203results about How to "Reduce the impact of accuracy" patented technology

The invention discloses a self-adaptation three-dimensional attitude positioning method based on microinertia and a geomagnetic technology, comprising the following steps of: (1) inducing a motion attitude of a carrier by utilizing a sensor of a microinertia measuring device; (2) setting an initial attitude of the microinertia measuring device and accelerated speed and geomagnetic field information under a global coordinate system; (3) solving the attitude value of the microinertia measuring device; (4) predicting the attitude value of the microinertia measuring device by utilizing triaxial micro-gyroscope sensor data; (5) carrying out confidence judgment on the triaxial micro-acceleration sensor data and triaxial magnetic field sensor data, detecting the interference of a surrounding environment, and setting self-adaptation parameters; (6) obtaining the attitude error value of the microinertia measuring device by utilizing the triaxial micro-acceleration sensor data and the triaxial magnetic field sensor data which are processed in the step (5); (7) fusing the attitude predictive value obtained from the step (4), corrected information obtained from the step (6) and the self-adaptation parameters obtained from the step (5) to obtain the attitude value of the microinertia measuring device; and (8) outputting attitude information.

The invention discloses a network monitoring method and a network monitoring system to solve the problems that the processes for performing analysis and fault elimination in the current network optimization and fault treatment flows have low real-time, poor accuracy and low efficiency. The network monitoring method provided by the invention comprises the following steps that: a network side receives session information reported by each terminal when sessions performed by the terminals respectively have abnormal events; according to the session information reported by a first terminal in the terminals, the network side acquires network side signaling information, which is received and transmitted during the time when the session information reported by the first terminal happens and belongs to corresponding sessions of the terminal, from acquired signaling information received and transmitted by a network side network element; and the network side performs fault analysis according to the session information reported by the first terminal and the acquired network side signaling information.

The invention relates to a device for testing nozzle characteristics. The device comprises a testing device box, a nozzle, a testing platform system, a gas channel system, a water channel system and a laser imaging system. The device is used for testing aerial fog characteristics such as the drop grain diameters, the drop speed, the flow density at the different positions and drop grain diameter distribution of the different nozzles at different spraying heights at different air pressures and different water pressures. The device is high in testing accuracy, wide in adjusting scope, stable in operation and capable of testing the aerial fog characteristics of single-phase liquid flow and gas liquid two-phase flow.

The invention discloses a robust learning model and an image classification system. The robust learning model comprises the steps that a training set is initialized so that an initial category tag matrix is obtained, and training samples in the training set include the samples of which the categories are known with calibration of category tags corresponding to the categories and the samples of which the categories are unknown without calibration of the category tags; the training samples are processed by a construction method based on neighboring definition and reconstruction weights, a reconstruction coefficient matrix is constructed according to similarity between the samples, and symmetrization and normalization processing is performed; soft tags without calibration samples are determined by utilizing the reconstruction coefficient matrix and the initial category tag matrix, and l2,1 normal regularization is performed on the soft tags of the training samples by adopting an iteration method so that a projection matrix and a soft tag matrix are obtained; mapping is performed on samples under test by utilizing the projection matrix so that the soft tags of the samples are obtained; and the samples under test are the samples of which the categories are unknown without calibration of the categories. Influence of mixed signals in an original space can be effectively reduced by the model so that classification accuracy can be enhanced.

The invention discloses a Lamb wave damage probability imaging method for damage monitoring of a composite plate structure, which belongs to the field of monitoring of engineering structure health of composite plates. According to the invention, the technology of excitation / sensing arrays is utilized; reference signals of the composite plate structure in a sound state and monitoring signals of the composite plate structure in a damage state are acquired by loading Lamb wave; then the means of wavelet analysis is employed to extract an energy difference coefficient of damage, and the position where damage is located is predicated; finally, a structure damage image is obtained by using the probability imaging method. The invention enables rapid and effective damage identification of the composite plate structure to be realized and high precision in damage identification to be obtained, and has a good engineering application value.

The invention provides a point cloud based indoor dynamic scene SLAM (Simultaneous Location and Mapping) method, which comprises the steps of carrying out multi-scale based laser frame and probability graph registration; updating easy layer of historical probability graph by combining information of the current frame, wherein the implementation mode of the step comprises the sub-steps of converting the current frame to a coordinate system where the historical probability graph is located according to a registration result so as to acquire a converted laser frame, carrying out probability graph generation so as to generate a corresponding probability graph, and updating the historical probability graph by combining the current frame probability graph and the historical probability graph so as to acquire a current historical probability graph; terminating iteration until all of the laser frames are processed, and acquiring a moving trajectory of a laser and a two-dimensional map of the indoor dynamic environment. The technical scheme provided by the invention can carry out composition on a dynamic scene quickly and accurately and accomplish self-locating.

The invention discloses a method for manufacturing cement-base conducting composite materials, which belongs to the technical field of civil engineering. The method is characterized in that a plane stainless steel net is manufactured into a solid stainless steel net of different waveforms. When a cement-base conducting composite material specimen is manufactured, the wave crests of the electrodes of two solid stainless steel nets are arranged oppositely; and as the basal body of the cement-base conducting composite material can be made of cement paste, cement mortar or concrete, and the conductive phases incorporated can be carbon fiber, powdered carbon, steel fiber, steel scrap, plumbago or carbon nanometer tubes, the grid sizes of the plane stainless steel net, the waveforms, the cycles and the amplitudes of the waveforms of the electrodes of the solid stainless steel nets can be selected on a rational basis according to the types of the material of the base body and the conducting material besides considering the type and the size of the specimen in practical application. The invention has the advantages that the contact resistance between the electrodes and the base body can be reduced; the influence of polarization effect is reduced; the bonding effect of the electrodes and aggregate is improved; and the accuracy of the resistance test of the cement-base conducting composite material is improved.

The invention relates to a full-automatic chemiluminescence immunoassay analyzer, which comprises a reaction cup loading module, a transmission track module, a sample filling module, a reagent filling module, a reaction disc module, a magnetic separation disc module and a detection module, wherein after the sample filling module rotates a sample needle to a sampling position of the transmission track module, the sample needle is inserted into a liquid level for sampling, and after the sampling is completed, the sample needle is lifted and rotated to a sample adding position of the reaction disc module for sample adding; and the reagent filling module is used for quantitatively filling a reagent into a reaction cup of the reaction disc module for uniform mixing reaction, after the reaction is finished, the reaction disc module is used for transferring the reaction cup into the magnetic separation module, and components participating in the reaction are adsorbed on the wall of the reaction cup. Compared with the prior art, the full-automatic chemiluminescence immunoassay analyzer has the advantages that the structure is simple, the operation is convenient, the size of the whole machine can be reduced, the occupied space is small, the production cost is reduced, and the like.

The invention discloses an analog measuring device and method for the surface potential of an insulator in a GIL. The analog measuring device comprises a closed cavity, the insulator, an electrostaticprobe and a high-voltage electrode guide rod. Air inlet and outlet are formed in the cavity wall surface of the closed cavity. A guide rail is installed in the closed cavity, and a first shielding cylinder and the insulator are arranged at one end of the guide rail. The insulator is rotatably disposed in the first shielding cylinder through a high-voltage electrode support rod. A second shieldingcylinder is arranged at the other end of the guide rail, and the second shielding cylinder is internally provided with the high-voltage electrode guide rod. The first shielding cylinder and the second shielding cylinder are both grounded. The electrostatic probe is movably disposed in the closed cavity, and the electrostatic probe can measure the surface potential of the insulator. The analog measuring device can complete the analog measurement of the surface potential of the insulator in the high-pressure closed cavity, and can simulate the working environment of the insulator of the GIL inreality, the measurement result is high in credibility, and theoretical guidance can be provided for actual production.

The invention discloses an operational-amplifier-free band-gap reference circuit with a pre-voltage-stabilizing structure, and solves the problems in the prior art that the power supply rejection ratio is low and the offset voltage is high. The operational-amplifier-free band-gap reference circuit comprises a band-gap core unit (1), an output unit (2), a voltage pre-stabilizing unit (3) and a starting unit (4). The band-gap core unit generates a negative temperature coefficient voltage VBE and a positive temperature coefficient voltage delta VBE; the output unit generates band-gap voltage VBGand zero temperature coefficient voltage VO; the pre-voltage-stabilizing unit generates stable pre-adjusting voltage VREG by detecting feedback voltage VFB and supplies power to the band-gap core unit. The starting unit detects the bias voltage VBIAS outputted by the pre-voltage-stabilizing unit and outputs a starting signal Vstart to the pre-voltage-stabilizing unit. The pre-voltage-stabilizing structure is adopted so that the power supply rejection ratio of reference voltage is increased; the operational amplifier structure is not adopted, and the influence of offset voltage on reference voltage is weakened. The circuit is simple in structure and can be used for an analog integrated circuit.

The invention discloses a band-gap reference circuit free from an operational amplifier. The band-gap reference circuit comprises a band gap core unit, a clamp unit and a starting unit; the band gap core unit is used for generating zero temperature coefficient reference voltage VREF; the input end of the clamp unit is connected with clamp voltage VA and clamp voltage VB, and an output end is connected with the reference voltage VREF and formed to be a negative feedback along with the band gap core unit; the clamp unit is used for guaranteeing that the clamp voltage VA and clamp voltage VB output by the band gap core unit are equal to each other; an input end of the starting unit is connected with the clamp voltage VB, and an output end outputs a starting signal Vstart to the clamp unit forguaranteeing that the clamp unit and the band gap core unit can rapidly enter a normal working state in the process of electrifying. For not adopting the operational amplifier structure, the influence of offset voltage on the reference voltage VREF is weakened, and the temperature coefficient of the reference voltage is reduced by a resistance proportional compensation technology. The band-gap reference circuit is featured by low temperature offset, low power consumption, and simple structure; the band-gap reference circuit can be applied to a chip with high requirement on static power consumption.

The invention discloses a time-phased refined short-term load prediction method based on BOA-SVR and fuzzy clustering. The method comprises the following steps: S1, analyzing factors affecting the load change according to the historical load data and characteristics of a region, and proposing a time-phased comprehensive prediction method through considering the real-time influence factors; s2, clustering the historical data by using a fuzzy C-means clustering method to obtain different types, and evaluating the clustering effect of the historical data; s3, improving a bat algorithm by using adynamic adaptive weight method and a Cauchy distribution inverse accumulation partial function; s4, optimizing the parameters of the SVR by using an improved bat algorithm, searching the optimal parameters to establish an SVR time-phased refinement model, and taking the average load value of each time period to be predicted as the output of the SVR to obtain a prediction result; according to the method, the prediction precision can be effectively improved, the prediction precision for different types of days can reach 96% or above, the scale of the prediction model is effectively reduced through the improved bat algorithm, and the performance of the prediction model is obviously improved.

The invention belongs to the field of low-temperature plasma parameter diagnosis equipment, relates to a probe system and particularly relates to a probe system for low-temperature plasma potential diagnosis. The probe system comprises a metal wire, an insulating sleeve, two metal posts, wires, two sampling resistors, an oscilloscope and a direct-current stabilized power source. The insulating sleeve is provided with two passages for fixed mounting of the metal columns; the two metal columns are conducted through the metal wire; the other end of each metal column is connected to the direct-current stabilized power source and the two sampling resistors through the wires; the two sampling resistors are serially connected and then parallelly connected with the direct-current stabilized powersource; a measurement probe of the oscilloscope is mounted at the center of the two sampling resistors in series connection. By symmetry of a parallel circuit and plasma sheath in a contact process oflow-temperature plasmas with metal equipment, real-time, quick and accurate measurement of low-temperature plasma potential is realized.

The invention discloses a behavior identification system based on sub-action perception and a work method thereof. The system comprises a wearable terminal used for acquiring inertia sensor data, if the wearable terminal has sufficient local processing capability, a behavior identification algorithm based on sub-action perception operates, action of a user is identified, if the wearable terminal has insufficient local processing capability, data is forwarded to a cloud processing server for processing; communication equipment used for processing information transmission between the wearable terminal and the cloud processing server, and the cloud processing server used for acquiring the inertia sensor data acquired by the wearable terminal in real time, the behavior identification algorithm based on sub-action perception operates, and user action is identified. The method is advantaged in that accuracy generally does not decrease when action of a new user is identified, problems caused by customized user action can be excellently solved, required training data is small, training cost is small, an identification speed is fast, calculation is simple, and cost is low.

The invention discloses an image classification model processing method. The image classification model processing method comprises the following steps: mapping source domain image samples and targetdomain image samples in an original space to the same characteristic space, and correspondingly obtaining the mapped source domain image samples and the mapped target domain image samples; performingdimensionality reduction on the mapped source domain image samples and the mapped target domain image samples, and correspondingly obtaining a source domain subspace and a target domain subspace; aligning samples in the source domain subspace and the target domain subspace, and determining reduced-dimension source domain image samples when a distribution difference between a source domain and a target domain in the characteristic space satisfies a preset minimum difference condition; performing weighting processing on each type of samples in the reduced-dimension source domain image samples, and obtaining aligned and weighted source domain image samples; applying the aligned and weighted source domain image samples and corresponding label information to a model for classifying new image samples in the target domain. The invention also discloses an image classification model processing device and a storage medium.

The invention provides a spatially correlated cylindrical shape error monitoring method and system and a medium, and the method comprises a surface measurement step of measuring the surface of a cylinder, and obtaining the three-dimensional coordinate data of a measurement point; a data processing step of obtaining an error value of the measurement point according to the obtained three-dimensional coordinate data of the measurement point; and a spatial correlation judgment step of judging whether the error value of the measurement point has the spatial correlation or not according to the obtained error value of the measurement point, and obtaining a spatial correlation judgment result. The monitoring method and system of the cylindrical geometric error with spatial correlation provided by the invention can reflect the form of the processing error more accurately, remove the influence of the spatial correlation on the measurement result and effectively improve the recognition rate of unqualified products, and improves the defect that the traditional cylinder quality monitoring cannot reflect the geometric shape of the cylinder and the defect of the spatial correlation caused by a large number of measuring points.

The invention aims at providing a technology of monitoring a state of a CDN node. The technology is characterized by acquiring state detection information corresponding to the CDN node from a slave equipment terminal, and according to the acquired state detection information and node state information corresponding to the CDN node, determining detection state change information of the CDN node; according to acquired one or more detection state change information sent by slave equipment, determining integration state change information of the CDN node by a master equipment terminal; and when the integration state change information changes, updating the node state information. Compared to the prior art, by using the technology of the invention, an influence of single point detection misjudgment on a CDN node state is effectively reduced and monitoring accuracy of the CDN node state is increased. And the slave equipment can be deployed in a or network or a network segment where the monitored CDN node is located so that an influence of network time delay or state jitter on single point detection accuracy is effectively reduced.

The invention discloses an HgCdTe infrared photoconductive detector with a reference element structure, wherein the reference element comprises a substrate, an epoxy glue layer, a first-side HgCdTe oxidation layer, a bulk material HgCdTe layer, a second-side HgCdTe oxidation layer, a silica insulating layer, a negative photoresist insulating layer, a metal indium layer, and a metal gold layer; and the reference element is characterized in that an edge photosensitive surface is covered with a double-layer insulating layer, an infrared absorption layer and an infrared light reflection layer. The HgCdTe infrared photoconductive detector with a reference element structure has the advantages: the problem of temperature drift existing in the HgCdTe infrared photoconductive detector is solved, and because a differential circuit is adopted to replace a temperature-compensation circuit, the influence of ambient temperature variation on the accuracy of the HgCdTe detector can be reduced, and the circuit structure can greatly be simplified; the detector can be used for measuring the temperatures of low-speed moving objects and high-speed moving objects; and the detector is quick in response speed and has a self-compensating function.

The invention discloses a magnetic core surface defect detection method based on deep learning, which comprises the following steps of: firstly, acquiring all front and back images of a magnetic core, preprocessing, extracting defect parts, inputting a deep convolutional generative adversarial network improved by a Gaussian mixture model, training an extracted defect set to generate a new defect image, performing poisson fusion on the obtained new image and the intact magnetic core image, making a standard data set, and dividing a training set and a verification set; inputting the training set image into a YOLO-v3 neural network for training, proposing a new training strategy, setting training parameters, and taking the trained network as a magnetic core surface defect detection model so as to identify category information and position information of the magnetic core. The method provided by the invention is faster and more accurate in identification.

The invention relates to an entity recognition method, terminal equipment and a storage medium. The method comprises the steps of S1, constructing a word graph containing domain entities correspondingto a to-be-recognized text; S2, expressing each word in the to-be-recognized text as a vocabulary tensor through a word vector embedding layer; S3, extracting candidate entities corresponding to theto-be-recognized text from the constructed word graph through a graph neural network module according to all vocabulary tensors of the to-be-recognized text, wherein the graph neural network module comprises a graph attention network layer and a bidirectional graph convolutional network layer; S4, converting a vocabulary tensor and a candidate entity of the text to be recognized into an intermediate calculation tensor containing context information through a bidirectional recurrent neural network layer; and S5, inputting the intermediate calculation tensor into a CRF decoding layer for decoding to obtain an entity contained in the finally recognized text to be recognized. According to the method, the secondary graph structure of the entity boundary is modeled, and the relationship betweenthe entity boundary and the graph neural network is analyzed, so that influence of insufficient judgment of the entity boundary on the result accuracy is reduced.

The invention relates to a novel single-phase earth fault starting method. The method comprises the steps of carrying out wavelet packet transformation on a zero sequence voltage signal of a bus to obtain a specific frequency band component; further processing the history data of a high-frequency component and a low-frequency component to obtain a self-adaptive threshold value for diagnosing whether a system is disturbed or not; and rapidly detecting the single-phase earth faults according to whether the system is disturbed or not and the disturbance continuing time. The method is capable of rapidly and correctly recognizing the single-phase earth faults and effectively distinguishing the single-phase earth faults and non-fault behaviors of switching on-load circuits and switching capacitor banks, has relatively strong adaptation capacity and has relatively high recognition correctness under the working condition of noise interference.

The present invention discloses air pollutant detection equipment, and belongs to the technical field of environments. The air pollutant detection equipment comprises an environment parameter adjustment device, a sensor device, an index calculation device and a positioning device, wherein the environment parameter adjustment device is used for collecting external gas through a gas inlet and adjusting the environment parameter value of the collected gas to a preset value range, the sensor device is arranged at the gas outlet position of the environment parameter adjustment device, and is used for carrying out pollutant detection on the parameter-adjusted gas, and the index calculation device is respectively connected to the sensor device and the positioning device, and is used for determining the concentration value of the preset pollutant in the parameter-adjusted gas according to the detection data detected by the sensor device, acquiring the position information detected by the positioning device, and performing corresponding storage on the concentration value, the position information and the detection time. With the air pollutant detection equipment of the present invention, the accuracy of the detection on the pollutant concentration can be improved.

The invention provides an octane number loss prediction method based on a particle swarm algorithm and a neural network. The method comprises the steps: firstly, collecting the original data and octane number loss value data of an operation variable in a catalytic cracking gasoline refining desulfurization device, and carrying out the preprocessing of the original data of the operation variable; secondly, performing feature screening on the processed data by adopting decision tree regression and Pearson correlation coefficients to obtain feature variables; then training the four-layer BP neural network by utilizing sample data and octane number loss value data corresponding to the characteristic variables to obtain an octane number loss prediction model; and finally, performing iterative optimization on the sample data corresponding to the characteristic variables by using a particle swarm algorithm and an octane number loss prediction model, and outputting the value of the characteristic variable corresponding to the minimum octane number loss value. According to the method, the particle swarm algorithm and the BP neural network are combined to search for the value of the characteristic variable corresponding to the minimum octane number loss value, so that the repeated training process is avoided, and the prediction efficiency is improved.

The invention discloses an underwater target signal feature extraction method based on probability latent semantic analysis, and the method comprises the following steps: processing an underwater target signal mixed with noise to obtain a noise reduction spectrogram, and carrying out the spectrum reliability weight calculation through an original spectrogram v and the noise reduction spectrogram;obtaining a noise reduction spectrogram Sn of the nth frequency band after weighting; carrying out short-time Fourier transform on Sn to obtain a signal pair, filtering the signal pair by using a frequency spectrum triangular filter, carrying out matrix decomposition on the whole frequency spectrum pair after filtering, and extracting first L feature vectors with the highest contribution rate to form an acoustics subspace; performing normalization calculation on the extracted first L feature vectors, and connecting the normalized L feature vectors to construct an acoustic feature vector. According to the method disclosed by the invention, the characteristics of the sound signal are reflected by combining time and frequency, and the noise pollution can be reduced, so that the sound signal can be more effectively represented.

The invention discloses a dual-learning-based method for counting pedestrians at a subway station scene. The method is characterized by comprising the steps of (1) based on the subway video foreground extraction method of a Gaussian hybrid model, extracting the active computing area of a video pedestrian group for each frame of the video; (2) based on the dual-area division method for the subway monitoring video, dividing each frame of the video into a far-view area and a near-view area; (3) based on the far-view pedestrian estimation method of a super-strong learning machine (ELM), extracting the global scene characteristics of the far-view area, and establishing a pedestrian number estimation model based on the ELM and a Gaussian process regression model; (4) conducting the pedestrian number estimating method for pedestrians in the near-view area based on aggregated channel characteristics (ACF). According to the technical scheme of the invention, the video data acquired by each single monitoring camera in a subway station is adopted as the input, and the number of pedestrians obtained through counting pedestrians in each frame of a video data stream is adopted as the output. The counting result of the method is good in accuracy and real-time performance.

The invention relates to the field of damage detection, in particular to a novel structural damage positioning method based on acoustic emission. The method comprises the following steps: S1, collecting acoustic emission signals emitted when a to-be-detected structure is damaged through each acoustic emission sensor; S2, calculating the time difference of the acoustic emission signals arriving at any two acoustic emission sensors; S3, inputting all the time differences in the step S2 into a trained neural network model, and obtaining the damage position of the to-be-detected structure, wherein the damage position is the relative position between the damage position of the to-be-detected structure and one acoustic emission sensor S. The detection method provided by the invention can reduce the influence of a frequency dispersion phenomenon, a propagation path and the environmental noise on the accuracy of the detected damage position in the propagation process of the acoustic emission signals.

A comparator type power supply controller that controls an output voltage. The power supply controller includes a comparator that compares a feedback voltage, which corresponds to the output voltage, with a reference voltage, wherein the comparator instructs power supply when the feedback voltage decreases when the feedback voltage decreases from the reference voltage. An offset application circuit is coupled to the comparator. The offset application circuit controls an input offset voltage of the comparator and decreases the input offset voltage as time elapses when the feedback voltage becomes lower than the reference voltage after the power supply ends. An amplifier amplifies a voltage difference of the feedback voltage and the reference voltage. The offset application circuit increases a change speed of the input offset voltage as the voltage difference increases in accordance with an output voltage of the amplifier.

The invention discloses an air temperature and humidity automatic control system and method for a cloud chamber. The method is characterized in that air to enter the cloud chamber is divided into two parts; one part of the air is dried to form dry air; the other part of the air is humidified to form moist air; the temperature and the humidity of the dry air and the temperature and the humidity of the moist air are monitored; and according to required and preset air temperature and humidity, the dry air and the moist air are proportionally mixed and then are pumped into the cloud chamber. The system and the method may provide air with specific humidity for cloud physics and manual influence weather research atmosphere ice nucleus or ice crystal equipment.

The invention discloses a baseline processing method and device of a pulse signal and a particle detection system. The baseline processing method of the pulse signal comprises the following steps of:sequentially acquiring data of each sampling point in a to-be-processed pulse signal; acquiring an average value of the data of a current sampling point and a continuous preset number of sampling points before the current sampling point, and taking the average value as a baseline value corresponding to the current sampling point; and subtracting the corresponding baseline value from the data of each sampling point to obtain baseline-removed data of each sampling point, thereby obtaining a baseline-removed target pulse signal. Through the mode, the influence of baseline fluctuation on the accuracy of pulse recognition can be reduced, so that the accuracy of pulse signal recognition is improved.