27results about How to "Extract comprehensive" patented technology

The invention discloses a power generation equipment maintenance knowledge question-answering system and method based on a multi-modal large model. The system comprises an equipment question-answering robot module, a data lake module, a large data model module and the like. A data lake knowledge base is constructed; receiving user demand questions for semantic analysis and classification, and determining data types in the demand questions; according to the data type, selecting an appropriate trained RAG large model to carry out comprehensive processing and reasoning, and generating a question answer; and presenting the question answer to the user as a reply. According to the invention, data intercommunication between equipment knowledge questions and answers and maintenance questions and answers is realized through the data sharing module, so that the accuracy of fault diagnosis and the maintenance efficiency are improved; the application of the multi-modal large model and the data management module provides powerful data support for equipment maintenance; and through the real-time monitoring function of the fault alarm module, the equipment fault can be found in time, the downtime is reduced, and the production efficiency is improved.

The invention provides a lightweight human shape detection method and device based on improved RTDETR, a medium and a product, and the method comprises the steps: obtaining a data set for human shape detection, and dividing the data set into a training set, a verification set and a test set; an improved RTDETR model is constructed, an ES Block is used in the improved RTDETR model to replace a Basic Block in the RTDETR feature extraction network before improvement, and the ES Block comprises an OfficientVIT Block and a SimAM attention module which are connected in sequence; the NWD Loss is used as a loss function of the improved RTDETR model; training the improved RTDETR model by using a training set, verifying the improved RTDETR model by using a verification set and testing the improved RTDETR model by using a test set to obtain a trained improved RTDETR model; and inputting the to-be-detected image into the trained improved RTDETR model for reasoning to obtain a human shape detection result of the to-be-detected image. By means of the technical scheme, the human shape detection precision and small target human shape detection in a security and protection monitoring complex scene can be improved.

The invention discloses a wool and cashmere fiber identification method based on a near infrared spectrum. The method comprises the following steps: step 1, establishing a near infrared spectrum data set; step 2, preprocessing to obtain a spectrum sample set; step 3, generating enhanced samples to form positive and negative sample pairs; step 4, constructing a pre-task model of self-supervised learning; step 5, constructing a downstream task model of self-supervised learning; and step 6, the pre-task model and the SVM classifier form a self-supervised learning model based on the near infrared spectrum, and classification is carried out by using the self-supervised learning model based on the near infrared spectrum. The system is simple in structure and reasonable in design, sample enhancement is performed based on band scaling, local smooth disturbance and spectral peak offset, and positive and negative sample pairs are automatically generated, so that the cost and time of data annotation are reduced; deeper features are extracted through the pre-task model, the precision and efficiency of wool and cashmere fiber identification are improved, and the method has a wide application prospect.

This invention discloses a method and system for fault diagnosis of wind turbine gearboxes, comprising: performing VMD variational mode decomposition on the original vibration signal to obtain intrinsic mode components (IMFs), and taking the correlation coefficient ρ. i The largest intrinsic mode component (IMF) is subjected to Hilbert transform to obtain the Hilbert-Huang spectrum. The Hilbert-Huang spectra corresponding to each original vibration signal are divided into source domain data samples and target domain data samples. A deep residual network model is constructed, and the deep residual network model is trained using the source domain data samples and the target domain data samples to obtain a transfer diagnostic model for the target domain. The target domain test data samples are input into the transfer diagnostic model to obtain the diagnostic results. The MK-MMD loss function is used to adjust the deep residual network model so that the model can achieve high-precision fault diagnosis with only a small number of samples.

The application relates to the technical field of collaborative allocation of warehouse resources, and discloses a warehouse resource adaptive collaborative allocation system based on multi-time scale deep reinforcement learning, which comprises a warehouse data acquisition unit, which acquires equipment operation data and environmental parameters and standardizes the same; a warehouse resource state feature extraction unit, which extracts resource state features to form abnormal features; a reinforcement learning state space construction unit, which analyzes the features to form state space construction parameters; a resource allocation strategy generation unit, which constructs a model to generate a strategy and divides influence factors; and a strategy adaptive optimization unit, which updates and optimizes the strategy by using an algorithm. In addition, the system is provided with a multi-source information fusion unit and a task queue dynamic adjustment module. The system realizes adaptive collaborative allocation of warehouse resources, improves warehouse operation efficiency, adapts to dynamic changes of a warehouse system, and meets the demand of modern warehouse intelligent operation.

The present invention proposes a method and system for detecting distracted driving based on local and global knowledge distillation. By introducing group convolution technology in the teacher network, the model's parameter amount and computational complexity can be effectively reduced, while maintaining the effectiveness of feature extraction, so that the model can still extract rich feature information under limited computing resources. The introduction of the attention mechanism enables the model to adaptively focus on the areas in the image that are most relevant to distracted driving behavior, significantly improving the representation ability of features, so that the model can focus more on key information; the combination of local distillation and global distillation enables the student network to more comprehensively analyze the driver's behavior in the inference stage, improve the accuracy of detection, and enable the student network to achieve detection accuracy similar to that of the teacher network while maintaining a low parameter amount.

The application provides a natural language-based public opinion tracking method and device, obtains original Internet public opinions including multiple public opinion samples according to a target theme; obtains a label list of each public opinion sample based on a label library corresponding to the target theme and a title and content of each public opinion sample; classifies the label list of each public opinion sample based on the label library and a keyword corresponding to the target theme, and obtains first-class samples associated with the target theme; labels corresponding key information labels for each first-class sample based on the label library, and obtains an event occurrence time of each first-class sample; aggregates all first-class samples into one or more public opinion sets according to the key information labels and the event occurrence time, and tracks public opinions of each public opinion set. The application can alleviate the problems of key information loss and complex processing in the existing public opinion processing mode.

The application discloses a network protocol design knowledge automatic extraction method based on RFC documents, and comprises the following steps: acquiring a chapter list of the RFC document and a line list in each chapter; acquiring a structured description in the chapter list and the line list to generate a key line list; analyzing the structured description in the key line list to obtain a formatted field list containing partial information and / or an undefined name list of an automaton; supplementing the formatted field list containing partial information and / or the undefined name list of the automaton based on chapter content to which a key line corresponding to the structured description belongs; and obtaining a network protocol design knowledge extraction result based on the formatted field list containing complete information and the automaton migration list. The application can automatically extract relevant network protocol design information.

The application relates to a tumor intraoperative evaluation method and system based on multi-wavelength photoacoustic mathematical statistics analysis, wherein the method comprises the following steps: irradiating a tumor tissue with a pulsed laser during operation, collecting photoacoustic signals at different wavelengths in time sequence according to a preset wavelength sequence, performing mathematical statistics spectrum analysis after pretreatment to obtain multi-wavelength statistical parameters reflecting the characteristics of the photoacoustic signals; performing parameter mapping on the multi-wavelength statistical parameters to obtain a parameter imaging graph reflecting the scattering characteristics and microstructure differences of the tissue, obtaining a parameter spectrum curve based on the parameter imaging graphs at different wavelengths, and extracting auxiliary features based on the parameter imaging graphs; performing visual analysis based on the parameter imaging graphs and the parameter spectrum curve, spatially positioning a lesion area, selecting a ROI and visually displaying the lesion area; and taking the multi-wavelength statistical parameters and the auxiliary features as inputs of a tumor benignity and malignancy evaluation model to output a tumor classification result. Compared with the prior art, the application has the advantages of being capable of realizing rapid and accurate intraoperative evaluation.

The present invention relates to the technical field of traditional Chinese medicine detection, and in particular to a detection method and quality control method for a Tangut rhubarb product. The detection method for the Tangut rhubarb product provided by the present invention uses octadecylsilane bonded silica gel as a filler, uses methanol-phosphoric acid aqueous solution as a mobile phase for gradient elution, selects a specific elution program, and obtains 18 common characteristic peaks, which not only significantly improves the number of characteristic peaks, but also achieves good separation of these common characteristic peaks, the elution program is simple, the obtained characteristic spectrum baseline is stable, each characteristic peak has a good peak shape, and the separation between each characteristic peak is high, so the detection method can effectively improve the accuracy of the characteristic spectrum of the Tangut rhubarb product, and can fully reflect the integrity and characteristic of the Tangut rhubarb product, which is conducive to more comprehensive monitoring of the quality of the Tangut rhubarb product.

This invention discloses a method, device, electronic equipment, and product for monitoring the implementation of land spatial planning, relating to the field of image classification technology. The method includes acquiring aerial images of the target area of ​​the approved plan; extracting SIFT features from the aerial images and local color features of the neighborhood of each feature point in the SIFT features; fusing the SIFT features with the local color features of the neighborhood of each feature point in the SIFT features to obtain a first fused feature; extracting global color features from the aerial images; fusing the first fused feature with the global color features to obtain a second fused feature; and using the second fused feature as input to a pre-trained classification model to obtain the land use type of the target area, so as to monitor the implementation of land spatial planning in the target area based on the land use type. This invention can automatically and dynamically monitor the implementation of land spatial planning in real time, saving human resources.

The invention belongs to the technical field of feed detection, and specifically relates to a rapid detection method for pesticide, veterinary drug and fungal toxin residues in feed, comprising the following steps: using a mixed extract of methanol, acetonitrile and isopropanol to extract a sample under alkaline conditions and acidic conditions in turn, mixing, solid phase extraction adsorbent (including N-vinylbenzene modified ferroferric oxide functionalized magnetic nano-adsorbent material and octadecyl silica gel C 18 After purification by ultra-high performance liquid chromatography-tandem mass spectrometry (UPLC-MS / MS) in multiple reaction monitoring scanning mode, quantification is performed by external standard method or internal standard method. It can quickly and accurately detect 420 pesticides, 163 veterinary drugs and their metabolites and 13 mycotoxins in feed at the same time, with a quantitative limit of 0.01 μg / kg to 50 μg / kg, which is of great significance for ensuring food safety, public health and the sustainable development of animal husbandry.

The application discloses a medical image tumor segmentation model based on a TransUNet framework and a construction method and a segmentation method thereof, the construction method comprising constructing a segmentation model based on the TransUNet, wherein an encoder of the segmentation model comprises CNN modules and a self-defined VSS module connected in sequence, the output ends of each convolution block of the CNN modules are input into a decoder of the segmentation model through a jump connection layer, the self-defined VSS module comprises a plurality of novel conversion layers stacked in series based on a Mamba module and a focus linear attention module, the self-defined VSS module performs multi-level feature interaction and global modeling on first encoding features output by the CNN module to obtain second encoding features; and the decoder performs feature fusion and up-sampling on the second encoding features and the output of the jump connection layer to obtain a segmentation image for a sample image. The application can significantly improve the segmentation robustness and precision of breast ultrasound images with low contrast, high noise and fuzzy boundaries.

The invention discloses an electromyographic signal and gesture intention mapping model, and belongs to the field of artificial intelligence and biological signal processing. According to the method, for double-arm sign language recognition, synchronous electromyographic signals of a left forearm and a right forearm when a user executes a preset sign language gesture are collected, and a training data set is constructed; a double-flow fusion neural network model based on an attention mechanism is constructed to serve as a mapping model, an encoder part of the double-flow fusion neural network model comprises two long and short-term memory network branches for processing left and right arm signals respectively, and a decoder part of the double-flow fusion neural network model comprises a fusion attention layer used for dynamically calculating and fusing weights of double-arm features to generate a joint context vector. And finally mapping to a gesture intention category. The model effectively captures the space-time cooperation relation of the muscle activities of the two arms in the gestures of the two hands, and high-precision and robust mapping from original electromyographic signals to complex gesture intentions is achieved. The invention also protects the training method of the model and the application of the model in a sign language recognition system.

The application discloses a kind of meridian tire X-ray image disease flaw automatic detection method based on improved YOLO-v5 model, belong to meridian tire detection technical field, including following steps: step 1, meridian tire X-ray image is collected and is segmented processing, uniform resolution, model training sample data are made;Step 2, for the image strip, block missing caused by X-ray machine instability, thereby affecting the detection effect Problem, image restoration processing is carried out;Step 3, design improved YOLO-v5 model, including increasing the fourth detection layer, increase attention module, and improve loss function;Step 4, using meridian tire disease flaw data carries out model training;Step 5, using the model of training completion carries out actual application scene tire disease flaw detection.The application is a kind of automatic detection method, based on improved model can automatically identify multiple disease flaws, and different disease flaws are detected and classified, and detection efficiency and accuracy are higher.

The invention provides a sea clutter prediction method and device based on spatio-temporal correlation, and the method comprises the steps: collecting sea clutter data corresponding to a to-be-measured distance unit and two adjacent distance units, and corresponding environment parameter data; modeling is carried out on a time correlation acquisition graph according to the collected sea clutter data, the time sequence dependence characteristic of the distance unit to be measured is obtained, the environment dependence characteristic of the distance unit to be measured is obtained according to the collected environment parameter data, and the high-dimensional prediction characteristic of the distance unit to be measured is further obtained; a left adjacent dependency feature and a right adjacent dependency feature are obtained by combining data corresponding to two adjacent distance units, and temporal correlation is also used in the two adjacent distance units and used for fully mining spatial correlation; therefore, the sea clutter prediction result of the distance unit to be measured is obtained on the basis of the temporal correlation and the spatial correlation, and the sea clutter prediction precision is fully improved.

The application discloses a kind of mobile device image denoising methods based on double-branch residual sparse network, belong to image processing technical field, which includes: residual sparse module, by mixed dilated convolution and residual connection capture local feature in image, while reducing the number of parameters and computational complexity of model;Attention guiding residual sparse module introduces channel attention and pixel attention mechanism on the basis of residual sparse block to adjust the weight of feature map, pay attention to important area in image, improve denoising effect and image quality;Feature fusion module, by adding double-branch output after residual module processing, combined with attention mechanism and activation function, realize high-frequency detail preservation and noise suppression.The application carries out image denoising in unmanned aerial vehicle and other resource-limited scenarios, solves the problems of large parameter quantity, low computational efficiency, insufficient detail preservation and other problems of existing denoising model.

The present invention belongs to the technical field of coal rock carbon content detection, and specifically discloses a coal rock carbon content detection method and system based on multi-feature image spectroscopy technology, the method comprising: constructing a database of typical coal rock types and their corresponding characteristic spectra and texture features; textually annotating the image category of each pixel point in a mixed sample spectral image; using the data in the database as source domain data, and using the hyperspectral data of the coal rock sample to be tested as target domain data, and realizing coal rock classification of the coal rock sample to be tested in the target domain through transfer learning; for the identified coal sample, identifying whether each pixel point on the spectral image of the coal sample contains carbon content; for the pixel point with carbon content, using a trained carbon content prediction model to obtain the carbon content of each pixel point on the spectral image of the coal sample, thereby ensuring rapid and accurate detection of the carbon content of the coal rock.

The present invention discloses a sample automatic processing system and method based on a neural network, which belongs to the field of artificial intelligence and machine learning technology, and specifically includes: capturing sample images through a scanner, applying convolutional neural networks and semantic segmentation technology to locate and separate DNA sample areas; combining an automated workstation and real-time fluorescence quantitative method, the system extracts DNA from samples, and monitors and records extraction process parameters in real time. The extracted DNA is subjected to spectral analysis, and the obtained data is fused with the sample image and process parameters, and input into a pre-trained classification neural network model for training optimization; the trained model is embedded in a control system, and combined with reinforcement learning to realize automatic detection of DNA extraction quality and parameter adjustment; efficient and intelligent processing of forensic evidence samples is achieved, and the accuracy and efficiency of DNA analysis are improved.

This application provides an IQ imbalance calibration method, apparatus, signal transmitter, and zero-IF transceiver. The method includes: acquiring a reference baseband signal and a radio frequency (RF) signal output by the IQ modulation channel based on the reference baseband signal, and converting the RF signal into a modulation baseband signal; reconstructing the modulation baseband signal to generate a baseband feedback signal characterizing the distortion of the IQ modulation channel; obtaining pre-distortion parameters for compensating for IQ imbalance based on the reference baseband signal and the baseband feedback signal; performing pre-distortion processing on the reference baseband signal using the pre-distortion parameters to generate a baseband pre-distortion signal; and performing imbalance calibration on the IQ modulation channel based on the baseband pre-distortion signal. Because the IQ modulation channel is calibrated, the image rejection ratio and error vector amplitude performance of the output signal can be improved, thus increasing the calibration accuracy of the IQ modulation channel.

The application provides a multi-modal disease auxiliary reasoning system and method, a terminal and a storage medium, and particularly relates to the technical field of artificial intelligence, and comprises a multi-modal information acquisition module, a multi-modal information preprocessing module, an entity extraction module, a medical causal knowledge database and a multi-modal classification and discrimination module connected in sequence, realizes preprocessing, entity extraction and modal classification and discrimination of various modal information of a patient, matches the entity with medical knowledge data in the medical causal knowledge database, then extracts various modal features based on the matching result, splices the various modal features according to a set rule, inputs a classifier to obtain a reasoning result. The scheme fully utilizes various modal features, the extracted entity is more accurate and comprehensive, can improve the accuracy of matching the entity with the medical causal knowledge database, has better generalization, the reasoning result is more accurate, and is beneficial to improving the artificial diagnosis efficiency.

The application provides a digital radiograph image lead character recognition method and system, comprising: acquiring a digital radiograph image, and dividing the digital radiograph image into a training set and a test set; cropping the digital radiograph image in the training set to obtain a training set cropped image; manually labeling the training set cropped image; constructing a digital radiograph image lead character recognition network; constructing a loss function of the digital radiograph image lead character recognition network; training the digital radiograph image lead character recognition network to convergence using the training set cropped image; cropping the digital radiograph image in the test set, inputting the trained digital radiograph image lead character recognition network, and obtaining a recognition result. The application can obtain multi-granularity characteristics of lead characters, effectively increase the recognition accuracy of lead characters of different sizes, use a convolutional neural network to obtain image characteristics, and solve the problem of irregular placement of lead characters by using the rotation invariance of the characteristics extracted by the convolutional neural network.

The invention discloses a distribution network automation equipment group fault diagnosis method and system, equipment and a medium, and relates to the technical field of distribution network automation, and the method comprises the steps: obtaining fault data of a distribution network automation equipment group, and carrying out the fault feature extraction; inputting the fault features into the Cauchy distribution probability neural network to obtain an output result with scale parameters, performing parameter optimization by adopting a fractional order war strategy optimization algorithm, and updating the optimized scale parameters to obtain a trained Cauchy distribution probability neural network; and inputting the real-time fault data of the distribution network automation equipment group into the optimal Cauchy distribution probabilistic neural network to obtain various faults. According to the method, data with extreme values or abnormal values are processed through Cauchy distribution, output with scale parameters is better obtained, the convergence speed of the algorithm can be increased through adjustment of fractional calculus, the algorithm can be helped to jump out of a local optimal solution, and therefore the quality of an output result is improved, and the diagnosis effect is improved.

The invention provides an agent generation method and device, equipment and a medium, and relates to the technical field of artificial intelligence, in particular to the technical field of large models, agents and the like. The method comprises the steps that in response to a received agent generation instruction, processing path information used for agent generation is determined, and the agent generation instruction indicates a target webpage; and executing the following operations based on the processing path information to generate a target agent: calling a webpage rendering engine to load a target webpage through the data acquisition node, and obtaining rendered dynamic page data to obtain material data; processing the material data by using a large model through structured extraction nodes so as to extract multiple pieces of field information meeting the structured requirements of the agent generation interface; and through the agent generation node, calling an agent generation interface based on the multiple pieces of field information to generate a target agent.

The invention discloses an underground wireless energy and signal transmission system, and relates to the technical field of wireless power transmission. Comprising an electric-magnetic energy conversion module, a magnetic field reconstruction module, a primary side energy and signal transmitting module, a primary side signal receiving module, a primary side current and voltage sampling circuit, a secondary side energy receiving module, a secondary side signal receiving module, a secondary side current and voltage sampling circuit, a secondary side maximum power point tracking BMS module and a secondary side signal transmitting module. According to the invention, the problems of low energy transmission efficiency, strong signal interference and poor high-temperature adaptability in a thick metal shielding environment are solved, and reliable underground non-contact power supply and communication are realized.

The main purpose of the present invention is to provide a method for identifying illegal buildings using a drone based on a multi-threaded training mechanism. The method first uses a drone adaptive image capturer to collect images and obtain multiple original high-altitude building images. The multiple original high-altitude building images are then preprocessed to generate multiple standard high-altitude building images and multiple groups of authentic illegal building discriminant vectors of standard high-altitude building images. According to the multiple standard high-altitude building images, an illegal building identification model based on a multi-threaded training mechanism is constructed to obtain multiple groups of predictive illegal building discriminant vectors. The multiple standard high-altitude building images, the multiple groups of predictive illegal building discriminant vectors and the multiple groups of authentic illegal building discriminant vectors are used as data samples to train the illegal building identification model based on the multi-threaded training mechanism, and finally the efficiency and performance of the illegal building identification model are improved. The present invention solves the problems of slow training speed, unstable convergence and poor recognition performance in traditional illegal building identification methods.

This invention discloses a method for intelligent identification of elemental sensitivity analysis in shale sub-layers using elemental logging, comprising the following steps: Step S1, mining the identification features of shale sub-layers with multiple elemental single-double composite intersections; Step S2, screening homogeneous elements based on Siamese neural networks; Step S3, heterogeneous element sensitivity analysis based on the generalization property of residual neural networks. Using the method designed in this invention, the systematic, objective, and comprehensive extraction of sensitive elements from shale sub-layers can be achieved, automating the analysis of sensitive elements in shale sub-layers.