flight control system fault detection and separation method based on nonlinear adaptive observers. According to the invention, based onan unmanned aerial vehicle longitudinal nonlinear model, a set of nonlinear adaptive observers are designed for actuator and sensor faults of a longitudinal flight control system; a fault separationproblem is transformed into a model matching problem; by referring to the idea of a contribution analysis method, a fault estimation value and the fault direction, which are acquired by each adaptiveobserver, are used to contribute a contribution function, and standardization is carried out; the optimal matching model is determined by selecting the maximum standardized contribution function; andfinally, the maximum standardized contribution function is compared with a threshold value, and fault detection and separation results are acquired. According to the invention, a new reference index is provided for model matching by referring to the contribution analysis method; an analytical model is used to design the nonlinear adaptive observers to acquire a reliable fault estimation value andfault direction; and the problem that the fault direction is difficultly determined in a data driving method is overcome.

The invention provides a wave field separating method and device. The method comprises the steps that variation function values of anisotropy parameter distribution of initial models are calculated and fitted to obtain variation functions of the initial models; the variation functions and a reference point search strategy are utilized for selecting N reference models from the initial models; a vector elastic wave field of a spatial domain is converted into a wave number domain; following operation is carried out on each reference model selected from the N reference models, pseudo differential operators are truncated, the wave field separation is carried out on the wave number domain, the vertical and horizontal wave field is reversely converted into the spatial domain, and the wave field separation results of the reference models are obtained; the variation functions are utilized for calculating the weight coefficients of all the reference models relative to the initial models; weighting interpolation processing is carried out on the wave field separation results of all the reference models in the spatial domain to obtain the wave field separation results of the initial models. According to the method, on the basis of ensuring the accuracy of the wave field separation results, the effect of reducing the calculation amount is achieved.

The invention discloses a brain electrical signal independent component extraction method based on convolution blind source separation. The brain electrical signal independent component extraction method based on the convolution blind source separation includes concrete steps: building a brain electrical signal independent component extraction system based on the convolution blind source separation, which comprises an AD (analog to digital) sampling module, a short time Fourier transformation module, a frequency domain instantaneous blind source separation module, a sequence adjustment module and a short time inverse Fourier transformation module; using the AD sampling module to sample brain electrical signals; using the short time Fourier transformation module to transform the brain electrical signals from a time domain to a frequency domain; using the frequency domain instant blind source separation module to separate instantaneous mixing signals in the frequency domain; using the sequence adjustment module to perform sequence adjustment on independent components in a vector on each frequency domain segment; using the short time inverse Fourier transformation module to transform a frequency domain separation result into an independent component on the time domain. The brain electrical signal independent component extraction method based on the convolution blind source separation extracts the independent components of brain electrical signals based on a true convolution mixing model, uses a convolution blind source separation frequency domain algorithm, and is simple to achieve, good in separation effect, and low in calculation complexity.

This invention involves separation of brown algae; is a method of extract phycophacin from brown algae. The specific process is: brown algae as raw material, by 60-100% ethanol extraction, concentration, filtration to gain precipitation, macroporous resin column chromatography, spray drying to obtain products. Products content of 1% -20% (HPLC: High Performance Liquid Chromatography, HPLC). This invention is simple, separation result nice, and suitable for industrial production. The invention also involves brown algae in the health-food products, food and cosmetics field applications.

The invention discloses a gear case coupling modulation signal separation method based on inner product transformation and correlation filtering. According to the method, on the basis of the structure parameter (like the number Z of teeth) and the operational state (like the rotating frequency fn), a parsing dictionary Ds(t) capable of representing a stable modulation signal is constructed; an inherent frequency fd and a damping ratio Xi that correspond to the impact are identified in a gear case vibration acceleration signal x(t) by using a correlation filter method and an adaptive dictionary Dc(t) capable representing an impact modulation signal is constructed; and according to inner product transformation and correlation analyzing of the x(t), the Ds(t), and the Dc(t), sparse decomposition is carried out on the x(t) so as to obtain a primary function DS(t) and a coefficient Ai that are capable of expressing a stable modulation signal as well as a primary function DC (t) and a coefficient Bi that are capable of expressing an impact modulation signal, and separation is carried out to obtain a stable modulation signal Dpw(t) that is equal to sigma Ai DS (t) and an impact modulation signal Dcj(t) equal to sigma Bi DC (t). The method for constructing a parsing dictionary has advantages of both the parsing method and the adaptive method; and the physical significance is explicit. No complete orthogonality is required and the universality is high. The coupling modulation separation result is precise and reliable; and the anti-noise performance is good. A defect that the weak impact can not be extracted easily according to the traditional method can be overcome.