227results about How to "Easy to describe" patented technology

A system to measure fluid flow characteristics in a pipeline, meter, and methods includes a pipeline having a passageway to transport flowing fluid therethrough, a process density meter including at least portions thereof positioned within the pipeline to provide flowing fluid characteristics including volumetric flow rate, fluid density, and mass flow rate of the flowing fluid, and a fluid characteristic display to display the fluid characteristics. The process density meter includes a vortex-shedding body positioned within the pipeline to form vortices and a vortex meter having a vortex frequency sensor to measure the frequency of the vortices and to determine the volumetric flow rate. The process density meter further includes a differential pressure meter positioned adjacent the vortex-shedding body to produce a differential pressure meter flow rate signal indicative of the density of fluid when flowing through the pipeline. The process density meter also includes a thermal flow meter positioned adjacent the vortex-shedding body to produce a mass flow rate signal indicative of the mass flow rate of fluid when flowing through the pipeline. The process density meter produces an output of a volumetric flow rate, a flowing fluid density, and a mass flow rate to be displayed by the fluid characteristic display.

A system to measure fluid flow characteristics in a pipeline, meter, and methods includes a pipeline having a passageway to transport flowing fluid therethrough, a process density meter including at least portions thereof positioned within the pipeline to provide flowing fluid characteristics including volumetric flow rate, fluid density, and mass flow rate of the flowing fluid, and a fluid characteristic display to display the fluid characteristics. The process density meter includes a vortex-shedding body positioned within the pipeline to form vortices and a vortex meter having a vortex frequency sensor to measure the frequency of the vortices and to determine the volumetric flow rate. The process density meter further includes a differential pressure meter positioned adjacent the vortex-shedding body to produce a differential pressure meter flow rate signal indicative of the density of fluid when flowing through the pipeline. The process density meter also includes a thermal flow meter positioned adjacent the vortex-shedding body to produce a mass flow rate signal indicative of the mass flow rate of fluid when flowing through the pipeline. The process density meter produces an output of a volumetric flow rate, a flowing fluid density, and a mass flow rate to be displayed by the fluid characteristic display.

The prostate gland or other region of interest is stimulated with laser light, resulting in ultrasound waves (photoacoustic effect) which are focused by an acoustic lens and captured by a specific 1- or 2D sensor array and subsequently displayed as a C-scan on a computer screen. The amplitude of the ultrasound waves generated by laser stimulation is proportional to the optical absorption of the tissue element at that spatial location. Variability in tissue absorption results in C-scan image contrast. The photoacoustic imaging is combined with an ultrasound C-scan image produced with a plane ultrasound wave applied to the region of interest.

The invention relates to a convolutional neural network-based user attribute inference method and apparatus. The method comprises the steps of establishing a self-centered network according to attributes of user nodes and a friend relationship; and extracting attribute information of the user nodes in the self-centered network and hidden information comprised in the friend relationship by adoptinga convolutional neural network, and inferring missing attributes of a user by utilizing the hidden information. For a social network which cannot directly obtain the friend relationship or has relatively high obtaining difficulty, the missing attributes are subjected to classification prediction by only utilizing the attribute information of the user through adopting the neural network. The limitation of manual definition of a similarity function can be well avoided; and relationships among different attributes and different attribute dimensions can be better showed through convolution operation of a convolution kernel, so that the missing attributes of the user can be inferred efficiently and accurately.

The invention relates to a binocular initiative vision monitoring device for precision machining. The feature is that it fixes three supports on base, and sliding rails are installed in the three supports and plumbing to bottom plane, sliding blocks are installed in the sliding rail, the round guiding rail connecting to the three sliding blocks, two dollies installed on the round guiding rail, each dollies having a cradle head that is installed a video camera. The base line and location between the two video cameras would be adjusted by sliding on the guiding rail. The vision system could flexibly adjust the observing anger. It has the advantages of parallel structure, simple moving mode and simple vision computing mode.

The invention aims at providing a hyperspectral image classification method based on spatial and spectral features and sparse representation. The method comprises the following steps that: hyperspectral high-dimension data is read, dimension conversion is carried out, and the hyperspectral high-dimension data is subjected to normalization processing, wherein the contained sample class number is L; the image spatial vein features and the spectral features are respectively extracted to obtain the spatial vein features T1 and the spectral features T2; a spatial and spectral feature set T of images is obtained through merging the spatial vein features and the spectral features, wherein T={T1, T2}; a part of samples is respectively selected from T for L classes to form a training set A, and a test set is set to a state that the set of all of the L classes in the T is M; the M and A are utilized for solving a sparse representation coefficient S<^> of the hyperspectral data for image reconstruction, and in addition, corresponding redundancy in each class is calculated; and the classes of the samples are determined according to the redundancy. The hyperspectral image classification method has the advantages that the information in hyperspectral images can be sufficiently utilized, and the hyperspectral images can be perfectly depicted through the spatial and spectral features; the classification precision can be improved; and the method can be applicable to different hyperspectral images, and the applicability is high.

The invention discloses a commodity purchase prediction modeling method. The method comprises the steps that a purchase record marking training sample is used to predict whether to purchase or not; a sliding window commodity purchase sample is constructed; commodity purchase features are designed based on a time preference; a gradient improvement decision tree algorithm is used for training prediction; after the sample and the features are constructed, feature processing and selection need to be performed, and then the features are input into the gradient improvement decision tree algorithm for training prediction; and feature selection indicators include feature value distribution and relevancy, feature information gains, feature calling frequency, influences of feature knockout, etc. Ordering is performed on feature importance by integrating the indicators, and redundant features with low importance are eliminated. According to the method, a sliding window sample construction method and a feature system based on the time preference are proposed, the accuracy of a commodity purchase prediction model is effectively improved, and the method is used for realizing commodity personalized recommendation in a big data background to precisely recommend proper commodities to a user at a proper time and a proper place.

The invention discloses a lithium ion storage battery temperature combinational circuit model including an energy balancing circuit and a voltage responsive circuit. The energy balancing circuit predicts the continuous operation time and charged state SOC of a battery, and the voltage responsive circuit stimulates the transient response of the battery, and elements of the circuit are variables and change continuously according to the change of the environment temperature and SOC. The parameter identification method of the lithium ion storage battery temperature combinational circuit model includes four steps of charged state and battery open-circuit voltage relation identification, battery capability and temperature relation identification, self-discharge resistance identification, and internal resistance and polarized parameter identification. The lithium ion storage battery temperature combinational circuit model and the parameter identification method thereof make the stimulation approximate to the real response of the storage battery at different temperatures and give guidance to battery state estimation, and have the advantages of distinct characteristic, visual construction, clear physical meaning, convenient parameter identification, and easy application in engineering.

The invention discloses a QR Code image super-resolution reconstruction method based on sparse representation, which solves the problem of super-resolution reconstruction of low-resolution QR Code two-dimensional code images. Using the edge gradient feature and texture feature of the QR Code two-dimensional code, the operator extracts the image feature to obtain the feature block; through the dictionary learning algorithm, obtain the learning dictionary of the low-resolution image block and the corresponding high-resolution image block; use sparse representation Theoretically, sparsely encode the input low-resolution image feature blocks, and combine the learning dictionary to obtain the high-resolution image blocks corresponding to the input low-resolution image blocks; apply global constraints to synthesize all high-resolution image blocks into the final high-resolution QR Code two-dimensional code image to achieve super-resolution reconstruction. The invention utilizes image characteristics to effectively reconstruct a high-resolution QR Code two-dimensional code image with clear edges and retains a large number of high-frequency details, and is suitable for super-resolution reconstruction of the QR Code two-dimensional code image.

The present invention relates to a microwave radiation antenna capable of making percutaneous puncture to cure tumor of several regions of human body. Said invention is formed from antenna connector, hard coaxial line, ceramic tube, water inlet tube, water outlet tube, temperature-measuring system, stainless steel sleeve and handle. Its output power can be up to above 100w, and is therapeutic effect is reliable and safe.

The prostate gland or other region of interest is stimulated with laser light, resulting in ultrasound waves (photoacoustic effect) which are focused by an acoustic lens and captured by a specific 1- or 2D sensor array and subsequently displayed as a C-scan on a computer screen. The amplitude of the ultrasound waves generated by laser stimulation is proportional to the optical absorption of the tissue element at that spatial location. Variability in tissue absorption results in C-scan image contrast. The photoacoustic imaging is combined with an ultrasound C-scan image produced with a plane ultrasound wave applied to the region of interest.

The invention discloses a plant leaf area lossless measuring method. According to the plant leaf area lossless measuring method, a leaf blade image collection method and an indoor leaf area measurement calculation method are mainly used for carrying out measurement and calculation; according to the leaf blade image collection method, a digital camera, a shooting background base plate, coordinate paper, a steel tape and other tools are used as tools, and are easy to operate and convenient to carry, especially when shooting is carried out on the situation that the shooting resolution and the image storage pixel level set when the digital camera leaves the factory are utilized, and the shooting angle is close to 90 degrees, a clear leaf blade image can be obtained, and leaf area measuring accuracy does not have significant difference; according to the indoor leaf area measurement calculation method, the shot leaf blade image is stored in a computer preloaded with AutoCAD software, and the leaf area of a plant is measured and calculated by running the AutoCAD software. Lossless measurement of the leaf area can be achieved by combining the digital camera and the AutoCAD software; compared with a popular leaf area determinator at present, the plant leaf area lossless measuring method has the advantages of being easy and convenient to operate and low in cost.

The invention discloses a method for reconstructing a compressively sensed spectral image based on structural clustering sparse representation, which method solves the difficulty of the existing method for reconstructing the spectral image that a partial structure of the spectral image is hard to be recovered accurately since spatial and spectral correlations are not fully used. The method comprises implementation steps as follows: 1, performing back projection upon input encoded sensing data of an spectral image to obtain an initially reconstructed spectral image; 2, segmenting the reconstructed spectral image to obtain a series of overlapped three-dimensional spectral image blocks; 3, de-noising the three-dimensional spectral image blocks by using a sparse representation method based on structural clustering; 4, recovering the whole spectral image by using the spectral image blocks subjected to de-noising procession; 5, updating the spectral image by using a back projection technology, iterating the steps 2-5 so as to obtain a final reconstruction result. Experiment results show that the method disclosed by the invention can reconstruct a relatively fine spectral image structure, and the reconstructed spectral image has a relatively high signal to noise ratio.

The invention discloses a rapid and high-precision forward modeling method for a gravitational field of an arbitrary density distribution complex geological body. The rapid and high-precision forward modeling method achieves the unification of efficiency and precision of gravitational field forward modeling through the steps of complex geological body model representation and prism body combined model gravitational field calculation (including weighted coefficient calculation, two-dimensional discrete convolution calculation and gravitational field value synthesis). The rapid and high-precision forward modeling method solves the problems that the existing gravitational field forward modeling method cannot guarantee calculation efficiency and calculation precision simultaneously, and cannot meet the requirements of large-scale gravitational field three-dimensional density inversion and man-machine interaction modeling and interpretation.

The invention provides a source code vulnerability detection method based on deep learning, and the method comprises the steps: automatically completing the feature extraction of a source code based on deep learning, and constructing a vulnerability detection model by using a random deep forest algorithm in combination with a code measurement index and an automatically extracted source code feature. The source code vulnerability detection method based on deep learning provided by the invention has higher degree of automation, reduces dependence on domain expert knowledge, greatly reduces the code auditing cost and improves the code auditing efficiency. Compared with other methods for vulnerability detection by using deep learning, , grammar and semantic information of the code are reservedto the maximum extent by combining multiple representations of the code, the code can be better depicted by the characteristics automatically extracted by the deep learning algorithm, and meanwhile,the detection effect is further improved by combining common code measurement indexes as detection characteristics.

The invention relates to an RBF neural network indoor positioning method based on sample clustering. According to the method, signal packet loss rates under different transmitting power are taken as basic data, a clustering algorithm is adopted for screening out a training sample set with similar feature points, then the sample set is trained through an RBF neural network, and finally the position coordinate of an unknown mobile node is predicted. Due to the relation of communication distances and the packet loss rates, the sample set of the RBF neural network indoor positioning method is rich in information, and relation of signals and the distances can be depicted better; meanwhile, the clustering algorithm is adopted for screening out the position similar feature points and RBF neural network training data, so that data under large-scale and large-range conditions are convenient and easy to collect, the purpose of practicability is achieved really, and meanwhile the algorithm has the advantages of being high in convergence rate, accurate in positioning and the like.

The invention provides an anti-jamming model in unmanned aerial vehicle communication and a Stackelberg game subgradient algorithm. The model is characterized in that in a process of unmanned aerial vehicle group communication, a jammer jams user communication and mutual jamming also exists in the unmanned aerial vehicle group; the jammer and the unmanned aerial vehicle group achieve the maximum utility by adjusting respective transmission power. The algorithm is characterized in that firstly a Stackelberg game model is built, wherein the game leader is a jammer, and game followers are unmanned aerial vehicle users; unmanned aerial vehicle users perform communication; the intelligent jammer adjusts own transmission power to the optimum according to the transmission power of the unmanned aerial vehicle users, and then the unmanned aerial vehicle users adjust respective transmission power to the optimum according to the transmission power of the intelligent jammer; power adjustment is conducted circularly and alternately until the unmanned aerial vehicle users and the intelligent jammer converge to the optimal transmission power. The model is complete and the physical significance isclear; anti-jamming scenes in unmanned aerial vehicle communication can be better described.

The invention discloses a method and system for detecting apnea events based on BCG signals, which are used to accurately detect apnea events during sleep. This method locates the potential occurrence interval of respiratory events by identifying the arousal segment, and then divides the event potential occurrence interval into apnea segment, respiratory effort segment, and arousal segment, and then extracts fine-grained features that can describe the breathing pattern from the three segments. , and finally with the help of machine learning methods, it is judged whether the potential occurrence stage of the event contains an apnea event. The system mainly includes: a signal acquisition module, a data processing module, and a detection result output module. The method and system can automatically and accurately locate the potential occurrence interval of apnea events and automatically divide the interval into three different stages, which facilitates the description of breathing patterns from multiple aspects and in a fine-grained manner, and greatly improves the detection accuracy of apnea events Rate.

The invention discloses a classifier integration method based on floating classification threshold, which is characterized by obtaining T optimal weak classifiers are by means of training after T iterations and then combining the T optimal weak classifiers to obtain an optimal combined classifier. In case of aiming at a bi-classification problem, training the T optimal weak classifiers comprises the steps of: (3.1) training the weak classifiers based on a training sample set S with weight omega<t>, wherein t is equal to 1,..., T; (3.2) based on the result of the step (3.1), adjusting sample weights omega<t+1>=omega<t>exp(-yiht(xi)) / Zt; (3.3) judging whether t is smaller than T, if so, enabling t to be equal to t + 1 and returning to the step (3.1) until t is equal to T; in case of aiming at multi-classification problem, training the T optimal weak classifiers comprises the steps of: (3.1) training the weak classifiers based on the training sample set S with weight omega<t>, wherein t is equal to 1,..., T; (3.2) based on the result of the step (3.1), adjusting sample weights shown in the description; (3.3) judging whether t is smaller than T, if so, enabling t to be equal to t + 1 and returning to the step (3.1) until t is equal to T. Compared with the prior art, the classifier integration method of the invention can overcome the defect that fixed classification threshold-based classifiers have unstable classification at points adjacent to classification boundary.

The invention discloses a form feature describing and indexing method of an image object, which comprises the following steps of: inputting a to-be-indexed image object P of a binary image form; carrying out form feature extraction on the to-be-indexed image object P; calculating the similarity of the to-be-indexed image object P and a to-be-compared image object Q in a system example base on the basis of the form feature description expressed by each layer of image object; and selecting one example or one group of examples with highest similarity with the to-be-indexed image object P as an index result to be output from the system example base. The form feature description integrates form features of multiple layers of image objects, has better describing and indexing property, is flexible and adjustable in the indexing precision and the algorithm complexity, can adapt the processing demands of a larger range of object types, can effectively process the influences of interference conditions of noise, local deformation and the like in the image indexing, and is suitable for various image indexing and recognition processing software systems.

The invention provides a depth estimation method, device and system, and a computer readable storage medium. The method comprises the steps of: correcting and transforming the acquired color image andinfrared image respectively to obtain an undistorted virtual perspective image; estimating the parallax between the corrected and transformed color image and the corrected and transformed infrared image to obtain a parallax image; filtering the parallax image; and inversely correcting and transforming the filtered parallax image to obtain a depth image aligned with the original color image. The depth estimation method, device and system can estimate the depth on a mobile terminal, and can accurately obtain the depth information of a scenario.

The invention discloses a wind power prediction method based on a variable mode decomposition and a long-short memory network. Firstly, the wind power signal is decomposed into three modal componentssuch as a long-term sub-component, a fluctuation number and a random component according to the characteristics of the wind power. Then, the long-short memory network is used to learn the modal components of wind power, and the prediction model is established. Finally, the prediction values are synthesized. The method can obtain the random characteristics of wind power and improve the accuracy andstep size of wind power forecasting.

The invention discloses a detection system for tool chattering in milling and a detection method thereof, characterized by comprising following steps: 1), selecting a suitable displacement sampling window; 2), determining a center of a tool and vibration changes corresponding to the center of the tool by using a radius constrained least square method; 3), subjecting a tool vibration signal to empirical mode decomposition, and obtaining a time-frequency spectrum using HHT (Hilbert Huang transform); 4), verifying whether the time-frequency spectrum meets chattering rule. The invention has the advantages that the use of a non-contact detection method, using a laser displacement sensor, overcomes limitations such as size of a workpiece under processing, mass, a mounting mode, and high precision is provided; by studying the influence of milling chattering and tool eccentricity upon positional changes of the center of the tool, positional flutter of the center of circle of the tool is determined, and thus chattering change conditions of the tool are directly reflected; by using Hilbert Huang transform, the limitations of Fourier transform can be overcome and a time-frequency spectrum of signals can be better described.

The invention discloses a heterogeneous information network and Bayesian personalized sorting-based recommendation method and apparatus. According to the method, a heterogeneous information network is used to describe a user, articles and related information, and an iterative method based on random walk with restart is used to calculate scores of the user to the articles. For weights of edges in the heterogeneous information network, a Bayesian personalized sorting-based machine learning method is used for automatic learning. In the learning process, an optimal solution is selected from a plurality of candidate solutions obtained by multi-round iteration through multi-time iterative solving. The invention furthermore discloses the heterogeneous information network and Bayesian personalized sorting-based recommendation apparatus. According to the recommendation method and apparatus, the weights of the edges in the heterogeneous information network can be automatically learned from data, the adaptability to different data is greatly improved, and user preferences can be better described, so that a better personalized recommendation result is obtained.

The invention provides a quick and high-precision numerical simulation method for a two-dimensional magnetic field of a strong magnetic body. According to the method, through two-dimensional model expression of the strong magnetic body, Gaussian parameter design, calculation of the discrete offset wave number, weighting coefficient calculation of a wave number domain, calculation of magnetizationintensity, one-dimensional discrete Fourier inversion, magnetic abnormality calculation of a spatial domain, iteration convergence judgement and other steps, unification on efficiency and precision ofnumerical simulation of the two-dimensional magnetic field of the strong magnetic body is achieved. According to the method, the problems are solved that an existing numerical simulation method for the magnetic field of the strong magnetic body is low in calculation precision and cannot meet fine inversion imaging of large-scale strong magnetic measurement data, and the method helps to carry outresearch of two-dimensional magnetic susceptibility fine inversion imaging of the large-scale strong magnetic measurement data, man-machine interaction modeling and interpretation.

The invention provides a method and a device for geologic random inversion based on multiple-point geostatistics, and relates to the technical fields of geology and petroleum exploration. The method includes the following steps: determining an initial lithofacies model by a multiple-point geostatistical stochastic simulation algorithm; performing probabilistic disturbance updating to generate a probabilistic disturbance updated lithofacies model; counting the distribution functions of reservoir physical parameters in different lithofacies, and converting the probabilistic disturbance updated lithofacies model into a pseudo-reservoir physical property parameter model; converting the pseudo-reservoir physical property parameter model into an elastic parameter model; carrying out time-depth conversion to obtain a time domain elastic parameter model, determining a reflection coefficient, and synthesizing a seismic record through forward modeling a convolution model; comparing the seismic record with a pre-determined actual seismic record, and determining a lithofacies model result and a pseudo-reservoir physical property parameter model result according to a maximum disturbance numberand a maximum simulation number; and carrying out quantum annealing optimization on the pseudo-reservoir parameter model result to generate a reservoir physical property parameter inversion result.

The invention provides a second generation curvelet transform-based static human detection method, which mainly solves the problem of high detection false-alarm rate in the conventional human detection technology. The detection process comprises the following steps of: acquiring a negative sample through bootstrap operation of the negative sample, and forming a training sample set by the negativesample and other positive samples in a database; calculating curvelet transform-based feature vectors of all training samples to form a training sample feature set; performing classification trainingon the sample feature set by adopting an AdaBoost algorithm to obtain a classifier; inputting a to-be-tested image with any size, calculating curvelet transform-based feature vectors of all scan window images in the to-be-tested image; inputting the curvelet transform-based feature vectors of all scan window images into the obtained classifier for classification; and according to classification results, combining all scan windows classified as human by utilizing a main window merging method to form the final human detection result. The method has the advantages of high detection accuracy and low false-alarm rate, and can be used for classifying and detecting human in the image.

The invention relates to a data recommendation method based on face recognition, a device, a server end and a client end. Acquiring a stored face feature vector corresponding to the user identification according to the user identification; A human face feature vector is extracted from a human face image, whether the extracted human face feature vector matches the stored human face feature vector is judged, and if the extracted human face feature vector matches, recommendation data corresponding to the stored human face feature vector is obtained, and the recommendation data is sent to a smallprogram for display. By using face recognition as the basis of data recommendation, Convenient to use, effectively prevent information leakage problem, can better portray the user portrait, through the combination of small procedures for data recommendation, without installation and uninstallation application, can achieve rapid data analysis and processing, reduce the occupation of resources, reduce the development cost and development threshold, can be widely used in life service scenarios.

The invention discloses a near-infrared-spectrum-based tobacco leaf part feature extraction and discrimination method, which comprises: providing K tobacco leaf samples, collecting the spectrums of the samples through a near-infrared spectrometer, and pre-treating; sampling N times, and calculating the correlation p between each wavenumber point and the part in the sample and the ratio d of the distance in the same part to the distance between the parts in each wavenumber point sample; recording the correlation matrix P and the distance matrix D after the n sampling, and calculating the average values of P and D and the standard deviations Pm, Ps, Dm and Ds; determining the wavenumber points meeting threshold conditions by giving thresholds t1, t2 and t3; combining the wavenumber points meeting threshold conditions as feature points, and modeling by using the spectrums of the feature points and the part labels; and collecting the near-infrared spectrum of the sample to be determined, predicting by using the established model, and determining the tobacco leaf part of the sample to be determined. According to the present invention, the modeling is performed by screening the feature wavenumber points related to the tobacco leaf part in the spectrum so as to identify the part of the tobacco leaf in the same production place.

The invention provides a regional traffic signal lamp control method based on a graph neural network. A traffic flow predictor and a traffic signal lamp controller are trained at the same time, a future traffic flow change prediction value under a current intervention action is predicted by using the traffic flow predictor to help the traffic signal lamp controller to generate a new control scheme, the evaluation information of the traffic flow predictor to a value of the current action is used to assist in training the traffic signal lamp controller to maximize the long-term and short-term earnings of the traffic signal lamp control scheme. The traffic flow predictor and the signal lamp controller are built based on a depth message propagation graph network. According to the method, a system can be continuously optimized to adapt to changing traffic flows, and the road network smoothness degree and the traffic efficiency are improved.