33results about How to "Efficient and accurate extraction" patented technology

The invention provides a remote acquisition system for operation data of gas equipment, and relates to the technical field of data processing, and the remote acquisition system comprises the steps: capturing the time-varying rule of the state of the gas equipment through learning the long-term dependency relationship in a sequence, and generating key features containing time sequence information; fusing the time sequence characteristic representation output by the long short-term memory network with the equipment attributes and relationships in the knowledge graph to form a characteristic vector reflecting the comprehensive state of the gas equipment; comprehensively analyzing the feature vector to obtain a hidden mode related to the risk of the gas equipment; and based on the hidden mode related to the gas equipment risk, adopting a multi-classification support vector machine to predict the risk level so as to obtain a final risk level prediction result. According to the invention, risk assessment of the operation state of the gas equipment can be realized, so that the safety and the operation efficiency of the equipment are improved.

The invention relates to a power transmission line multi-defect detection method and system based on improved YOLOv8s, and the method comprises the steps: collecting insulator defects and bird nest images on a power transmission line, and obtaining an initial data set; adopting an SAM large model to preprocess the initial data set to obtain a target data set; dividing the target data set into a training set, a verification set and a test set; taking the YOLOv8s network model as a reference model to construct an improved YOLOv8s network model as a power transmission line multi-defect detection network model; inputting the training set into a model for model training; inputting a verification set into the trained power transmission line multi-defect detection network model for model evaluation; and inputting the test set into the model to obtain a power transmission line defect detection result, and evaluating the model. According to the improved YOLOv8s network model, the problems of missing detection and false detection of defects can be avoided, the detection precision of the model is improved, and the light weight of the model is considered at the same time.

The invention relates to a portrait generation and feature similarity recommendation method and system based on a large language model, and the system comprises a data collection module, a feature extraction module, a similarity calculation module and a recommendation module, the data collection module collects multi-modal data from a social media platform, and covers audio, video, text and interactive data information; the feature extraction module performs deep feature analysis and processing based on the collected multi-modal data, and extracts features in multiple dimensions; the similarity calculation module generates a training sample based on the feature matrix generated by the multi-dimensional feature extraction module in combination with a large language model, constructs a similarity matrix of audio, video, text and interactive data features, and calculates similarity scores to determine candidate objects; the recommendation module provides personalized recommendation based on the comprehensive similarity score generated by the similarity determination module. Through combination of a large language model and multi-modal features, the system can accurately analyze behaviors and expressions and generate personalized recommendation.

The invention discloses a high-voltage live cable identification method and system based on a frequency mixing signal, and belongs to the technical field of cable detection, and the method comprises the steps: carrying out the signal collection of a to-be-identified cable based on an electromagnetic induction principle, and obtaining an electric signal; wherein the electric signal comprises a dual-frequency signal for identification and a power frequency noise signal; performing analog-to-digital conversion processing on the electric signal to obtain a digital signal; processing the digital signal based on a phase-sensitive detection technology, and respectively extracting the amplitude and phase of the target dual-frequency signal; and cable identification is completed based on the amplitude and the phase of the target dual-frequency signal. The method can solve the problem that when noise interference is strong, traditional cable identification equipment is affected and cannot accurately carry out cable identification, and the identification accuracy and the identification efficiency can be improved.

The invention discloses a raw material sampling device for photoresist production, which relates to the field of photoresist production and comprises a bottom plate and a storage box fixedly mounted on one side of the top of the bottom plate. In the sampling process, a driving motor drives a first crank through a belt transmission piece to promote a first push rod to perform circular motion so as to drive a first driving block to slide, and the mechanism ensures the swing of a swing rod, so that the reverse acting force of a second driving block is triggered, and a second crank also moves along with the second driving block; the rotating rod transmits force to the sliding block to trigger swinging of the movable rod, the action directly controls movement of the second fixed rod and the sampling piece, it is ensured that the sampling pipe can accurately cut into the sample storage box, stable and consistent sampling is achieved, sampling is conducted under the same standard and condition each time, and therefore the consistency of the sampling amount is ensured, and the sampling efficiency is improved. In addition, due to the automatic sampling process in the closed environment, the external pollution risk is reduced to the maximum extent.

The invention provides a form highlighting method and device based on a large model, computer equipment and a storage medium. According to the method, a to-be-recognized form screenshot is input into an image recognition model to obtain an image recognition result, then a unique identifier is added, a first mapping relation is determined, and a text character string and an identifier character string are further constructed. And then instructing the large language model to generate result information required by a highlighting function based on the information, and finally, responding to a highlighting request of a user based on each piece of result information to carry out rapid positioning and highlighting. According to the scheme, a complete process from the form image to structural information extraction and visual display is realized, and the efficiency, accuracy and convenience of form highlighting processing are improved by effectively depending on a large language model.

The invention provides a method for extracting chemical reaction information from journal papers, and relates to the technical field of chemical informatics. The method comprises the following steps: firstly, extracting a reaction information unit from a journal paper PDF (Portable Document Format) file; carrying out reaction information modal segmentation on the reaction information units; automatically segmenting the extracted reaction information unit into three types of picture information including a chemical reaction expression, a table and a conditional description text; respectively analyzing the segmented chemical reaction expression picture, the table picture and the conditional description text picture; and finally, integrating the analyzed information into complete chemical reaction information. Meanwhile, the extracted, segmented, analyzed and integrated reaction information is stored in an optimized mode, and subsequent use and updating are facilitated. The method can be used for efficiently and accurately extracting reactants, products, reaction conditions, yield, remarks and other information related to the chemical reaction in the journal paper.

The utility model provides a disc type multichannel ultrasonic sample processing device which comprises a shell, a circular groove is formed in the upper surface of the shell, a reagent tube mold is fixedly installed in the circular groove and is in a circular ring shape, and a reagent tube is slidably connected into the reagent tube mold. A radial vibration disc is slidably connected to the circular ring center of the reagent tube mold, and a longitudinal vibration transducer is fixedly mounted at the bottom of the radial vibration disc. By designing the reagent tube mold, the longitudinal vibration energy converter and the radial vibration disc, when the reagent tube mold is used, the motor is started to drive the longitudinal vibration energy converter and the radial vibration disc to move upwards in the reagent tube mold so as to relieve extrusion on the inner wall of the reagent tube mold, a reagent tube can be conveniently inserted into the reagent tube mold, and then the longitudinal vibration energy converter is started to drive the longitudinal vibration energy converter and the radial vibration disc to move upwards. And the radial vibration disc and the reagent tube are driven to shake, so that samples are uniformly dispersed and mixed, non-contact and non-pollution of the samples are realized, and the samples can be extracted more efficiently and accurately.

The invention relates to the technical field of image processing, in particular to an improved YOLOv8n-based power transmission line insulator defect detection method and system, and the method comprises the steps: collecting an insulator defect image in a power transmission line, and carrying out the preprocessing of the image, and obtaining an initial data set; performing data enhancement on the images in the initial data set to obtain a target data set; dividing the target data set into a training set, a verification set and a test set; taking the YOLOv8n network model as a reference model to construct an improved YOLOv8n network model as an insulator defect detection network model; inputting the training set into a model for model training; inputting a verification set into the trained insulator defect detection network model for model evaluation; and inputting the test set into the model to obtain an insulator defect detection result, and evaluating the model. According to the invention, the improved YOLOv8n network model can efficiently and accurately detect and extract insulator defects, and the model has good mobility and generalization.

The invention discloses an innovative process for extracting effective components from tylosin waste liquid for liquid fertilizer, and relates to the technical field of fertilizer preparation, and the innovative process comprises the steps of tylosin pretreatment, extraction separation, concentration purification and nutrition proportioning. On one hand, the ultrasonic cavitation effect can effectively break agglomerated particles possibly existing in the tylosin waste liquid, so that filtering is more sufficient and efficient, on the other hand, the polydopamine coating modified ceramic membrane can better intercept tiny impurity particles, meanwhile, the adsorption and blocking phenomena of impurities on the surface of the membrane are relieved, and the filtering efficiency is improved. Compared with a traditional pure physical filtering mode, the tylosin waste liquid treatment device has the advantages that suspended particles, colloids, part of macromolecular impurities and the like in the waste liquid can be more effectively removed, and the tylosin waste liquid is magnetized, so that the treatment efficiency is improved, and the treatment cost is reduced. The interaction between effective components of the tylosin and other subsequently available components and substances such as an extracting agent can be improved.

The invention discloses a multi-modal cognitive screening method and device based on a brain-computer interface and a galvanic skin bracelet. The device comprises a brain-computer interface module, a galvanic skin bracelet, a signal feature extraction module and a cognitive evaluation module. The brain-computer interface module is used for acquiring an electroencephalogram signal sequence when a user performs a cognitive test; the skin electric bracelet is used for acquiring a skin electric signal sequence when a user performs a cognitive test; the signal feature extraction module is used for performing feature extraction processing on the acquired electroencephalogram signal sequence and the acquired skin electrical signal sequence to obtain an electroencephalogram feature sequence and a skin electrical feature sequence; the cognitive assessment module is used for performing cognitive impairment assessment processing on the electroencephalogram feature sequence, the galvanic skin feature sequence, the skin electric signal sequence and the electroencephalogram signal sequence to obtain a cognitive impairment assessment result.

The invention discloses a rapid extraction method for mixed phase seismic wavelets. According to the method, main feature information of seismic data is extracted by adopting a high-precision iterative complex field basis tracking method, an instantaneous spectrum of the seismic data is efficiently calculated by utilizing analyzed WVD time-frequency distribution, and frequency and phase information of the seismic data is quickly obtained and used for extracting final seismic wavelets. Under the condition of lack of logging data, the seismic wavelets containing the mixed phase information can be accurately and efficiently extracted from limited seismic data, uncertainty caused by insufficient data is effectively avoided, and a key basis is provided for subsequent seismic data processing.

The embodiment of the present application provides an anti-counterfeiting information embedding method, extraction method, device and storage medium, the anti-counterfeiting information embedding method includes determining N embedding areas in the page of the to-be-printed file according to a preset embedding rule; encoding the anti-counterfeiting information according to a preset coding rule to obtain the anti-counterfeiting information code; adding a positioning code and a fault-tolerant code to the anti-counterfeiting information code to obtain the anti-counterfeiting code, the positioning code and the fault-tolerant code are respectively used to locate and fault-tolerantly check the anti-counterfeiting code when extracting the anti-counterfeiting information; according to a preset point block generation rule, the anti-counterfeiting code is converted into an anti-counterfeiting point block set, and embedded in the N embedding areas respectively. By adopting the technical solution provided by the embodiment of the present application, multiple anti-counterfeiting information embedding areas are set in the confidential document, and in the case where the confidential document is torn, the anti-counterfeiting information can be accurately and efficiently extracted based on the spliced ​​file fragments.

The invention provides an emergency lane flatness monitoring and early warning system and a control method thereof. The system comprises a road surface flatness module, a cloud platform data analysis module, a user interaction interface module and a dynamic safety early warning module which are connected in sequence. Firstly, the system carries out initialization and self-inspection, and then enters a vehicle monitoring stage. Then, the system scans the road surface, and uploads and analyzes the collected data. And according to a data analysis result, the system formulates a security and maintenance strategy and updates user information at the same time. In a construction area, the system can give out safety early warning, the system is continuously monitored to ensure normal operation of the system, and the early warning system is reset. The performance of the emergency lane flatness monitoring and early warning system is remarkably improved through a plane optimization algorithm, the safety management process is optimized, and efficient and accurate data support is provided for road maintenance.

This invention discloses a method for constructing a parameter extraction model for active ingredients in traditional Chinese and Tibetan medicines. The method involves acquiring multi-source sample data of traditional Chinese and Tibetan medicines and constructing key derived features from the preprocessed multi-source sample data. A random forest model is used to generate a corresponding importance score for each key derived feature, forming a feature importance vector. The feature importance vector and the key derived features are weighted and summed to obtain a fusion feature, and the variance and correlation coefficient matrices corresponding to the fusion feature are calculated. The average values ​​of the variance and correlation coefficient matrices are weighted and summed to obtain a comprehensive score. The depth and width of the parameter extraction model to be trained are determined based on the comprehensive score. The number of fully connected modules and neurons in the parameter extraction model is determined by the depth and width to obtain the corresponding target architecture. The parameter extraction model under the target architecture is trained using the fusion features and extracted parameter data to obtain a trained parameter extraction model.

The invention discloses a harmonic state evaluation method for an AC / DC hybrid power system, and the method comprises the steps: collecting the voltage and current waveforms of the AC / DC side of the hybrid power system, carrying out the frequency domain analysis, extracting a harmonic component, building an LCC-HVDC power transmission AC system measurement equation, and a converter transformer branch positive and negative sequence harmonic model, and obtaining an LCC-HVDC power transmission system measurement equation; establishing a linear harmonic state estimation model based on a power network, an element, a measurement, a bus voltage measurement equation, a branch current measurement equation and a power transmission system measurement equation; based on the correlation between the measured value and the state quantity, considering the N-1 criterion, and establishing observability constraint of the AC / DC hybrid power system; and based on the observability constraint and the linear harmonic state estimation model, obtaining a stability index of the AC / DC hybrid power system. According to the invention, harmonic state evaluation and risk analysis can be carried out on the to-be-evaluated system according to the harmonic state stability index, and a harmonic treatment optimization suggestion is given.

The invention provides an intelligent question-answering system construction method supporting image enhancement and an intelligent question-answering system, and the method comprises the steps: constructing a user knowledge base, and extracting a picture and a picture title in a document; the extracted picture is stored; searching a title text corresponding to the picture title in the analyzed text segment, and embedding a storage path of the corresponding picture into the title text; representing each text segment as a low-dimensional vector, and storing the low-dimensional vector and the text segments in a vector database; constructing a user question context based on the user knowledge base according to the received user question, and obtaining a text answer of the large language model; and obtaining picture information based on a picture storage path contained in the context of the user question, and displaying the text answer of the large language model and the obtained picture information in a user interaction interface after combining the text answer and the obtained picture information. According to the intelligent question and answer system, the answers not only comprise texts, but also comprise related pictures, so that the user can understand the answer content more intuitively.

The invention relates to a control method of a microwave power amplifier of an unmanned aerial vehicle, and the method comprises the steps: sequentially transmitting signals of different types and intensities to the unmanned aerial vehicle, and enabling an unmanned aerial vehicle signal detection module to recognize the types of received signals, detect the intensities, and give different gains; the microwave power amplifier carries out signal amplification, and the execution system executes corresponding actions according to received signals; acquiring various data of the test flight time period of the unmanned aerial vehicle, and extracting corresponding gain data in which the execution action state of the unmanned aerial vehicle meets specified requirements; constructing a training data sample and inputting the training data sample into the gain prediction model for training; collecting a received signal in real time and inputting the received signal to a gain prediction model for gain prediction; performing corresponding gain amplification processing on the received signal; and monitoring the action execution state of the unmanned aerial vehicle in real time, feeding back the action execution state to the gain prediction model, and optimizing parameters of the gain prediction model. Different amplification processing of signals of different types and intensities is realized, the signal quality is enhanced, and the stability and reliability of the signals are improved.

The application provides a kind of multimodal synergistic industrial voice control method.Application to intelligent manufacturing technology field, the method includes collecting industrial field original voice stream, generates noise spectrum entropy value by frequency domain feature analysis, according to the noise interference level determined by dynamic matching array beam forming weight and adaptive filter coefficient, and gets pure voice instruction by noise reduction;Extract its voiceprint features and semantic text, generate instruction semantic structure tree combined with industrial knowledge graph, calculate identity matching degree and semantic analysis confidence respectively, and get instruction comprehensive effectiveness coefficient by weighting;Satisfy the benchmark time to map it to sub-instruction, evaluate process conflict and equipment risk through digital twin pre-play, determine execution strategy;Real-time acquisition of running data in execution, whether to trigger abnormal rollback is determined according to trajectory deviation rate.This improves the efficiency of industrial voice control.

The invention relates to a video semantic segmentation method and model based on a time sequence difference attention mechanism, and the method comprises the following steps: marking an obtained video image data set, and distributing a pixel category for each pixel in an image; preprocessing the image data set; extracting multi-scale spatial features of a target frame image in the video based on an encoder; according to the reference frame image and the target frame image, utilizing a time sequence difference attention module to extract multi-scale time sequence features of the images; fusing the time sequence features of different scales based on a self-adaptive time sequence feature fusion pyramid; splicing the spatial features and the fused time sequence features to obtain spliced features; and inputting the splicing features into the prediction layer, and outputting a prediction result. Compared with the prior art, the method has the advantages of high segmentation accuracy, good robustness and the like.

The invention aims to provide a rolling bearing weak fault feature enhancement extraction method, and belongs to the field of rolling bearing fault diagnosis. Firstly, a 1 / 3 binary tree filter bank is adopted to divide an original signal into a plurality of frequency bands, then spectrum amplitude modulation (SAM) is carried out on each filtered signal, a discrete normalized envelope spectrum peak factor (DNESP) index is established, the most MO range is selected in a self-adaptive mode, and then a normalized square envelope spectrum is reconstructed. And finally, constructing a normalized envelope spectrum peak factor (NESP) as an evaluation index, quantizing and reconstructing the richness of fault information in the normalized square envelope spectrum, and adaptively selecting a demodulation band with significant fault information according to the richness, thereby realizing enhanced extraction of fault features. The method provided by the invention can effectively extract weak fault features of the rolling bearing, and has good robustness for various interferences and noises.

The invention provides a mine intelligent navigation LED indication system and an indication method, relates to the technical field of mines, is applied to a mine field scene, and comprises a plurality of LED display screens and a plurality of visual identification cameras which are deployed on a mine road in a mine field. Capturing a vehicle image through a visual identification camera, and identifying a license plate number based on a deep learning model; searching a corresponding loading area according to a preset vehicle loading schedule; calculating a first driving path by using the graph model and the weight, and sending the first driving path to the nearest LED display screen; according to the method, the LED display screens and the visual identification cameras are deployed on the mine roads in the mine area, congestion judgment is conducted through the visual identification cameras, when congestion occurs, the Dijkstra algorithm is used for re-planning the path, and a second driving path with the least time is obtained and sent to the LED display screens. And effective management and scheduling of the traffic in the mine are realized.

The application discloses a vegetation evolution three-dimensional dynamic feature extraction and visualization method based on remote sensing data, and belongs to the technical field of remote sensing information processing and ecological environment monitoring. The method comprises the following steps: step 1, basic data acquisition and processing; step 2, horizontal gradient vegetation coverage spatial centroid migration calculation; step 3, vertical gradient vegetation evolution feature identification; and step 4, vegetation evolution three-dimensional dynamic feature extraction and visualization. The method disclosed by the application is based on multi-dimensional space fusion analysis of remote sensing vegetation index and digital elevation model data, can identify the horizontal migration characteristics and vertical gradient evolution characteristics of vegetation coverage, comprehensively consider the horizontal and vertical vegetation dynamic change characteristics, construct a vegetation three-dimensional curved surface model, efficiently and accurately extract and visualize the spatial structure and dynamic change trend of large-scale vegetation, and provide high-precision spatial analysis technical support for ecological environment monitoring, surface vegetation dynamic research and ecological restoration evaluation.

The invention discloses a method for removing fishy smell of marine fish protein by utilizing combined treatment of low-temperature plasma and kaempferia philippica polyphenol, which comprises the following steps of: performing structural modification on myoplasma protein and myofibrillar protein of marine fish by utilizing the low-temperature plasma which is efficient and simple and convenient to operate, so that the structure is expanded to a certain extent and becomes loose; and substances with fishy smell can be removed more easily. In addition, the binding capacity of kaempferia philippica polyphenol and protein hydrophobic groups is higher, kaempferia philippica polyphenol and fishy smell substances compete for binding sites, so that the fishy smell substances are released, the kaempferia philippica polyphenol and the fishy smell substances have a synergistic effect, the marine fish protein fishy smell can be more thoroughly removed, and kaempferia philippica polyphenol substances not only have a fishy smell removal effect, but also have oxidation resistance and antibacterial activity; the effects of efficiently removing fishy smell and inhibiting the growth of microorganisms to a certain extent are achieved.

The invention discloses a zero abnormal sample fault diagnosis method and system for a low-frequency power transformer, and belongs to the technical field of low-frequency power transformer operation state detection. According to the zero abnormal sample fault diagnosis method for the low-frequency power transformer, original voiceprint data are preprocessed by constructing a voiceprint fragment generation model, a voiceprint feature extraction model, a sample fault diagnosis model and a zero abnormal sample fault diagnosis model; dimensionality reduction compression and accurate and efficient feature extraction of the original voiceprint data are realized; meanwhile, the voiceprint multi-dimensional characteristic parameter matrix is used as model input, a sample fault diagnosis model is trained and tested, a zero abnormal sample fault diagnosis model is formed, efficient and accurate recognition of abnormal voiceprint samples of the low-frequency transformer can be achieved, therefore, under the condition of zero abnormal samples, sound samples which are not seen can be accurately diagnosed, and the fault diagnosis accuracy of the low-frequency transformer is improved. And the generalization capability is high, so that the accurate detection and evaluation of the operation state of the low-frequency power transformer can be realized.

This invention belongs to the field of signal analysis and processing technology, and provides a fault diagnosis method and system based on broadband data compression processing. The method includes: acquiring the fault current signal of the optical-pile-load system; decomposing the fault current signal using wavelet transform to obtain the wavelet packet coefficients and decomposition levels for each frequency band of the fault current; calculating the corresponding wavelet energy entropy based on the wavelet packet decomposition results of the fault current signal under various combinations of wavelet functions and decomposition levels, thereby determining the optimal combination of wavelet mother function and decomposition level; performing wavelet packet decomposition on the normal current signal and the fault current signal based on the optimal combination of wavelet mother function and decomposition level, reconstructing the current signals of each frequency band before and after the fault, and calculating the Tsallis entropy ratio of the current signals of each frequency band before and after the fault to determine the characteristic frequency bands of the fault current signal. This invention can accurately extract the fault characteristic frequency bands from the MHz-level broadband electrical state data of the optical-pile-load system.

The invention belongs to the technical field of PDF analysis, and provides a method and device for analyzing a flow chart in a pdf and a medium. According to the method, an open source framework PDFBOX of Apache is utilized to read a PDF file, various closed graphs are identified through a depth-first search algorithm (DFS), duplicate removal is performed on the repeated closed graphs, a formed longest link is searched through the DFS, appearing repeated elements or intersecting links are merged, and finally, all intercepted rectangular images are output. Compared with the prior art, the method has the advantages of high efficiency, accuracy, capability of processing various complex conditions, easiness in expansion and the like.

The present invention relates to the field of medical big data technology in artificial intelligence, and in particular to a tumor knowledge base construction system and method based on a large model and prompt word optimization, the system comprising: a data acquisition module, a prompt word generation and optimization module, an entity recognition and relationship extraction module, a knowledge graph construction module, a clinical decision support module, and a knowledge base update and maintenance module. The present invention uses prompt words that are initially designed and continuously optimized to ensure that the subsequent large language model LLM accurately and efficiently extracts required information, and uses the large language model LLM to quickly and accurately extract related entities and the association relationship between related entities from target text information, so as to construct an accurate hierarchical tumor knowledge graph database. When the hierarchical tumor knowledge graph database is subsequently applied, a personalized clinical report on recommended treatment plans can be generated by inputting the clinical data of the target patient, thereby providing better decision support for medical personnel.

A method and system for extracting entity relationships from an operation and maintenance work order. The method includes obtaining text data from an operation and maintenance work order, annotating the text data according to a preset operation and maintenance knowledge base to obtain text annotation data; performing word segmentation processing on the text annotation data to obtain a word sequence, and determining an annotation word vector of an entity annotation word in the word sequence; determining a semantic vector of the entity annotation word according to the annotation word vector and an entity pair specified by the word sequence, and determining an entity relationship extraction result of the operation and maintenance work order according to the semantic vector and a preset entity relationship extraction model. The solution of the present invention improves the accuracy of entity extraction from an operation and maintenance work order.

The application provides a dynamic gesture recognition method based on light-weight multivariate space-time convolution, which comprises the following steps: constructing a space feature extraction module based on a pseudo 3D gated attention fusion network; extracting multi-scale space features by using the space feature extraction module; injecting a guided heat map by using a gated attention fusion module to enhance key region features and suppress background interference, so as to obtain a space refinement feature sequence; decomposing the space refinement feature sequence into multiple sub-variables; capturing long-range and local time dependence in parallel by using a modern convolution module; and modeling the intra-variable and inter-variable relationships respectively by using a decoupling feature interaction network, so as to obtain a dynamic gesture recognition result. By combining the multivariate feature decomposition strategy with the modern convolution, the double-branch design advantage of the modern convolution is to balance long-range dependence and local details, solve the limitations of traditional convolution in capturing long-range time dependence, and model complex dynamics.