48results about How to "Meet computing requirements" patented technology

The invention discloses an industrial CCD color imaging system based on super-resolution reconstruction and automatic registration of an FPGA and a DSP. The imaging system consists of a hardware system and a software system, wherein the hardware system comprises a video acquisition module which uses a color CCD camera and a monochrome CCD camera to perform real-time acquisition of different resolutions, a module used for analog video digitalization and system conversion, a video pre-processing and automatic and reconstruction parameter registration module based on the FPGA, an automatic registration and trous wavelet fusion reconstruction algorithm realization module taking the DSP as the core, and an Ethernet communication interface module; and the software system is mainly a fusion algorithm which realizes a path of low-resolution color video and a path of high-resolution monochrome video through automatic registration and then adopts a fusion algorithm of combining trous wavelet and IHS to reconstruct a needed high-resolution color video.

The invention discloses an electrospark wire-electrode cutting pulsed power supply and a control method thereof. The electrospark wire-electrode cutting pulsed power supply comprises an alternating-current power supply, a PFC convertor, a DC/DC convertor, a main circuit, a detecting circuit, a driving circuit, a control circuit and an upper computer, wherein a three-way staggered in-parallel Buckcircuit is adopted as the main circuit, supplies unipolar voltage to a gap and replaces current-limiting resistance with energy-storage inductance. According to the control method of the electrosparkwire-electrode cutting pulsed power supply, feedback closed-loop control is conducted by taking the sum of gap voltage and gap currents as the unique control quantity of feedback closed-loop control.Therefore, the machining efficiency and the switching frequency of a system are improved, the dynamic response capability of the system is improved, the current-resistant requirement on a switching tube is lowered, and on the condition that the discharge frequency is kept constant, the single-time discharge energy can be maintained consistent, and high-speed uniform micromachining can be achieved.

The invention discloses a task load scheduling method, which comprises the following steps of: analyzing a received task load scheduling request to obtain a target task to be scheduled; sending the target task to a hybrid heterogeneous distributed computing system; wherein the hybrid heterogeneous distributed computing system comprises a plurality of computing devices with different computing architectures, and carrying out load scheduling processing on the target task by utilizing the hybrid heterogeneous distributed computing system according to a pre-established equipment topological structure diagram. By applying the technical scheme provided by the embodiment of the invention, the calculation requirements of a multi-modal artificial intelligence algorithm model are met, efficient calculation cooperation among the calculation devices of different calculation architectures is realized, and the overall performance of the hybrid heterogeneous distributed calculation system is improved. The invention further discloses a task load scheduling device and equipment and a storage medium, which have corresponding technical effects.

The invention provides a method and a system for extracting facial features. The method comprises the following steps of: 1, acquiring and comparing grayscale values of each pixel of a facial image and surrounding pixels in an image library, and determining grayscale difference values between the pixels and the surrounding pixels in the facial image; 2, performing probability density statistics on the grayscale difference values to obtain the accumulated probability distribution of the grayscale difference values, and encoding the grayscale difference values according to the accumulated probability distribution of the grayscale difference values and a requirement for the number of feature vector encoding bits to obtain grayscale difference value codes; 3, performing computation on the grayscale difference value codes in a linear local binary mode to obtain a feature vector P1 in a horizontal direction and a feature vector P2 in a perpendicular direction; 4, performing computation on the grayscale difference value codes on a statistical local binary mode to obtain a feature vector P3 in the horizontal direction and a feature vector P4 in the perpendicular direction; and 5, combining the feature vectors P1, P2, P3 and P4 to represent the facial features of the facial image.

The invention discloses an equivalent model construction method for a fuel battery power generating system. In the fuel battery power generating system, a fuel battery outputs electric energy to a power grid via an inverter, and the fuel battery power generating system is characterized in that external characteristics of the fuel battery power generating system in a grid-connected operation environment can be described by an equivalent circuit shown as a figure 3 in the specification, and a mathemathical model of the system is represented by the following second-order dynamic differential equation in the specification, wherein IL.d and IL.q are components of axis d and axis q of the output current of the inverter; Vinv is outlet voltage of the inverter; Vgrid.d and Vgrid.q are components of axis d and axis q of the voltage at PCC (Point of Common Connection) between the fuel battery power generating system and the power grid; and R and Lf are equivalent resistance and equivalent inductance respectively. By adopting the equivalent model of the fuel battery power generating system, stable-state and transient-state characteristics of the fuel battery power generating system under a grid-connected operation condition can be simulated accurately.

The invention discloses an optimal edge computing node selection method and system in an Internet-of-vehicles environment. The optimal edge computing node selection method comprises the steps of obtaining and analyzing vehicle computing task attributes in the Internet-of-vehicles environment; obtaining and analyzing attributes of different edge computing nodes; calculating the matching degree between the task attributes and the node attributes; analyzing calculation requirements of different tasks, and assigning weights to different types of matching degrees; and selecting a node with the optimal sum of the products of the matching degree and the weight as an optimal edge calculation node to calculate the vehicle calculation task. According to the method, rapid matching of vehicle computing task attributes and node attributes in the Internet-of-vehicles environment is achieved, the optimal edge computing node of the vehicle computing task is rapidly selected according to different computing requirements and matching results, the computing efficiency is improved, the computing requirements are met, and meanwhile node computing resources are saved.

The invention discloses a face feature data obtaining system. The system comprises a comparison unit, an encoding unit, a calculation unit, a calibration unit and a combination unit, wherein the comparison unit obtains and compares grey values of pixel points and surrounding pixel points of a face image in an image library; the calculation unit is used for calculating grey difference value codes in a local binary pattern so that feature vectors in the horizontal direction and feature vectors in the vertical direction can be obtained. The face feature data obtaining system can well meet requirements of an on-line network system for the calculation speed and the recognition speed of a face recognition algorithm. Besides, grey difference value codes obtained by means of the face feature data obtaining system are more accurate, and the recognition rate is much improved though calculation complexity is not improved.

The invention discloses a protein rapid detection method based on a matched microalgae protein characteristics sequence label and a system thereof; the method includes steps of collecting a microalgae sample, and acquiring its protein sequence set, converting it to a standard protein sequence set, performing splitting and configuration treatment, and converting the protein sequence set to a key/value pair list specific to a given sequence segment; after summarizing the key/value pair list, performing an intersection operation; when the result is not empty, confirming that one data of the data set contains the characteristic protein sequence. The system mainly includes a detection data input module, a key/value pair list module, a detection identification module, a characteristic value intersection operation module, and an output module. Through indexing the protein sequence, the method and the system judge the protein homology and similarity level; by comparing with the traditional method of comparing data one by one, the algorithmic identification accuracy is and efficiency are significantly improved, the indexing efficiency is greatly promoted; the method and system have good practicability, and the algorithmic time complexity is low, the performance is high; therefore, the method and the system are significant to study proteomics.

The invention discloses a fast power distribution network loop opening feasibility judgment method and device. The device comprises an initialization module, a positioning load position module, an outlet power calculation module and a safety assessment checking module, wherein the initialization module is used for obtaining the position information needing to perform loop opening, and recording the power values of the two ends of a loop-opening feeder line at the moment; the positioning load position module is used for carrying out feeder line topology analysis, determining a feeder line segment section with measurement, performing measurement based on switches on the two sides, and estimating the size of the section load; according to the position of the loop-opening point, a feeder lineoutlet of a transformer substation needing checking is determined; the position of the load of the feeder line segment region is determined through the topology information; the outlet power calculation module is used for carrying out power flow calculation of the network after the loop-opening only comprising the additional load, and the power of the feeder line outlet of the transformer substation is saved; the power of the feeder line outlet of the transformer substation is calculated through the superposition principle with the power flow calculation; and the safety assessment checking module is used for correcting the load flow of the feeder line by utilizing the calculated outlet power, and judging the loop-opening safety. The invention provides a method and a means for achieving theimplementation of the loop-opening power regulation application.

The invention discloses a quadrangle-structured-grid dividing method based on the non-equal-dividing blocking rule. The quadrangle-structured-grid dividing method includes the steps that the equidistance dividing mode is adopted on boundaries, and finally-formed grids can be even; angles of boundary grids are adjusted, abnormal grids in all the divided grids cannot appear accordingly, and the regularity of the grids can be kept; the whole computational area is divided into different sub-areas according to boundary characteristics of the whole computational area, the evenness of the grids can be kept, and it is avoided that when the grids are directly and equally divided, as the boundaries are twisty, grid intersection appears. According to the quadrangle-structured-grid dividing method based on the non-equal-dividing blocking rule, complex areas are divided through structured irregular quadrangle grids, the advantages of structured grids and non-structured grids are considered, accurate fitting of complex boundaries is achieved, and the calculating efficiency and the calculating accuracy are improved. In future parallelization-computational-process conducting process, the structured grids have the quite-large advantages relative to the non-structured grids, and the computational efficiency of the structured grids can be further improved through the parallel computation mode.

The invention discloses a realizing method of a distributed networking relay protection setting calculation system. The realizing method mainly comprises the following steps of: (1) starting the setting calculation and the power grid boundary equivalent calculation; (2) exchanging a boundary constant value; (3) forming a local power grid distributed relay protection setting calculation data environment; (4) generating a relay protection setting calculation cooperative relationship; (5) generating a relay protection setting calculation case; (6) calculating a setting calculation case, and calculating to obtain a setting calculation constant value of a local power grid according to imported boundary setting data; and (7) outputting the setting calculation constant value of the local power grid, and stopping a distributed relay protection setting calculation flow path. In the actual calculation, the calculation results obtained by wholly importing the ex-situ power grid data into the calculation results are almost to be same as one another completely; and the whole data of the ex-situ power grid is only represented by several equivalent impedances, so that the calculation speed can meet the requirement on the setting calculation.

The invention provides a template phased array seeker digital beam forming processor, comprising a fiber interface FPGA processing board and a double-DSP processing board. The fiber interface FPGA processing board and the double-DSP processing board are externally connected with an SRIO switch. The fiber interface FPGA processing board and the double-DSP processing board are interconnected throughutilization of a VPX backboard. According to an MGT transceiver employing fiber modules and an FPGA, high reliability fiber transmission of ultrahigh bandwidth baseband preprocessed data is realized,and 28 ways of receiving and 16 ways of transmitting are realized. On the premise of sufficiently and rationally utilizing hardware resources of the FPGA and DSPs, a requirement for adaption to digital beam forming computing complexity and a requirement for real-time processing under a large array can be satisfied; through flexible programming of software, functions of a product are expanded; andmodule software implementation is realized.

The invention discloses a plain river network water volume replacement rate calculation method based on a remote sensing image, and a terminal. The method comprises the following steps: establishing a water surface area-water level fitting relation between the water surface area of a plain river network enclosure water system and the water level of the water system through the remote sensing image; calculating a first water volume from the lowest water level of the plain river network enclosure water system to the water level of each water system by a difference method, calculating a second water volume below the lowest water level by a volume generalization formula method, and adding the first water volume and the second water volume to obtain the total water volume of the enclosure water system corresponding to each water level of the water system; establishing a water level-water volume fitting relation and a water surface area-water volume fitting relation of the plain river network enclosing area water system; and then obtaining the total water volume of the current enclosed area water system at the current moment according to the fitting relationship, so as to calculate the water volume replacement rate. According to the method, the water volume replacement rate can be obtained in real time, the precision of the method meets the calculation requirement of river network water system water volume mass exchange, the surrounding area water volume scheduling condition can be mastered in time, and the method has high practicability.

According to the method for calculating the hot spot temperature of the axial double-hole copper bar rotor coil of the generator, the calculation process is modularly designed according to functions,the whole calculation process is divided into an input data module, a wind path iterative calculation module and a loss and temperature rise calculation module, and the invention has the advantages that the function modules are clear, and calculation programs are easy to use and upgrade; for temperature rise calculation of the double-hole copper bar structure, the pertinence of the invention is high, and the calculation precision is high. Meanwhile, outlet surface temperature rise calculation of radial air hole surface heat dissipation and axial air hole surface heat dissipation is considered,and the method is suitable for temperature rise design of an existing two-pole 800MW-level to 1260MW-level generator and can be used for developing generators adopting axial double-hole copper bar rotor coils in more capacity ranges; and a new rotor hot spot temperature calculation method which is higher in application degree and capable of guaranteeing the precision in product optimization and improvement can be met, so that the calculation requirement of an axial double-hole copper bar rotor generator is met.

The invention provides a vector data online calculation method and device. The method comprises the steps of constructing a vector data thematic map according to vector data; performing standardized grid slicing processing on the vector data thematic map to obtain a plurality of tiles; according to the corresponding relationship between the corresponding hierarchy of each tile and the precision, selecting an optimal tile hierarchy adaptive to the statistical precision requirement according to the preset data statistical precision requirement; in a set space region range, according to space range intersection judgment, obtaining an intersection tile data set of the required tile hierarchy; taking a single tile as a minimum calculation unit for the tile data set, obtaining tile gray scale statistical results one by one through a vector data gray scale statistical algorithm, and summarizing the gray scale statistical results of the whole tile data set; according to the gray scale statistics results of all the tiles and the pixel area of each tile under the corresponding hierarchy of the tiles, obtaining the gray scale statistics area of a set space region range; vector data calculation requirements of different precisions can be met, and the vector data calculation efficiency is improved.

The invention relates to a mechanical structure health monitoring method, in particular to a quantitative analysis method for response characteristics of the aging state of a rubber O-shaped ring sealing structure. The problem that an existing finite element method cannot achieve high-precision modeling and efficient solving of the aging state of a rubber O-shaped ring sealing structure is solved. According to the technical scheme, the method comprises the following steps: firstly, obtaining rubber material samples with different aging days, and evaluating the aging degree of the rubber material through a mechanical property test; then a finite element model of the rubber O-shaped ring sealing structure is established in a modeling-importing-pretreatment mode, different aging states of an equivalent model are achieved by setting different compression elasticity moduli of a rubber O-shaped ring, and different compression ratios of the rubber O-shaped ring are achieved by setting boundary conditions; and finally, through modal analysis and random vibration analysis, obtaining a change rule between the aging state and the response characteristic under different compression ratios.

The invention discloses a robust precoding method for a multi-beam satellite communication system based on machine learning, and the method comprises the steps: constructing a multi-beam satellite downlink channel vector model containing the positioning uncertainty of a user position; acquiring a statistical channel model about the channel autocorrelation matrix; constructing a robust precoding optimization design problem of a multi-beam satellite system and rate maximization; equivalently converting a robust precoding optimization design problem of a multi-beam satellite system and rate maximization into a power minimization problem under user signal to interference plus noise ratio guarantee and single antenna power constraint; obtaining an optimal precoding vector by combining a Lagrange function of an equivalent optimization problem and a KKT condition of the Lagrange function; and predicting a Lagrangian multiplier required by the optimization problem based on a channel autocorrelation matrix by combining with a machine learning method. According to the method, the implementation complexity of a channel autocorrelation matrix prediction problem algorithm can be reduced, and the transmission performance of a multi-beam satellite communication system and the robustness of a positioning angle estimation error are remarkably improved.

The invention discloses a data processing method and device, equipment and a medium. The method comprises the steps of obtaining an aggregation result corresponding to each server in at least two servers, wherein the aggregation result corresponding to the server is determined after the server aggregates multiple pieces of target data of the server within a preset time period based on a preset aggregation condition and multiple numerical intervals; based on a preset merging condition, merging the aggregation results of the at least two servers to obtain at least one merging result; and performing quantile statistics on each combination result in the at least one combination result to obtain a quantile corresponding to each combination result. According to the embodiment of the invention, the calculation efficiency of quantiles in a big data distributed application scene can be improved.

The invention discloses a plastic warm air blower pressure chamber plate assembly which comprises a front mounting plate, a front cover connecting seat, a gutter channel and a rear mounting plate which are sequentially arranged from front to back, and the front mounting plate, the front cover connecting seat, the gutter channel and the rear mounting plate are all made of plastic; the gutter channel comprises a U-shaped channel main body and two water baffles, the U-shaped channel main body is sunken downwards in the front mounting plate and the front cover connecting seat, the two water baffles are arranged at the left end and the right end of the U-shaped channel main body respectively, and the left end and the right end of the U-shaped channel main body can be blocked by the two water baffles; the U-shaped groove body is provided with a water flowing hole, an air conditioner outer circulation butt joint opening and a windscreen wiper installation receding seat. The number of parts can be reduced, the assembling efficiency can be improved, the corrosion resistance is high, the overall weight can be greatly reduced, the raw material cost is reduced, automobile light weight is facilitated, and the cruising ability of an automobile is improved.

The invention discloses a wind turbine generator operation control method based on a prony algorithm, and belongs to the technical field of wind power generation. The method comprises the following steps that firstly, detected operation parameter signals are filtered to obtain filtered operation parameter signals; the filtered operation parameter signals are approximately fitted into a plurality of models with exponential functions of the generator rotating speed oscillation amplitude, the generator rotating speed oscillation frequency, the generator rotating speed oscillation phase and the generator rotating speed oscillation attenuation factor by using the prony algorithm; mathematical transformation is performed on the model to obtain a linear difference equation; after the linear difference equation is solved, the generator rotating speed oscillation frequency is obtained through calculation; and whether the calculated generator rotating speed oscillation frequency is within a preset value range or not is judged, and the operation state of the wind turbine generator is adjusted according to the judgment result. According to the method, mechanical and electrical parts of the wind turbine generator can be prevented from being in a periodic oscillation state to the maximum extent under the specific wind condition, and the situation that the wind turbine generator is damaged due to oscillation caused by instability is avoided.