55results about How to "Contributes to accuracy" patented technology

A gas measurement system of this invention includes a detector assembly having a beamsplitter adapted to separate infrared radiation into a first beam and a second beam and a mirror adapted to receive the first beam from the beamsplitter. The first and second beams are directed to first and second detectors that are disposed in a common plane. In one embodiment, the optical elements are provided on or in an optical block. In another embodiment, the gas measurement system includes a housing that contains an infrared absorption gas measurement assembly, a luminescence quenching gas measurement assembly, and a processor that is programmed to measure gas constituents of a gas flow based on an output of the infrared absorption gas measurement assembly and the luminescence quenching gas measurement assembly.

The invention relates to an electric car driving motor torque closed-loop control method. The electric car driving motor torque closed-loop control method is composed of a whole vehicle control unit,a motor controller, an inversion switching circuit, a permanent magnet synchronous motor, a position sensor and a three-phase current sensor. The electric car driving motor torque closed-loop controlmethod is characterized in that whether active short circuit control is carried out according to an active short circuit control instruction (ASC-Enable) and the motor mechanical rotating speed, the output torque of the permanent magnet synchronous motor is calculated by means of an active short circuit method without depending on motor parameters such as stator resistance, rectangular axis voltage and the like, and the calculated value is more accurate; the motor controller then carries out PI regulation control according to a torque command emitted by the whole vehicle control unit and the torque on-line calculated value, the accuracy of the torque output absolute value can be improved, and the torque closed-loop response time is shortened.

The invention relates to functional masterbatch of a halogen-free intumescent flame retardant, and a preparation method thereof. The halogen-free intumescent flame retardant in the functional masterbatch is a phosphorus-nitrogen series intumescent flame retardant, and a matrix in the functional masterbatch is a polyolefin resin. The functional masterbatch comprises, by weight, 50-90 parts of the halogen-free intumescent flame retardant, 20-45 parts of the polyolefin resin, 1-5 parts of a compatibilizer; 1-5 parts of a lubricant and 1-5 parts of a surface treatment agent. According to the present invention, dust pollution due to the flame retardant is fundamentally overcome; the flame-retardant masterbatch and polypropylene have good dispersions, such that purposes of environmental protection and no pollution are achieved; heavy metal contents of the functional masterbatch accord with requirements of the report of societe generale de surveillance S.A. (SGS); and the functional masterbatch has a good flame retardant property and high economy.

The invention discloses a method for predicting permeability of carbonate acid fracturing self-supporting cracks. The method comprises the following steps: collecting basic parameters of carbonate acid fracturing self-supporting cracks; collecting a rock sample to simulate a reaction process of the acid rock, and acquiring a Young modulus and Poisson modulus and a poisson ratio of carbonate rock dissolved by acid; establishing an equation of crack width variation under acid fracturing self-supporting crack stress sensitivity; establishing a gas effective viscosity prediction model consideringthe stress sensitive effect and confined space effect of carbonate acid fracturing self-supporting cracks; and establishing a dynamic prediction model for the permeability of the carbonate acid fracturing self-supporting cracks considering the stress sensitive effect and the influence of the viscosity variation in a micro-nano scale confined space of the self-supporting crack for the gas, acquiring a prediction result of the permeability of the carbonate acid fracturing self-supporting cracks. The prediction model is closer to the actual formation conditions, thereby facilitating accuracy forpredicting the permeability of the acid fracturing self-supporting cracks.

The invention discloses an ultrasonic signal processing method based on an auto-encoder. The method includes preprocessing an ultrasound signal, and then superposing samples of the ultrasonic signals,establishing and classifying a sample library of the ultrasonic signals, compressing and decompressing the ultrasonic A scanning signals through an automatic encoder established through a neural network, obtaining related weights and bias variables of the feature functions based on deep learning, and performing noise reduction processing on subsequent ultrasonic signals by taking the related weights and bias variables as noise reduction parameters. The method is good in signal noise reduction screening compatibility, small in operand and low in calculation cost, and the feature extraction accuracy of the noise reduction auto-encoder is improved.

The invention relates to an intelligent diagnosis diagnosis method and device based on an ensemble learning framework, equipment and a medium, and the method comprises the steps of collecting multi-source operation data of a target metrological verification automatic system, and carrying out the preprocessing of the multi-source operation data to obtain a corresponding fault data sample set; constructing a readable target data set of a preset fault diagnosis model according to the fault data sample set, wherein the preset fault diagnosis model adopts an XGBoost model, training the XGBoost model according to the target data set, and taking the corresponding XGBoost model after training as a target fault diagnosis model; and acquiring online operation data of the target metrological verification automatic system in real time, and analyzing the online operation data according to the target fault diagnosis model to predict the fault condition of the target metrological verification automatic system. According to the invention, the fault condition of the metrological verification automatic system can be accurately predicted on line in real time, and the fault diagnosis efficiency and accuracy are improved.

The embodiment of the invention provides a three-dimensional reconstruction method and a three-dimensional reconstruction system for mixed pictures. The method comprises the steps of classifying all pictures in a mixed picture set through a trained neural network model to acquire a picture set corresponding to each target in a to-be-modeled scene, wherein the mixed picture set includes pictures atdifferent angles of the to-be-modeled scene and characteristic points of multiple target scenes; and performing three-dimensional reconstruction on each target according to the picture set corresponding to each target to obtain a three-dimensional reconstruction image of the to-be-modeled scene. According to the embodiment, because multiple targets are generally included in the mixed pictures, the mixed pictures need to be classified at present to obtain the picture set corresponding to each target, the three-dimensional reconstruction is performed on the target according to the picture set corresponding to each target so as to obtain the three-dimensional reconstruction image of the to-be-modeled scene; and the embodiment is beneficial to the accuracy and extensibility for three-dimensional reconstruction of the target.

The objective of the invention is to provide a method for determining the content of fluorine in mold flux for continuous casting by using a fluoride ion selective electrode. The method comprises thefollowing concrete steps: (1) drawing a standard fluorine curve: with cyclohexanediaminetetraacetic acid-dipotassium phosphate-potassium dihydrogen phosphate as a total ion concentration buffer solution (with a pH value of 5.8), separately determining the potential values of the fluorine ion selective electrode in solutions ranked according to a sequence from low concentration to high concentration at a temperature of 30 DEG C with a saturated calomel electrode as a reference electrode, and drawing a figure E(mv)-lg(cF<->) (mg / l) of the potential values and the logarithmic values of fluorinions so as to obtain the fluorine standard curve; (2) in a muffle furnace, subjecting a sample of the mold flux for continuous casting to decomposition in a corundum crucible under the action of a mixedflux of Na2O2 and ZnO; (3) subjecting slag in the crucible to leaching with hot water, filtering and washing, and carrying out fixing to a constant volume; and (4) taking a sample solution, adding thetotal ion concentration buffer solution, carrying out fixing to a constant volume, determining the potential of the electrode in the formed test solution, finding out a corresponding fluorine concentration in the standard curve, and calculating the content of fluorine in the sample of the mold flux for continuous casting.

The invention discloses a festival and holiday scenic spot passenger flow prediction method based on machine learning, and the method comprises the steps: taking the festival and holiday scenic spot passenger flow as a target variable, obtaining a feature variable, carrying out the data preprocessing, and storing the data in a database as sample data; randomly dividing the sample data into a training set and a test set; selecting a support vector regression algorithm as a training model, and training the training model by using the training set to obtain a festival and holiday scenic spot passenger flow prediction model; and performing model prediction by using the test set, evaluating the generalization ability of the model, and evaluating the performance of the model. According to the method, the artificial intelligence algorithm and tourist travel business characteristics in the tourism industry are effectively combined, the influence of various factors and real-time network data on the number of people in the scenic spot in a specific time period is counted by integrating previous data, and the prediction result is corrected by incorporating emergency factors, so that the scenic spot passenger flow condition in festivals and holidays is more accurately predicted.

The invention relates to the field of medical examination, and particularly relates to a sample dilution composition for detecting a sugar antigen 72-4, a detection reagent and a kit containing the detection reagent. The kit mainly comprises a magnetic particle suspension coated with a CA72-4 antibody, an enzyme conjugate, a sample diluent and a calibrator. The sample dilution composition can be used for auxiliary diagnosis and postoperative monitoring of tumor diseases such as gastric cancer, ovarian cancer, colorectal cancer, pancreatic cancer and cervical cancer. The kit is short in detection time, and the time of an inspector is greatly saved; the kit adopts a double-antibody sandwich method, is wide in linear range and high in clinical practicability; the kit adopts the specific sample diluent, so that the difference between new and old samples can be eliminated, false positive interference caused by HAMA is effectively reduced, and the detection accuracy of CA72-4 is greatly improved; and the kit adopts a two-step method, so that the risk of HOOK in clinical practice is reduced.

The embodiment of the invention discloses a dialogue robot reply method combining RPA and AI, a retrieval model training method and device used for a dialogue robot combining RPA and AI, equipment anda medium. The dialogue robot reply method comprises the steps of receiving a questioning statement of a user; based on a preset retrieval model, determining a target similar problem meeting a similarity requirement with the questioning statement; and returning the answers of the knowledge points to which the target similar questions belong to the user. The method is advantaged in that the presetretrieval model is obtained by training similar questions in a preset knowledge base so as to establish an association relationship between the similar questions and the similarity. By adopting the method, question answering accuracy of the dialogue robot can be improved.

The invention relates to the technical field of energy monitoring, and in particular relates to an office building independent room energy consumption collecting, monitoring, analyzing and alarming system. The office building independent room energy consumption collecting, monitoring, analyzing and alarming method comprises the steps that an office layout diagram and a corresponding office function distribution area are obtained, and the office layout diagram is split into corresponding energy consumption monitoring areas according to the office function distribution area; energy consumption use condition data of each energy consumption monitoring area are collected, and the energy consumption use condition data are monitored to obtain a corresponding monitoring result; the monitoring result and the corresponding energy consumption reference value are monitored and analyzed to obtain a corresponding analysis result, and the analysis result comprises standard exceeding energy consumption and normal energy consumption; and if the analysis result shows that the energy consumption exceeds the standard, a corresponding alarm message is triggered. The system provided by the invention hasthe effects of monitoring the energy consumption in an office and giving an alarm.

The embodiment of the invention provides a risk assessment method, a risk control system and risk assessment equipment. The risk assessment method comprises the following steps: obtaining a first riskpoint constructed for an assessment object and an index item associated with the first risk point; determining an evaluation parameter of the first risk point according to the change information of the index item under the condition that the index item is changed; and calculating a risk quantitative value corresponding to the first risk point according to the evaluation parameter. According to the technical scheme provided by the embodiment of the invention, dynamic risk assessment in a service full life cycle is realized; compared with existing subjective assessment, objective assessment factors are increased, subjectivity of risk assessment is weakened, and accuracy of risk assessment is improved.

The invention discloses an expression recognition method based on adaptive weighted fusion of significant structural tensors and LBP features. The method makes the image features contain both texturedetail information and texture structure information by fusing the structural tensor features and LBP features, so that the image features have higher description ability. The results show that this method improves the accuracy of facial expression recognition.

The invention provides a LIBS soil detection method, system and equipment for realizing self-adaptive spectral line balance. The method comprises the following steps of exciting laser to be emitted ata maximum energy value, attenuating according to a first attenuation degree and then emitting to the surface of a to-be-detected sample to obtain an original spectrum, calculating the maximum intensity value of the original spectrum, comparing the maximum intensity value with a spectrograph measurement range standard value, and adjusting the first attenuation degree, obtaining a standard spectrumfor the to-be-detected sample, obtaining a laser splitting value, and calculating a laser energy measurement value, acquiring a spark image of the to-be-detected sample, determining the boundary of the spark image, and calculating the spark area and spark intensity, and taking the spark area and the spark intensity as operation parameters, forming a sequence with a standard spectrum and a laser energy measurement value, and entering a neural network algorithm to obtain a final spectrum. By means of the method, spectrum correction is conducted through the spark area, the spark intensity, the standard spectrum and the laser energy measurement value parameters, the obtained final spectrum line is balanced, reliability is high, and the reliability of spectrum analysis is improved.

The invention discloses a method and device for obtaining the resistivity of an anisotropic formation of a horizontal well, and relates to the technical field of geophysical exploration. The method comprises the steps that the relation among the preset horizontal resistivity, the deep and medium induction value difference, the log value of the deep induction value and the resistivity anisotropic coefficient of an experiment formation under different hole drift angle conditions is established through a numerical simulation method; and then the resistivity anisotropic coefficient of the target formation is obtained through fitting and calculating. The anisotropic resistivity obtained through the method can provide more accurate data support for formation mud containing situation analyzing and formation horizontal or vertical resistivity calculation so as to make an analyzing and calculation result more approximate to the practical formation situation, and the accuracy of reservoir assessment is facilitated accordingly.

The invention relates to functional masterbatch of a halogen-free intumescent flame retardant, and a preparation method thereof. The halogen-free intumescent flame retardant in the functional masterbatch is a phosphorus-nitrogen series intumescent flame retardant, and a matrix in the functional masterbatch is a polyolefin resin. The functional masterbatch comprises, by weight, 50-90 parts of the halogen-free intumescent flame retardant, 20-45 parts of the polyolefin resin, 1-5 parts of a compatibilizer; 1-5 parts of a lubricant and 1-5 parts of a surface treatment agent. According to the present invention, dust pollution due to the flame retardant is fundamentally overcome; the flame-retardant masterbatch and polypropylene have good dispersions, such that purposes of environmental protection and no pollution are achieved; heavy metal contents of the functional masterbatch accord with requirements of the report of societe generale de surveillance S.A. (SGS); and the functional masterbatch has a good flame retardant property and high economy.

The invention belongs to the technical field of pipeline leakage detection, and discloses an oil depot pipeline leakage detection method, which is characterized in that an oil depot system state acquisition module detects various operation parameters of oil in an oil depot and an oil path pipeline through a sensor, and detects environmental parameters around the oil depot and the oil path pipeline through detection equipment; the data preprocessing module preprocesses various collected data through a data preprocessing program, and the central control module coordinates the normal operation of each module through a controller and correspondingly processes the data in the whole oil depot system; the pipeline leakage diagnosis module judges whether the pipeline leaks or not through a pipeline leakage diagnosis model; and the display module displays the acquired data and the diagnosis result through a display screen. By arranging the oil depot system state acquisition module, the operation states of the whole oil depot and the oil path pipeline can be comprehensively detected, comprehensive detection data are obtained, and the accuracy of the oil depot pipeline leakage detection result is improved.

The invention relates to the technical field of machine learning, and discloses a knowledge graph representation learning method and device, electronic equipment and a storage medium, which can improve the accuracy of vector representation in a knowledge graph. The method specifically comprises: clustering entities in the knowledge graph to obtain at least two entity clusters; determining the distance between any two entities in the knowledge graph based on the vector representation of the entities; obtaining a score function of triads based on triads describing relationships between entities, each triad including a vector representation of a head entity, a vector representation of a tail entity, and a vector representation of a relationship; on the basis of the distance between the entities, the at least two entity clusters and the scoring function, constructing an objective function, wherein the objective function is in positive correlation with the distance between the two entities belonging to the same entity cluster, and the objective function is in negative correlation with the distance between the two entities belonging to different entity clusters; and updating vector representations of entities and relationships in the knowledge graph by minimizing the objective function.

The invention discloses a speaker embedding learning method, which comprises the steps of performing frame-level feature extraction on a speaker voice segment to obtain a plurality of frame-level depth features, performing second-order pooling processing on the plurality of frame-level depth features to obtain segment-level depth features, and converting the segment-level depth features into segment-level speaker embedding by adopting an affine layer. According to the embodiment of the invention, a second-order pooling processing mode is adopted when the frame-level depth features are converted into the segment-level depth features, therefore, not only is the frame-level speaker feature considered, but also the dynamic information in the voice segment is considered, so that the finally learned speaker embedding can reflect the speaker feature more accurately, the quality of the speaker embedding feature is improved, and the accuracy and reliability of a speaker recognition task performed on the basis of the speaker embedding feature are facilitated.

The invention discloses a commercial vehicle AMT control system and method based on vehicular navigation information. The system comprises an engine controller, a vehicle instrument, an AMT controller and a vehicular navigation system. The engine controller, the vehicle instrument and the AMT controller are all connected with a CAN bus. The controller of the vehicular navigation system is connected with a CAN transceiving module. The CAN transceiving module is connected with the CAN bus. According to the system, the security of the vehicle is improved, the oil consumption is reduced, and according to the scheme, the extra cost of the system is not increased.

The invention belongs to the technical field of traditional Chinese medicinal material identification, and particularly relates to a medicinal material or decoction piece identification method based on an image annotation deep learning algorithm model. In order to overcome the defects in the existing technology for identifying medicinal materials or decoction pieces by artificial intelligence, the technical scheme of the invention is as follows: the method comprises the following steps: (1) image coding: acquiring a picture of a to-be-detected medicinal material or decoction piece, carrying out feature detection on the picture, and extracting to obtain feature data; (2) image decoding: converting the feature data obtained in the step (1) into feature description word vectors through a recurrent neural network model; and (3) identification: identifying the type of the to-be-detected medicinal material or decoction piece through the feature description word vector obtained in the step (2). The invention is suitable for identifying and classifying the types of the medicinal materials or the decoction pieces.

The invention relates to the field of speech synthesis, and provides a speech splicing method and device, and the method comprises the steps: segmenting a rhythm phoneme sequence of a target text, generating a plurality of clause sequences, and enabling the rhythm phoneme sequence to comprise a plurality of phonemes corresponding to the target text and rhythm identifiers located between the adjacent phonemes; voice synthesis is carried out on the clause sequences, multiple pieces of first clause voice information are generated, and the first clause voice information comprises each rhythm identifier and a first duration corresponding to a phoneme; and based on the first duration and the segmentation sequence of the clause sequence corresponding to the first clause voice information in the rhythm phoneme sequence, splicing the plurality of pieces of first clause voice information, and generating target voice. According to the voice splicing method provided by the invention, the naturalness and fluency of the splicing part of the adjacent first clause voice information can be improved on the basis of not needing to preset a voice splicing unit library and not needing to carry out smooth processing on the voice units to be spliced.

The invention discloses to control on output modulation in a hearing instrument. The hearing instrument comprises a main body, and a through hole is formed in front of the main body; the through holeis communicated with the internal of the main body, a plurality of blind holes are formed on the inner wall of the through hole, and the spaces between each two adjacent blind holes are equal; the inner end of the blind hole is communicated with the internal of the through hole, a guiding hole is formed on the outer end of each blind hole, and the guiding hole is collinear with the center line ofeach corresponding blind hole; a plurality of cylindrical slots are formed at the internal of the front of the main body, and the cylindrical slots and the blind holes are in one-to-one correspondence; the cylindrical clots are collinear with the center lines of the corresponding blind holes, and each guiding hole is respectively communicated with the cylindrical slot and the internal of the blindhole; through the mutual cooperation of a first gear, a third gear, a crank, a rack, a second gear and a movable rod, the adhesive tape can swing once when rotating the rotating shaft each time, thereby attracting the attention of the infant, so that the infant cannot be irritable and cry when the medical workers adjust the mode, the infant can cooperate to finish the check, the check can be fastand efficiently finished, and the control is beneficial for the accuracy of a checking result.

The invention discloses a method and a device for determining a text line region, and relates to the technical field of computers. According to one specific embodiment, the method comprises the steps of obtaining a text image, wherein a text in the text image is presented in a text line mode; detecting to obtain center line information of text lines by using a text line detection model based on center lines; and obtaining a text line contour according to the center line information so as to determine a text line region. According to the embodiment, the problem of text adhesion can be effectively solved, text lines can be accurately segmented by extracting center lines of the text lines especially under the condition that the text lines are very close to one another or have overlapped regions, the text line segmentation precision can be improved, and then the accuracy of subsequent text recognition is improved; moreover, the center lines of the text lines are determined firstly, then the text line contour (region) is obtained through detection according to the center line information, and the method is particularly suitable for text detection of long text lines.

The invention relates to the technical field of artificial intelligence, and provides an item recommendation method and device based on reinforcement learning, a model, a medium and equipment. The method comprises the steps that a recommendation value of a target item is predicted according to vector representation of the target item and historical browsing items in a current item list, and the current item list is composed of items containing user browsing behaviors; obtaining a relative feature between the target item and the historical browsing item; based on a reinforcement learning model, determining whether the current item list needs to be filtered or not according to the relative features and the recommendation values; and in response to the need of filtering the current item list, determining to-be-filtered items in the current item list according to the relative characteristics based on the reinforcement learning model, so as to perform item recommendation based on the filtered optimized item list. According to the technical scheme, whether the current item list needs to be filtered or not can be quickly and effectively distinguished, the to-be-filtered item can be determined, and the item recommendation accuracy can be improved.

The invention provides a road traffic sign recognition method based on a neural network. According to the method, equalization processing is carried out on training set images to obtain consistency difference distribution, first conversion, second conversion, third conversion and image enhancement are carried out on the images, a neural network model is improved, and the parameter quantity of the neural network model is reduced. After the model is optimized, compared with a traditional model, the parameter quantity is only 9.23% of that of the original model, but the recognition precision is not reduced, and a better effect is obtained.