93results about How to "Calculation time is short" patented technology

The invention discloses a three-dimensional structure sound source radiation sound field forecast method under the shallow sea channel and belongs to the sound source radiation sound field forecast field. The method is characterized in that firstly, a structure sound radiation multi-physical field coupling model under the shallow sea channel is established, the sea sound propagation theory is selected according to analysis frequency range, calculation distance and shallow sea environment factors to establish a two-dimensional axisymmetric channel transmission function G, through orientation coupling solution development, a channel three-dimensional sound field transmission function is acquired, secondly, the shallow sea channel multi-boundary sound field coupling calculation theory is employed to establish a multi-physical field coupling model, a structure sound source surface speed U under the shallow sea channel under the action of multi-boundary influence is acquired through calculation, and lastly, the Fourier wave superposition method theory is utilized to forecast the sound field information P of a structure sound source at any field point under the channel is forecasted. Themethod is advantaged in that multiple bottleneck problems of high computational complexity, multiple coupling physical fields and complex channel environments existing during researching three-dimensional elastic structure sound source radiation sound field characteristics under the complex shallow sea channel in the prior art are effectively solved.

The invention relates to a stud connection-based steel-concrete composite beam calculation model analysis method. The method comprises the following steps of 1, building a stud shear connector composite beam calculation analysis model; 2, proposing that the overall shear stiffness Dsc of a stud comprises the shear stiffness Ds of the stud and the stiffness Dc of concrete in the range of a concentric circle with the diameter nds; 3, introducing a reduction coefficient beta of a working state of the overall shear stiffness of the stud to obtain equivalent lateral stiffness Deq; and 4, accordingto a lateral stiffness equivalence principle, making the stud equivalent to a short steel beam unit, and proposing that different values of the equivalent diameter deq of the short steel beam unit areused for simulating the mechanical performance of the stud in different working states. The method has the advantages that the calculation process is high in efficiency; a calculation result can quantitatively reflect interface slip of a steel-concrete composite structure and stress states of parts in stress periods of the composite structure; the practicality is high; and the method is easily understood and mastered by structure designers.

The invention discloses a robot hand-eye calibration method based on a non-minimized and non-optimized algorithm. The method is based on the non-minimized and non-optimized algorithm, calibration can be implemented by feature value calculation, and an initial value does not need to be set. The robot hand-eye calibration method comprises the following steps of selecting suitable hand-eye movement data by using a movement selection algorithm; and performing the non-minimized and non-optimized algorithm, namely, establishing a linear equation of a hand-eye calibration system by using a tensor product to obtain an optimized cost equation and solving an optimal solution of a feature matrix (y) by feature value calculation. By the robot hand-eye calibration method, feature value calculation replaces the traditional 2-norm computing, calculated amount is reduced, and problems caused by improper selection of the initial value are avoided. In a noise environment, rotation errors and translation errors of the method can meet ordinary application requirements.

The invention provides a coupled FEM-PE-based sound radiation prediction method of an elastic structure under an ocean channel. Firstly, a multi-physical-field coupling theory model of sound radiationof the structure under an ocean channel environment is established; a finite-element method (FEM) is utilized to acquire a radiation sound field on one-dimensional transversals of a depth direction;and a piecewise cubic Hermite interpolation method is adopted to carry out perfect matching, namely a FEM-PE coupling condition, of FEM sound field information and grid points of a parabolic-equation(PE) method on space, then the radiation sound field information is used as an initial condition of PE finite-difference-method (FDM) calculation, and then quick prediction of the sound field at any field point of the elastic structure under the ocean channel is carried out after setting related parameters of the PE method. According to the method, adaptability on the elastic structure and the ocean channel is high, a calculation result is accurate, efficiency is high, use is simple, and promotion is easy, and many bottleneck problems that large calculation amounts, multiple physical fields, complex channel environments and the like currently encountered when sound radiation prediction of elastic structures under ocean channels are studied are effectively solved.

The invention discloses an image recognition method and apparatus based on a multilayer convolutional neural network, an image acquisition method and apparatus, computer equipment, and a non-volatilecomputer readable storage medium. According to the image recognition method and apparatus based on the multilayer convolutional neural network, the image acquisition method and apparatus, the computerequipment, and the non-volatile computer readable storage medium, a multilayer convolutional neural network model including three convolution layers and two pooling layers is constructed, the multilayer convolutional neural network model is trained for a training image of a first resolution by employing resolution normalization, the multilayer convolutional neural network model is tested for a test image of a second resolution by employing resolution normalization, the recognition of an image scene can be realized without using a full connecting layer, the complexity of a scene recognition algorithm is reduced, the calculating amount of scene recognition is small, and the time consumption of calculation is low.

The invention discloses a voice DOA estimation method based on ResNet, which comprises the following steps of: 1, simulating a training data set by using MATLAB, traversing a measurement range by using a plurality of voice signals in the data set, and storing corresponding angles and voice signals; 2, after each simulation signal is subjected to framing processing, calculating GCC and performing phase transformation; cutting according to the array model parameters, weighting and summing each voice frame; storing the weighted features and the corresponding incident angle as the data set; 3, initializing ResNet by using MATConvNet and training by using the data set; 4, carrying out coarse positioning on the signal to be measured by using broadband MUSIC to obtain a coarse positioning result,and selecting a group ResNet with a center point closest to the broadband MUSIC result to carry out subsequent accurate positioning according to the coarse positioning result to obtain a DOA estimation result. The method can effectively solve the problem of inaccurate voice DOA estimation under the condition of strong noise reverberation, and is a DOA estimation method suitable for any array structure.

The invention discloses a gait parameter measurement system and a data processing device and a data processing method thereof. The method comprises the following steps of: acquiring initial positions of a marking point in two image coordinate systems corresponding to two video acquisition devices; synchronously acquiring images from videos output by the two video acquisition devices; respectively determining the positions of the marking point in the two synchronously acquired images through motion estimation and block matching search; calculating the three-dimensional coordinate of the marking point through the positions of the marking point in the two synchronously acquired images according to a double-camera imaging theory; and calculating a gait parameter according to the three-dimensional coordinate of the marking point. The device comprises five modules corresponding to the steps. The system comprises the data processing device, the marking point and the two video acquisition devices, wherein the marking point is arranged on a leg of measured personnel; and the two video acquisition devices are respectively connected with the data processing device and used for respectively capturing the walking scene of the measured personnel according to two different positions in the front of the measured personnel. The system and the device are low in cost and portable.

The embodiment of the invention provides a camera external parameter calibration method and device and electronic equipment, and the method comprises the steps: carrying out the distortion correctionof an original image of a calibration image based on the internal parameters of a camera, and obtaining a distortion correction image; obtaining pixel coordinates of four intersection points formed bythe four straight lines in the distortion correction image; and calibrating camera external parameters through a preset coordinate transformation strategy based on the pixel coordinates of the four intersection points. Through the scheme, the external parameter calibration speed of the camera and the accuracy of the calibration result can be improved.

The method divides sound field into a numbers of regions and to present sound field at each region as linear superposition of region node sound pressure according to sound pressure at several known measuring points in certain sound field of known sound source shape. The regional node sound pressure can be obtained according to sound pressure of measuring point through least square method and singular value decomposition, and then sound pressure for the complete sound field can be rebuilt through sound pressure of regional node. The method is more accurate and efficient than HELS method in describing sound field.

The invention relates to a reinforced concrete simply supported beam fire damage identification method based on an improved support vector machine, and belongs to the technical field of fire damage identification. The method comprises the following steps: (1) establishing a simply supported beam fire damage model, and obtaining modal parameters of the reinforced concrete simply supported beam at each fire moment; (2) constructing combination parameters of the training samples based on the SVM; (3) defining an output vector of the FA-SVR as a fire receiving time prediction value; (4) the support vector regression machine tool box called in the FA-SVR algorithm having a regression index, and the damage identification effect of the model is measured; and (5) carrying out regression predictionby virtue of an SVM, and carrying out optimization by adopting a firefly algorithm. According to the method, the diagnosis efficiency and the convergence speed are remarkably improved, the complexityof the adjustment model and the experience risk are reduced, the calculation is more accurate, the calculation time consumption is short, the parameter optimization capability is improved, the experience error is greatly reduced, and the result reliability is higher.

The invention discloses a GPU parallel-based fast density inversion method of gravity gradient tensor data. Compared with single gravity data inversion, the method is higher in resolution. Compared with a traditional conjugate gradient method, a provided fast preprocessing method is closer to a set model in a density value and a geometric position. Efficiency of a parallel program is evaluated collectively through an iteration number and total computation time, and compared with serial CPU computation of a traditional algorithm, the method enables three-dimensional inversion to converge faster, is short in time used for computation, and achieves approximately a speed-up ratio of 25 times.

The invention discloses a lithium battery electrode surface defect detection method based on background standardization and a centralized compensation algorithm, being implemented by a measurement system constructed by lighting equipment, a linear array camera, a conveying device, a labeling machine, an industrial personal computer and a PLC. The lithium battery electrode surface defect detectionmethod comprises the following specific implementation steps: acquiring a lithium battery electrode surface image by using the linear array camera; preprocessing the electrode surface image by using abackground standardization algorithm; carrying out defect rough detection, and carrying out automatic threshold segmentation on the preprocessed electrode surface image to obtain a possible defect area; and finally, carrying out accurate defect detection, and carrying out accurate detection on areas where defects possibly exist by using an automatic centralized compensation algorithm so as to obtain defect images. According to the lithium battery electrode surface defect detection method provided by the invention, on the premise of ensuring the lithium battery electrode surface defect detection effect, the requirement of online real-time defect detection in the actual lithium battery industrial production process can be met, and the production efficiency of the lithium battery is improved.

The invention provides a power distribution network two-stage reconstruction method and device based on an ordered optimization algorithm, and relates to the field of electric power. The method comprises the following steps: obtaining load data in a future time period; carrying out load clustering on the load data in the future time period and constructing a static reconstruction model of a powerdistribution network; obtaining an optimal solution corresponding to each clustering center; within the future period of time, carrying out short-time prediction of load data on the power distributionnetwork; with the dissimilarity degree as an index, calculating the dissimilarity degree between the short-time prediction load data and the load data of the C clustering centers; determining the static reconstruction model corresponding to the clustering center with the minimum dissimilarity degree as a power distribution network reconstruction result in a first preset time period; and reconstructing the power distribution network according to the result. According to the method, the network loss after dynamic reconstruction of the power distribution network in actual operation is minimized,the reconstruction calculation amount is small, the operation is simple, the reconstruction calculation time is short, the dynamic change load can be well adapted, and the method has high practical value.

The invention discloses a method for rapidly generating a grait and gravel mixed concrete two-dimensional random aggregate model. A concrete two-dimensional random aggregate model only with grait or gravel is the special circumstance of the grait and gravel mixed concrete two-dimensional random aggregate model. The method comprises the steps that a circular random aggregate model and a Thiessen polygon with the boundary are combined, and the circle center of the circular aggregate is adopted as a control point to generate the Thiessen polygon similarly meeting the aggregate gradation; the peak of each Thiessen polygon is adopted as a control point, a closed B spline is generated in parallel to represent the shape of the grait aggregate, points are sampled randomly and uniformly on the closed B spline curve, and therefore the polygon is generated to represent the shape of the gravel aggregate. By adopting the parallel mode, area zooming is performed on the aggregate inside each Thiessen polygon, and the grait and gravel freely mixed concrete two-dimensional random aggregate model meeting the given gradation is rapidly generated.

The invention provides a coal seam floor water abundance prediction method based on multi-attribute inversion, and belongs to the field of coal exploration. The method comprises the step of extracting multiple seismic attributes; the step of performing mathematical manipulation on apparent resistivity and the seismic attributes and performing correlation intersection so as to acquire sensitive seismic attributes; the step of screening the sensitive seismic attributes to obtain an optimal model; the step of obtaining the apparent resistivity body according to a neural network training model and the three-dimensional seismic data of an area to be predicted; and the step of selecting the datum plane to perform slice extraction of the apparent resistivity body. The method has high adaptability for various seismic data, short calculation time consumption and wide range of application.

The invention relates to an artificial intelligence technology. The invention discloses a large-scale data clustering method. The method comprises the following steps: receiving a data sample set input by a user, calculating an average contour coefficient of a cluster center number K value according to the data sample set, selecting the K value with the maximum average contour coefficient, randomly determining K cluster centers, and storing the K cluster centers and the data sample set into a database according to a row priority storage form; calculating loss values of the K cluster centers and the data sample set according to a minimum square error algorithm, and judging a size relationship between the loss values and a preset threshold value; and when the loss value is smaller than the threshold value, outputting the K cluster centers to complete a clustering result. The invention further provides a large-scale data clustering device and a computer readable storage medium. Accordingto the invention, an accurate large-scale data clustering function can be realized.

The invention discloses a power grid fault diagnosis method considering false faults caused by measurement tampering attacks. The method comprises steps that firstly a suspected fault element is determined through a junction line analysis method, and the diagnosis range is reduced; secondly, whether each suspected fault element is subjected to a measurement tampering attack or not is judged by adopting a random matrix theory and a fuzzy C-means clustering algorithm, so that the problem of fault diagnosis misoperation caused by the measurement tampering attack is effectively solved. Finally, afault diagnosis model capable of comprehensively utilizing telemetry and telesignalling is established by utilizing a memory backtracking thought of a memory neural pulse membrane system, so the defect that modeling cannot be performed by utilizing the telemetry on the basis of an original production rule is overcome, and the fault type can be determined under the condition that a fault element isdiagnosed.

The invention discloses a power flow calculation method of a loop-network type power distribution system. The power flow calculation method comprises the following steps: building a circuit mode according to the characteristic of the loop-network type power distribution system; numbering nodes and branches of the circuit mode; representing initial data, specially data description of a circuit topological structure, of circuits by utilizing a matrix form to ensure that logical relations of graphics connection in the circuit mode are clear and definite; realizing programmed design of a path searching process due to introduction of a correlative matrix; using the initial data matrix as unique data input so that the whole program is designed directly and simply; reasonably expanding a loop ofthe loop-network type power distribution system by utilizing the knowledge of a graphic theory in a calculation process, and equivalently carrying out calculation for a radiation type power distribution system so that the calculation process is greatly simplified; and solving power flow of the various branches and voltage of various load nodes of a power distribution network by mainly utilizing algebraic operation of matrix elements, and therefore, the program calculation time is less, and the convergence speed is higher.

The invention discloses an HBF system beam selection method and device and a computer readable storage medium. The method comprises the following steps of respectively selecting at least one preselected beam with the maximum projection energy for each user according to the projection energy of a direct vision diameter of the user on the beam, wherein the number of all preselected beams is larger than the number of digital channels of an HBF system; and selecting a first number of target beams from all preselected beams according to a rate of arrival of the HBF system, wherein the first numberis the number of the digital channels. The HBF system beam selection method and device and the computer readable storage medium have the advantages of short time consumption of a calculation process and easiness in implementation, and the beam is selected from the global angle of a multi-user transmission system, so that the system has good overall performance.

The invention specifically discloses an early warning image ballistic missile direction estimation method under a single-satellite condition. The method comprises the following steps: S1, obtaining an original image of a target missile, and carrying out target detection on the obtained original image at a known time interval to extract a centroid coordinate of the target missile in the corresponding original image; s2, performing target tracking according to the center-of-mass coordinates of the target missile and the corresponding time information so as to obtain the motion trail of the target missile; s3, converting the motion trail attitude data of the target missile into an earth coordinate system, and finding out a corresponding trajectory point in the earth coordinate system; s4, establishing a normal vector equation set corresponding to the ballistic trajectory plane according to the ballistic trajectory points; and S5, establishing a coordinate equation of the target missile by using the collinear equation, and solving the missile direction according to the coordinate equation and the normal vector equation set. According to the invention, an accurate result can be obtained without ballistic prior knowledge, and the method has the characteristics that the required data is easy to obtain, the calculation amount is small and the calculation time is short.

The invention provides a space-time intelligent early warning method and system for soil landslide disasters. The method comprises the steps: acquiring landslide disaster point data and non-landslidedisaster point data; acquiring a landslide disaster early warning factor for intelligent learning according to the cause mechanism and the induction factor of the landslide disaster; acquiring original data corresponding to the early warning factor according to the landslide disaster point data and the non-landslide disaster point data; performing spatial analysis on the original data to extract values of landslide early warning factors for intelligent learning, and performing dimensionless processing on the obtained values of the early warning factors to construct a landslide disaster early warning sample data set; adopting an integrated machine learning stack generalization method to build an intelligent learning module used for landslide disaster early warning, wherein a landslide disaster early warning sample data set is input into the intelligent learning module to be trained and optimized, and early warning of landslide disasters is achieved. The method is high in accuracy, low in false alarm rate, short in calculation time and good in real-time performance during landslide prediction.

The invention relates to the field of automobiles, in particular to a CAE simulation analysis method for a car body aluminium alloy mortise-and-tenon joint structure. According to the method, beam elements are adopted to simulate an aluminium alloy profile, actual aluminium profile sections are adopted for section shape properties of the beam elements, spring units are adopted to simulate triangular plate welding reinforced fixed connection structures for a mortise-and-tenon intersecting structure of the aluminium alloy profile, and spring stiffness is based on experimental measurement. The method has the advantages that modeling is easy, preprocessing workload is small, the analysis result is accurate, calculation consumed time is short, and working efficiency is high.

The invention is suitable for the technical field of artificial intelligence, and provides a bionic robotic fish motion control method, device and equipment and a storable medium, and the method comprises the steps: constructing a kinematics model of a bionic robotic fish based on a central pattern generator; optimizing parameters of the kinematic model through a particle swarm optimization algorithm to obtain an optimized central pattern generator structure; and generating a joint angle according to the optimized central pattern generator structure so as to control the movement of the bionicrobotic fish. According to the optimized central pattern generator structure, stable output can be realized. In addition, the joint angle generated according to the optimized central pattern generatorstructure is short in time consumption and can be executed in real time. In addition, through computer simulation and bionic robotic fish actual test results, it is verified that the performance of the central pattern generator structure optimized through the particle swarm optimization algorithm is greatly optimized.

The invention relates to the technical field of alternating-current power transmission, in particular to a voltage sag level calculation method for a power system containing multiple STATCOM accessesunder a symmetric fault. The method comprises the following steps of: S1, establishing a STATCOM mathematical model for calculating a voltage sag level; S2, establishing a power system voltage sag level calculation model accessed by a plurality of STATCOMs under a symmetric fault; and S3, solving the voltage sag level of the power system containing a plurality of STATCOM accesses under the symmetric fault. According to the method, the blank of voltage sag level calculation of the multiple STATCOMs connected to the power system is filled, the method is accurate in calculation result and short in calculation time consumption, can be suitable for any node system and multiple STATCOM access conditions, and lays a theoretical foundation for STATCOM optimal configuration for voltage sag suppression.

The present invention provides a tobacco leaf image quality evaluation method and system, a storage, and an electronic device. The method comprises the steps of: extracting the component value of eachpixel of a tobacco leaf image to be evaluated in a color space, wherein the color space is formed by yellow color gamut components, bright color gamut components and a black color gamut components; calculating the color gamut indexes of the pixels according to the component values of pixels; determining the quality evaluation indexes of the pixel according to the color gamut indexes of the pixels; and calculating an average value of the quality evaluation indexes, comparing the average value with a preset quality evaluation index, and determining the quality condition of the tobacco leaf image to be evaluated. The tobacco leaf image quality evaluation method and system, the storage, and the electronic device employ the color gamut indexes to perform tobacco leaf image quality evaluation,and are lower in complexity, smaller in computing quantity, shorter in calculation time and better in quality evaluation result stability compared to a current tobacco leaf image quality evaluation method.

The invention relates to a method for filtering the background of a roadside radar in real time based on polygon modeling, which comprises the following steps of: carrying out geometric modeling on a road space based on a polygon, surrounding and closing the road space monitored by the radar by using a geometric model, reserving point clouds in the geometric model as detection objects, and filtering point clouds outside the geometric model as backgrounds. According to the method, the excellent geometric modeling performance of the polygon is adopted to adapt to various complex road space geometric bodies; according to the invention, various changes outside the road space do not interfere with each other; and through a table look-up method, the calculation load is low, the calculation time is short, and the background can be filtered in real time.

Disclosed is a particle damping structure vibration response estimating method based on the gas-solid two-phase flow theory. The method includes firstly acquiring the density and diameter of particle material, and measuring a coefficient of restitution to acquire the equivalent viscosity of collisions among particles and the equivalent viscosity of frictions among the particles; then utilizing the equivalent viscosity of the collisions among the particles and the equivalent viscosity of the frictions among the particles to acquire the viscosity of the gas-solid mixed flow in a particle damping structure; acquiring the cross section area of a particle damper, and utilizing the viscosity of the gas-solid mixed flow and a velocity amplitude value of a point, to be measured, of the particle damping structure to obtain an equivalent damping coefficient of the collision and friction effect among the particles in the particle damper; setting parameters in finite element software COMSOL Multiphysics, performing simulating response estimation on the particle damping structure, and optimizing the particle damping structure in a guided manner according to estimating results. The method is simple, is adaptable to analysis of complex shell or box structure problems, and can provide a solution to the problems of vibration and acoustic radiation response estimation of the particle damping structure.

The invention discloses a registration method and system for point clouds, and the method comprises the steps: carrying out edge extraction and downsampling of a first source point cloud and a first target point cloud to acquire a second source point cloud and a second target point cloud; randomly extracting a first coplanar point set from the second source point cloud, and determining a second coplanar point set in the second target point cloud; determining candidate rigid body transformation parameters according to a first coplanar point set and a second coplanar point set, converting the second source point cloud to acquire a third source point cloud; calculating the final overlapping degree of the third source point cloud and the second target point cloud; and determining a target rigid body transformation parameter from the candidate rigid body transformation parameters according to the final overlapping degree, and performing rigid body transformation of the first source point cloud and the second source point cloud. The edges of the first source point cloud and the first target point cloud are extracted and then downsampled to obtain the second source point cloud and the second target point cloud, so that the registration precision can be improved; by automatically modifying the initial overlapping rate value, the algorithm time complexity is reduced, the extraction process of the point set is accelerated, and the efficiency is high.

The invention discloses a power flow calculation method of a radiation type power distribution system; specific to the power distribution system in one direction of a power distribution network district, power flow calculation is carried out; firstly, according to the characteristics of the radiation type power distribution system, a circuit model of the radiation type power distribution system isbuilt, the topology structure of the radiation type power distribution system is reasonably numbered, and the initial data of the circuit is represented by means of a matrix form, especially data description of the circuit topological relation, so that the logic relation of the graph connection in the circuit model is clear and definite; the introduction of the incidence matrix realizes the programming design of the path searching process, and the initial data matrix is used as the only data input, so that the design of the whole program is direct and simple; and the voltage of each load nodeand the power flow of each branch of the power distribution network are solved through algebraic operation of the matrix elements in the calculation process, so that the calculation time of the program is relatively short, and the convergence speed is relatively high.

The embodiment of the invention discloses a well condition change identification method based on pumping unit electrical parameters. The well condition change identification method comprises the following steps that S1, the electrical parameters of a certain pumping unit under normal well conditions are collected; S2, the electrical parameters of the certain pumping unit under real-time well conditions are collected; S3, the rate of change of the electric parameters of the pumping unit under the real-time well conditions compared with the electric parameters of the pumping unit under the normal well conditions is calculated; S4, the sum of the rate of change of the electric parameters multiplied by the influence coefficients of the rate of change of the electric parameters is calculated, and a characteristic value of well condition changes is obtained; and S5, the characteristic value of the well condition changes is compared with an identification constant of the well condition change, and whether the well condition has obvious changes or not is determined. A calculation method is simple, software implementation is convenient, the efficiency of well condition identification is high, the calculation time is short, the advantage of identification efficiency is achieved especially for large-scale mass well condition identification, abnormal changes of well conditions can be detected in time by observing the electrical parameters, the advantages of being simple and quick are achieved, the test workload of test workers is reduced, investment is small, and equipment is not prone to being damaged.