38results about How to "Accurate discrimination" patented technology

The invention provides a bonding wire defect detection method based on deep learning, which can be used for identifying the defect type of a bonding wire and automatically detecting the bonding quality problem of the wire in an integrated circuit product, and aims to solve the problems of high labor intensity, low detection efficiency and the like. The process comprises the following steps: establishing a bonding wire acquisition physical platform, acquiring a depth image of a bonding wire through a stereo camera, preprocessing the acquired depth image to enhance the image quality, generating a normal diagram corresponding to the depth image through geometric transformation, and stacking the normal diagram and the depth diagram to obtain a final bonding wire data set. Performing data enhancement on the bonding wire data set to expand the data scale; a Depth-YOLO network model is provided, an input layer of an original YOLO model is reconstructed, and an input feature enhancement module is designed; and finally, training a Depth-YOLO model by using the self-built data set, and carrying out end-to-end bonding wire defect detection by using the model.

The invention proposes an entity extraction method and system based on a dynamic window attention mechanism, and the method comprises the steps: data collection, data preprocessing, and automatic labeling training, and during the automatic labeling training, based on the dynamic window attention mechanism, through improving the attention extraction weight of a TMN frame, the automatic labeling training is carried out. Trigger training, entity recognition model training and entity recognition model testing are carried out. From the perspective, the problem of key information omission or redundancy possibly caused by a fixed window is avoided by introducing a dynamic window attention mechanism, so that the model is more efficient when processing long sentences and complex sentence patterns, and the text information capture capability and the feature extraction effect of the model are remarkably improved. Besides, by combining the similarity matching optimization between the trigger and the sentence, the distinguishing capability of the model in a trigger classification task is remarkably improved, so that the distinguishing of a matched sample and an unmatched sample is more accurate.

The invention relates to the field of factory management, and provides an intelligent management system for a smart factory, and the system comprises a smoke foreground and background distinguishing module, a smoke detection module, a smoke region heating power recognition module and an alarm management module which are carried on an MES platform. The foreground and background distinguishing module is used for carrying out smoke foreground and workshop background distinguishing on workshop video images collected by the visible light camera; the smoke detection module generates a smoke recognition frame from the smoke foreground image through a smoke detection model based on improved YOLOv5, the smoke detection model is combined with an attention mechanism, and an improved loss function combined with an improved frame difference method is adopted; the smog area thermodynamic identification module generates a smog thermodynamic diagram representing smog temperature distribution from the smog identification frame through a smog segmentation network based on UNet; and the alarm management module displays the smoke thermodynamic diagram and manages fire alarm areas and grades according to the smoke thermodynamic diagram. According to the invention, visual intelligent management of the fire smoldering stage of the smart factory is realized.

The invention discloses individualized variable parameter closed-loop ultrasonic stimulation equipment and method based on deep learning, and belongs to the technical field of transcranial ultrasonic stimulation. The individualized variable parameter closed-loop ultrasonic stimulation equipment comprises an electroencephalogram signal acquisition module, an electroencephalogram signal processing module, an electroencephalogram signal deep learning judgment module and an ultrasonic stimulation module; the electroencephalogram signal acquisition module acquires electroencephalogram signals of an experimental subject within a specified finite time and transmits the electroencephalogram signals to the electroencephalogram signal processing module; the electroencephalogram signal processing module preprocesses and displays the electroencephalogram signals in real time, splits the electroencephalogram signals according to specified time, and stores all electroencephalogram signal segments in a picture format; the electroencephalogram signal deep learning discrimination module discriminates an electroencephalogram abnormal level and appoints effective ultrasonic stimulation parameters; and the ultrasonic stimulation module performs variable parameter ultrasonic stimulation on the experimental subject. According to the method, deep learning training is carried out on large sample electroencephalogram signals, an electroencephalogram signal deep learning model is established, the electroencephalogram abnormal level of any other experimental subject is judged, and variable parameter closed-loop ultrasonic stimulation is carried out.

The present application relates to the technical field of sleep monitoring, and provides a method, device, equipment and storage medium for detecting K-complex waves during sleep, wherein the method comprises: inputting EEG data of N channels into a K-complex wave detection model; wherein the K-complex wave detection model comprises a convolutional neural network, a downsampling network with skip connections, a feature fusion extraction network, a normalized stream and a classifier; the EEG data of N channels are processed by the convolutional neural network, the downsampling network with skip connections and the feature fusion extraction network in sequence, and then a first latent code is output; the vector space where the first latent code is located is transformed using the normalized stream to obtain a second latent code that obeys a standard Gaussian distribution; the second latent code is input into a classifier for processing, and after processing, a judgment result of whether a K-complex wave exists in the current EEG segment is output. The present application can improve the accuracy and speed of K-complex wave detection.

The invention discloses a rapid detection method of limestone mineral powder, and belongs to the technical field of road engineering material detection. The rapid detection method comprises the following steps: respectively acquiring a diffraction spectrum and a fluorescence spectrum of a mineral powder sample by using XRD and XRF scanning equipment and an image processing technology; determining mineral composition according to characteristic peaks in an actually measured diffraction spectrogram of the mineral powder sample, and calculating relative mass percent as an index for identifying whether the mineral powder sample is limestone mineral powder or not; the method comprises the following steps: carrying out comparative analysis on a characteristic peak in an actually measured fluorescence spectrum of a mineral powder sample and a standard fluorescence spectrum, and taking the relative mass percent of calcium oxide (CaO), magnesium oxide (MgO) and silicon dioxide (SiO2) in the mineral powder sample as an index for judging the quality of the mineral powder sample. According to the method, the property and the quality of the asphalt mixture mineral powder can be quickly and accurately judged, the operation is simple, effective supervision on the lithology of the rock processed by the limestone mineral powder is effectively realized, and a solid foundation can be laid for subsequently providing a limestone mineral powder quality grading method.

This invention discloses a graphic calibration system and method for CNC mold machining. The system includes: a data acquisition module configured to receive CNC machining tasks and acquire graphic data and measurement data of the mold to be machined based on the machining tasks; a data calibration module configured to perform geometric feature calibration on the graphic data based on preset calibration data, generate calibrated calibrated graphic data, and spatially register the measurement data and calibrated graphic data to obtain deviation data; a data processing module configured to map the deviation data to a machining parameter library in the machining task, calculate the correction values ​​of each machining parameter to obtain calibration parameters; and a data feedback module configured to feed back the calibrated graphic data and / or calibration parameters to the system, and perform the remaining graphic data input and machining parameter input for the machining task. This invention significantly improves mold machining accuracy and production efficiency, providing a high-confidence graphic input benchmark for CNC mold machining.

The present invention discloses a time difference culture device, which relates to the technical field of biological tissue culture; the time difference culture box is adapted to be equipped with a light source assembly, an observation assembly and a gas supply system; the light source assembly includes a first light focusing part, a light decaying sheet is arranged on the light incident side of the first light focusing part, and the light decaying sheet is provided with a plurality of light transmitting parts in sequence along the length direction of the light decaying sheet, and the light transmittance of the plurality of light transmitting parts increases in sequence along the length direction of the light decaying sheet; the observation assembly includes a microscope camera and a linear drive mechanism; the gas supply system includes a first pressure switch, a first reversing pressure relief valve and a first gas storage tank, the first pressure switch is used to monitor the air pressure in the time difference culture box and control the action of the first reversing pressure relief valve, and the outlet of the first gas storage tank is connected to the inlet of the first reversing pressure relief valve after passing through a pressure reducing valve. The present invention can ensure the clarity of the captured image, and the image has a three-dimensional sense and a sense of hierarchy, so as to accurately judge the morphology of biological tissues and ensure the accuracy of biological tissue evaluation, analysis and screening.

The application provides an image recognition-based personalized gait intention recognition method and system, and relates to the technical field of gait intention recognition.The application collects an infrared image pair and a distorted infrared image through a binocular infrared camera integrated with an equipment, respectively calculates and generates two depth maps, and performs weighted fusion to obtain a three-dimensional depth map of an environment under the feet of a user; multi-scale geometric analysis is carried out on the three-dimensional depth map, geometric features are extracted, and a multi-scale environment geometric feature tensor is generated; the tensor is input into a multi-task processing link, and user landing area semantic segmentation is executed in parallel, foot trajectory prediction is combined with real-time posture and historical gait of the user; according to the matching relationship between the semantic segmentation graph and the trajectory prediction sequence, the exploratory or confirmatory landing intention of the user is judged, corresponding intention instructions are generated, the personalized gait landing intention of an exoskeleton wearer can be accurately and robustly recognized, and reliable instruction basis is provided for intelligent control of the exoskeleton.

The application discloses a rock mass characteristic excavability prediction method, and belongs to the technical field of TBM and shield construction prediction. In view of the problems of low accuracy, difficulty in quantifying rock mass structure effect and inability to effectively support full-cycle decision of an existing prediction method, the method comprises the following steps: geological survey is performed on a tunneling section to obtain uniaxial compressive strength, rock mass brittleness index and volume joint number; a rock mass characteristic excavability index is calculated based on the above parameters; the tunneling section is divided into groups according to the index, and tunneling parameters are preset, and equipment selection is performed based on the thrust demand range predicted by the index; during the tunneling process, the actual thrust is compared with the predicted thrust converted based on the index to verify the accuracy, and the coupled function is corrected according to the actual data to realize dynamic optimization. The application is mainly used for equipment selection, tunneling parameter pre-planning and dynamic optimization in the construction process before tunneling, and can significantly improve the prediction accuracy and engineering guidance value.

The invention relates to the technical field of electric power system management, and provides an electric power system inertia safety domain evaluation method, device, equipment and medium, and the method comprises the steps: obtaining each generator speed regulator model parameter of an electric power system, and constructing an electric power system frequency dynamic response model based on each generator speed regulator model parameter; acquiring an active power disturbance quantity after active power disturbance occurs in the power system; on the basis of the active power disturbance quantity and a power system frequency dynamic response model, calculating critical inertia, including a frequency safety critical inertia value and a regulation economy critical inertia value; dividing an inertia operation domain of the power system into a plurality of inertia operation intervals according to the critical inertia; and the current inertia level of the power system is evaluated based on the system real-time inertia and the plurality of inertia operation intervals of the power system. According to the embodiment of the invention, objective quantification and multi-dimensional analysis of the inertia level of the system are realized by constructing a dual evaluation system fusing the frequency safety and the regulation economy.

The invention relates to the technical field of motor control, and discloses an electrical transmission system parameter self-tuning system based on a neural network, and the system comprises a state monitoring and gating module which is used for activating the system when the high-frequency energy characteristic of a torque instruction exceeds a threshold value; the homomorphic mapping and reference reconstruction module is used for constructing a homomorphic digital filter to convert a torque instruction into a homomorphic instruction aligned with a feedback speed time sequence; the momentum residual calculation module is used for calculating the difference value between the theoretical impulse and the actual momentum increment based on the momentum theorem; the morphological decoupling and reasoning module is used for calculating morphological similarity between the residual vector and the reference vector by using the neuron structure so as to output confidence; and the parameter updating module is used for adjusting the adaptive updating gain according to the confidence coefficient and iteratively correcting the inertia parameter. According to the method, through signal homomorphic reconstruction and waveform form reasoning, the influence of loop delay and load disturbance is effectively inhibited, and the accuracy and robustness of parameter setting are improved.

The invention provides a salted pork quality evaluation method based on flavor substance analysis, and the method comprises the following steps: S1, measuring the content of characteristic flavor substances of to-be-detected salted pork, taking OAV of the characteristic flavor substances as an independent variable, taking the flavor difference between the to-be-detected salted pork and representative Zhenjiang salted pork as a dependent variable, and adopting an OPLS-DA method to establish a discrimination model, obtaining a VIP value graph and a standardized regression coefficient graph; and S2, according to the positive and negative correlation of the characteristic flavor substance indexes in the standardized regression coefficient graph in the step S1, referring to the VIP value graph, taking the obtained VIP value as the weight of each characteristic flavor substance index, and adopting a weighted TOPSIS analysis method improved based on a grey correlation coefficient to finally obtain a relative closeness value. According to the method, the internal quality of the salted pork can be comprehensively evaluated, subjective assume is effectively eliminated, the reliability and accuracy of the evaluation result are enhanced, and a basis can be provided for quality control and protection of georationally marked products.

The application discloses a kind of owner actual residence big data detection method and system for smart city, it is related to data processing technical field.The method includes: collecting image to be detected;Obtain target building window area image;Significance detection is carried out, and it is judged whether it is bright window;If yes, then the target building window area image is marked;Otherwise, then obtain bright window template image and dark window template image;The similarity of target building window area image and bright window template image and dark window template image is calculated, and then it is judged whether it is bright window, if yes, then mark;Otherwise, then the corresponding target building window area image is handled using low-light image enhancement method, and then it is judged again.The application combines multiscale target detection method, multiple significance detection model and multiple region color histogram comparison method, and combines low-light image enhancement method, to realize the accurate discrimination of window area image.

The invention relates to the field of geological disaster recognition based on intelligent hidden danger judgment, in particular to a geological disaster monitoring system based on intelligent hidden danger judgment, which comprises an intelligent acquisition module, a displacement analysis module, a soil humidity analysis module, an intelligent hidden danger judgment module and a remote monitoring module, the soil humidity analysis module analyzes the hidden danger of the current geological disaster prone area according to the received soil humidity value, and the displacement analysis module analyzes the mountain displacement change of the current geological disaster prone area and sends the analysis result to the intelligent hidden danger judgment module; the intelligent hidden danger judgment module carries out feature extraction on the current geological disaster and judges the type of the current geological disaster, and the system can adapt to different geological conditions and environment changes and can be widely applied to monitoring and treatment work of various geological disaster prone areas.

The invention discloses a heavy metal element standard exceeding judgment method based on an EADC-Net model, and belongs to the field of element detection. The method comprises the following steps: S1, acquiring a soil sample, and determining an element set and sample information; s2, building a heavy metal element standard exceeding judgment model based on EADC-Net, and determining input and output of the model; s3, a heavy metal element standard exceeding judgment model based on EADC-Net is trained; and S4, judging whether the concentration of the heavy metal element exceeds the standard or not, and finishing qualitative analysis on the standard exceeding of the heavy metal element in the test sample. The method has the characteristics of being scientific, reasonable, simple in process, relatively low in cost, non-destructive, high in accuracy and the like, can solve the problems of X fluorescence energy spectrum peak value overlapping interference, low accuracy of a traditional method and the like, and can effectively, quickly and accurately judge whether the concentration of the heavy metal elements contained in the to-be-detected sample exceeds the standard or not.

The invention discloses a lesion discrimination method and system based on an internal image of a digestive tract, and relates to the technical field of medical image processing, and the method comprises the steps: collecting an endoscopic image containing intracavity pressure change, and carrying out the illumination standardization and polar coordinate system mapping of the image, so as to eliminate the illumination difference and the shooting angle influence. And through low-amplitude step inflation stimulation, obtaining a response frame of an inflation stage, and substituting the response frame into the health potential energy model to generate an anti-fact health image. Thirdly, deformation increment and non-conservation residual errors are calculated based on optimization of a joint energy function, topological difference analysis is conducted through multi-stage pressure trajectory analysis, and a focus area is judged according to structural connectivity changes. And in combination with persistence and abnormal intensity analysis, outputting a lesion judgment result. The problem that illumination change and image quality are inconsistent in a traditional method is solved, accurate judgment of the lesion area is achieved, accuracy and stability of lesion detection are improved, and the method has high clinical application value.

The invention discloses a virus inactivation evaluation method based on image recognition and data processing, and relates to the technical field of biotechnology and medical detection, and the method comprises the following steps: preprocessing a collected microscopic image, and adopting a gray conversion and contrast adjustment method to enhance the light and shade difference between virus particles and polysaccharide residues in the image, so as to obtain a virus inactivation evaluation result; obtaining first image data after primary processing; according to the virus inactivation evaluation method based on image recognition and data processing, through extraction and analysis of morphological characteristics such as boundary integrity and shape regularity of virus particles, objective judgment of a virus inactivation state is realized, so that uncertainty caused by manual subjective judgment is reduced, misjudgment and missed judgment are reduced, and the accuracy of virus inactivation evaluation is improved. The accuracy, the stability and the repeatability of a virus inactivation evaluation result are improved.

The invention provides a batch bar code data acquisition method, system and device and a readable storage medium. The method comprises the following steps: acquiring a to-be-acquired bar code set, a coding type of each bar code and a scanning time interval between adjacent bar codes; analyzing the logic continuity of each bar code; determining the bar code of which the logic continuity does not meet the preset requirement as an abnormal bar code; obtaining a first product batch code of the abnormal bar code; the prefix field of the first product batch code comprises a first production time field, a first production line number field and a first serial number field; according to the prefix field, checking the product batch to which each abnormal bar code belongs; marking barcodes which do not belong to the same product batch in the abnormal barcodes; and rejecting the marked bar codes so as to realize acquisition of batch bar code data. According to the invention, the validity of batch bar code data acquisition results can be improved, and the deviation of the bar code data on the overall level is reduced.

The invention belongs to the field of cables, and provides a high-voltage cable insulation eccentricity quality detection method and a deviation adjusting tool, and the method comprises the steps: S1, drawing an inner identification line at the outer side of a semi-conductive wrapping tape; s2, intercepting a detection sample section; s3, drawing a main outer identification line and an auxiliary outer identification line in the axis direction of the detection sample section; s4, cutting a plurality of test pieces, and insulating the thinnest point position and the thickest point position; and S5, observing the orientation distribution of the thinnest point and the thickest point of each test piece. Compared with the prior art, the method has the advantages that the inner marking line is applied before the conductor enters the three-layer co-extrusion machine head, the main / auxiliary outer marking line with the space direction definition is established on the surface of the take-up sample section, and multi-point slicing detection and space direction reduction are combined, so that the appearance quality of a certified cable is not influenced, and meanwhile, the production efficiency is improved. And reasons causing poor insulation eccentricity can be visually distinguished, so that an adaptive deviation adjusting mode can be selected according to the actual eccentricity condition, and accurate positioning of the real insulation eccentricity position and accurate judgment of causes are guaranteed.

The invention discloses a coal mine goaf type rapid discrimination method for mine disasters, and relates to the technical field of geophysical exploration. According to the rapid discrimination method, goaf type classification is carried out on measurement points in a known non-water-containing goaf, a water-containing goaf and a non-goaf range, a mapping relation between different parameters of each measurement point and goaf types is established, and a model training sample is formed. SPSS software is utilized to carry out Bayes modeling training on the established samples, and Bayes discrimination models of different types of goaf areas and three parameters are constructed. And performing Bayes model training by using SPSS software, and when the reestimation accuracy of a training sample exceeds 80%, inputting the optimal parameter combination calculated by the measuring points of the whole region into the Bayes model to realize rapid judgment of the goaf type of the measuring points of the whole region.

The present application relates to the field of deep learning technology, and discloses a method, device, equipment and storage medium for continuous blood pressure measurement prediction. The method includes: using a blood pressure measurement system to monitor the blood pressure of a target user according to multiple first blood pressure measurement sites, and performing spatiotemporal correlation node discrimination and node screening to obtain multiple second blood pressure measurement sites; constructing a variable structure dynamic Bayesian network model and performing blood pressure state prediction to obtain a first blood pressure state prediction strategy; performing node prior probability density estimation and influence weight adjustment to obtain a second blood pressure state prediction strategy; performing blood pressure state prediction strategy fusion to generate a target blood pressure state prediction strategy; performing blood pressure anomaly detection through an initial Volterra functional series model to obtain a target blood pressure anomaly detection result; performing model optimization and continuous blood pressure anomaly detection, and outputting continuous blood pressure anomaly detection results. The present application improves the accuracy of continuous blood pressure measurement prediction.

The application discloses a steel-cored aluminum stranded conductor broken strand defect identification method and device based on eddy current detection, relates to the technical field of steel-cored aluminum stranded conductor broken strand defect identification, and comprises the following steps: installing an annular vector coil array probe composed of a plurality of excitation coils on a steel-cored aluminum stranded conductor to be measured, constructing a vectorized excitation magnetic field on the surface of the steel-cored aluminum stranded conductor by controlling the driving parameters of the excitation coils; based on the vectorized excitation magnetic field, a multi-layer selective penetration strategy is constructed, and a set of excitation modes for detecting surface aluminum strands, inner layer aluminum strands and the surrounding area of the steel core is sequentially generated; eddy current signals are collected under each excitation mode and layered feature extraction is performed to form corresponding hierarchical feature vectors; fusion discrimination is performed based on the hierarchical feature vectors to generate an identification result of the broken strand defect of the steel-cored aluminum stranded conductor; and the application solves the problems of excitation magnetic field distortion, signal aliasing and insufficient recognition sensitivity of deep defects in the traditional eddy current detection in a multi-layer stranded structure.

The application discloses a kind of shield machine cutter wear detection analysis method, system, equipment and medium, it is related to geotechnical engineering technical field, including steps: according to cutterhead structure features cutterhead is divided into multiple independent areas, installs reference vibration sensor and cutter vibration sensor in each area, each sensor vibration data in the process of tunneling of shield machine is collected;Analysis obtains the vibration index of reference vibration and cutter vibration in different areas of cutterhead;Contrast analysis cutter vibration index and the difference of reference vibration index in different areas of cutterhead, and with difference value judges the wear state of each cutter in the area of current period;Further according to the size of relative difference rate of cutter vibration index and reference vibration index of each area of entire cutterhead, assesses the wear degree of cutter in different areas of cutterhead.The application can find the wear condition of shield cutter in real time, reduce the failure rate of shield machine, improve work safety.

The present disclosure relates to an energy recovery method, device, storage medium and vehicle, which is applied to a pure electric vehicle, can improve the accuracy of slope estimation in the energy recovery strategy, thereby improving the energy utilization rate of the vehicle, and guaranteeing the braking safety and driving comfort of the vehicle. The energy recovery method comprises the following steps: obtaining a first slope value of a position where a vehicle is located from a gyroscope; obtaining a second slope value of the position where the vehicle is located and a third slope value of a road section in front of the vehicle from a high-precision map module; when the driving road condition of the vehicle is a slope road condition, determining a target slope value of the position where the vehicle is located according to the first slope value, the second slope value and the third slope value; and performing energy recovery on the vehicle according to the target slope value.

This invention discloses a fake image detection method based on artifact decomposition. The method includes acquiring the image to be detected and inputting it into a trained three-branch artifact-aware encoder (scene consistency branch, imaging realism branch, and signal naturalness branch) for feature extraction, resulting in three artifact reflection maps. A trained cross-dimensional gated collaborative fusion module is used to fuse the features of the three artifact reflection maps, generating a unified artifact embedding representation. This unified artifact embedding representation is then input into a trained classification head, and the label classification head of the classification head outputs the true / false prediction probability of the image to be detected. This fake image detection method solves the problems of existing technologies, such as inability to uniformly detect multi-source heterogeneous fake images, poor generalization ability, and susceptibility to overfitting.

The present application relates to the technical field of algorithm identification, in particular to a sheep group identity recognition method based on facial features, comprising constructing a sheep face original data set after de-redundancy and data enhancement, realizing sheep face area and three facial key point detection by using an improved RetinaFace network, extracting a 128-dimensional normalized sheep face feature vector through a HTL-Net network fused with TripletAttention, constructing a feature database, developing an intelligent platform integrating detection and recognition, and completing identity matching based on Euclidean distance. The method solves the problems of traditional ear tags, RFID and other recognition methods, such as easy to fall off, high cost, harm to sheep, low precision of existing recognition models, and difficult deployment, and has a sheep face detection accuracy of 96.15%, a recognition accuracy of 91.90%, a lightweight model, supports multi-form real-time recognition, and is suitable for large-scale sheep group precision management requirements.

This application discloses a triple prediction method, apparatus, electronic device, storage medium, and program product, relating to the field of knowledge graph technology. The method includes: acquiring a target knowledge graph; acquiring a first feature and textual semantic representation of entities in the target knowledge graph, and a second feature and textual semantic representation of relations; for any entity, extracting all triples containing that entity and constructing a context representation; fusing the first feature, context representation, and textual semantic representation of the entity to obtain a first fused representation, and fusing the second feature and textual semantic representation of the relation to obtain a second fused representation; acquiring the triple to be predicted, and determining whether it is a missing triple based on the corresponding first fused representation and second fused representation. By implementing the technical solution of this application, the accuracy and robustness of knowledge graph link prediction can be significantly improved.

The application discloses a kind of marker compounds and screening method for distinguishing black tea origin and a kind of black tea origin identification method, belong to tea distinguishing technical field, the application first filters out kaempferol 3-O-(galloyl)-glucoside as the marker for distinguishing black tea origin as marker.It is stable to exist and higher in content in Sri Lanka black tea, while in the representative black tea in China, the content is significantly lower or difficult to detect, and the difference has stability and significance.The identification method of the application comprises: preparing standard extract of the black tea sample to be tested, and using high performance liquid chromatography-mass spectrometry to qualitatively and quantitatively analyze the extract.The scientific and accurate discrimination of the geographical origin of black tea can be realized by detecting the presence or absence of the marker compound or the amount of the marker compound.

The embodiment of the invention relates to the technical field of cloud office and big data pushing, and particularly provides a user behavior big data analysis method and system applied to cloud office resource pushing. And the non-directional pushing condition in the historical office resource pushing event can be accurately positioned. Session scene jump knowledge and collaborative interaction state characteristics are obtained from collaborative office session behavior information, so that the analysis dimension is enriched, and the analysis of the push event is more comprehensive. An original big data push analysis algorithm is combined with the features to obtain an initial push analysis viewpoint, and accurate judgment of a non-directional push result is facilitated. The target algorithm is obtained based on the prior push feedback data debugging algorithm, so that the algorithm accuracy is improved, the target push analysis viewpoint and the prior data are more consistent, a reliable basis is provided for directional push of cloud office resources, and the office resource push accuracy and the user office efficiency are improved.