30results about How to "Accurate feature extraction" patented technology

The invention discloses an intelligent customer service method for adaptively adjusting an interaction strategy by using reinforcement learning, and the method comprises the following steps: S1, obtaining a natural language text inputted by a user, and generating a multi-dimensional state vector; s2, grouping the users by using a real-time dynamic clustering algorithm, and updating a clustering result; s3, generating an action space in a specific state, and enriching the action space in combination with a rule-based strategy and a learning-based strategy; s4, aggregating neighbor node features through a graph convolutional network, and extracting local features; s5, weighting adjacent node features by using a graph attention network, and dynamically allocating attention weights; s6, carrying out interaction strategy optimization by adopting a spiral ladder type deep exploration network, and optimizing action selection; and S7, combining online learning with an experience playback mechanism, continuously updating an interaction strategy, and making a quick response to dynamically changing user demands. According to the method, reinforcement learning, the graph neural network and the sentiment analysis technology are combined, and intelligent customer service interaction strategy self-adaptive adjustment and optimization are achieved.

The invention belongs to the technical field of data processing, and discloses a data development method and platform based on multi-source data integration, and the method comprises the steps: collecting and preprocessing multi-source heterogeneous data, obtaining multi-source initial data, extracting initial features of different data sources, and forming a preliminary feature set; establishing and optimizing an event tree according to the collected multi-source heterogeneous data and the corresponding preliminary feature set, mining the multi-source data collected in real time based on the optimized event tree, analyzing the multi-source data collected in real time according to an established risk assessment model, performing comparative analysis on an actual risk event of an obtained risk prediction result, and obtaining a risk prediction result. Data access and operation are carried out through the user interface, efficient integration, processing, analysis and application of data are achieved, and the overall level of data development and utilization is improved.

The invention relates to the technical field of artificial intelligence, and discloses an artificial intelligence-based multi-modal document analysis system, which consists of an image preprocessing module, a multi-modal feature extraction module, a natural language processing module, a modal fusion and analysis module and a document structure identification module. According to the method, the image quality is improved through image sharpening and noise and blurring removal, text, table and image areas are extracted through masks respectively, different document contents are separated accurately, visual features and extracted area information are combined together, a comprehensive feature vector is generated, the document contents are understood more comprehensively, and the document quality is improved. According to the method, the text extracted from the image is analyzed, semantic analysis and keyword extraction are carried out, so that text content is understood, keywords are extracted from the analyzed text, main topics and key points of the document are recognized, and the beneficial effect that the analysis result is more accurate due to multi-modal analysis of the document is achieved.

The invention discloses an infrared expressway foreign matter detection method based on a Mama framework, and the method comprises the steps: continuously carrying out the data collection of expressway scenes under different illumination conditions and different weather conditions through an infrared sensor, obtaining infrared image data, carrying out the preprocessing of the infrared image data through data augmentation, generating training data, and carrying out the detection of foreign matter in an expressway. Dividing the training data into a training set, a test set and a verification set in proportion; constructing a detection network based on a Mama framework, wherein the detection network comprises a student network and a teacher network; training the teacher network through back propagation based on the training set; training a student network based on the result of the teacher network and the real data to obtain a target detection network; and inputting a to-be-identified image into the target detection network to obtain a detection result. According to the invention, the method achieves the quick and accurate recognition of potential risks, guarantees the high detection precision through the combination of the infrared imaging technology and the powerful processing capability of a Mama frame, achieves the real-time response to foreign body threats, and remarkably improves the efficiency and safety of expressway monitoring.

The invention discloses a railway dispatching safety data aggregation processing method based on AI model training, and the method comprises the following steps: S1, collecting data related to railway dispatching from a plurality of data sources, including train operation data, dispatching commands, track conditions, signal information and meteorological data, carrying out the cleaning and standardization processing of the collected data, and obtaining a standard data set; the method comprises the following steps: S1, selecting a data set, S2, integrating different types of data by using a multi-modal fusion method in deep learning, S3, extracting key features from the fused data set, performing time sequence feature extraction by using an auto-encoder method, and performing feature selection by using XGBoost, S4, performing AI model training by using the processed data set, and selecting a random forest for training; and S5, applying the trained AI model to real-time data. According to the invention, a solid technical support is provided for improving the safety and the intelligent level of a railway dispatching system.

The invention discloses a COD (Chemical Oxygen Demand) prediction method based on a two-channel convolution time sequence mechanism, which comprises the following steps of: acquiring COD original data for preprocessing to obtain COD input characteristics; constructing a prediction model, and importing data into a dual-channel convolution module; importing the first double-channel convolution output feature vector and the second double-channel convolution output feature vector into a double attention memory network to obtain a time sequence dependence feature; the time sequence dependence characteristics are imported into an output module, and predicted COD output is obtained; training the model, and in the training process, optimizing hyper-parameters of the model based on an improved pelican algorithm; and importing real-time COD original data into the trained model, and obtaining real-time COD output. According to the method, a high-efficiency effluent COD content prediction model of the sewage treatment plant is constructed by creatively fusing the two-channel convolution and the double-attention memory network.

The invention belongs to the technical field of image analysis, and particularly relates to a mulberry growth period identification method and system based on image analysis, and the method comprises the steps: obtaining image data of each growth period of a mulberry, preprocessing the image data, and inputting the preprocessed image data into a feature extraction module; performing feature extraction on the image data in each growth period, and establishing a feature comparison table of the images in each growth period of the mulberry trees; mulberry image data of each mulberry planting area to be identified is collected in real time, and after preprocessing, the mulberry image data is input into the feature extraction module for real-time image feature extraction; inputting the real-time image features into a feature comparison module, calling a mulberry growth period image feature comparison table, and matching the real-time image features with the mulberry growth period image feature comparison table; obtaining the mulberry growth period of each mulberry planting area to be identified according to the matching result of the real-time image feature and the mulberry growth period image feature comparison table; a preset mulberry growth period management measure table is called, management measures of the mulberries in the mulberry planting areas are obtained, mulberry planting management is facilitated, and the economic benefits of mulberry planting are improved.

The present invention discloses an infrared highway foreign body detection method based on Mamba architecture, comprising: continuously collecting data of highway scenes under different lighting conditions and different weather conditions through infrared sensors to obtain infrared image data, preprocessing the infrared image data through data augmentation to generate training data, and dividing the training data into training set, test set and verification set in proportion; constructing a detection network based on Mamba architecture, including a student network and a teacher network; training the teacher network through back propagation based on the training set; training the student network based on the results of the teacher network and real data to obtain a target detection network; inputting the image to be identified into the target detection network to obtain the detection result. Rapid and accurate identification of potential dangers is achieved, and the powerful processing capability of the infrared imaging technology and the Mamba framework is combined to ensure high detection accuracy, achieve real-time response to foreign body threats, and significantly improve the efficiency and safety of highway monitoring.

The invention discloses a clock travel time accuracy automatic detection method based on machine vision, and relates to the technical field of machine vision. The method comprises the steps of capturing an image and a video stream of a target clock in real time through a high-resolution camera, performing preprocessing, extracting clock dial and pointer characteristics through Hough transform and an image segmentation technology, calculating a pointer angle, matching the pointer angle with a standard dial image, and obtaining a static error index; the method comprises the following steps of: dynamically tracking a pointer motion track through an optical flow method and a Kalman filter, analyzing the influence of environmental factors on a travel time error, obtaining a dynamic error index, and finally, comprehensively analyzing a static error and a dynamic error, calculating the travel time error index of the clock, and comparing the travel time error index with an error threshold value, so as to judge the travel time accuracy of the clock. The method can automatically and accurately detect the timekeeper travel time error, improves the detection efficiency, reduces personal errors, and comprehensively analyzes the influence of environmental factors on the timekeeper travel time precision.

The invention provides a sensor array data transmission method which is applied to an FPGA, the FPGA is provided with a PDM digital interface and a DVP digital interface, multichannel PDM signal streams output by a digital sensor array are received through the PDM digital interface, the multichannel PDM signal streams are converted into multichannel PCM data streams in a synchronous digital modulation mode, and the multichannel PCM data streams are transmitted to the FPGA through the DVP digital interface. Dynamic fragmentation compression processing is carried out on the multi-channel PCM data flow through an equal probability sampling method; the method comprises the following steps: receiving a video data stream output by an image sensor through a DVP digital interface, and compressing the video data stream through a H264-like video data compression method; the multi-channel PCM data stream is subjected to combined coding, the coded PCM data stream is interspersed in the compressed video data stream through an interpolation method, and the coded PCM data stream and the compressed video data stream keep time synchronization; and packaging and uploading the interspersed and mixed data flow to a cloud server through a gigabit network communication protocol.

The present application provides an image processing method and system for rehabilitation training motion analysis, which relates to the field of image processing technology. By establishing a sample motion image set and annotating it, a feature recognition channel is established to perform rehabilitation motion fitting, and a time-series rehabilitation motion fitting result is established; a joint image capture of the user's rehabilitation motion is performed to establish a joint time-series image set including local images and global images; contour extraction is performed on the global image and the local image, and feature recognition is performed on the contour extraction result using the feature recognition channel to establish a feature recognition result; a user motion time-series set is constructed based on the corresponding image time sequence; and image annotation is performed on the user motion time-series set using the time-series rehabilitation motion fitting result. The present application solves the technical problem that the prior art cannot effectively distinguish and process image information from different perspectives and different parts when performing image recognition, resulting in low accuracy in rehabilitation training motion analysis, and improves the accuracy and efficiency of rehabilitation motion analysis.

The invention discloses a bolt pre-tightening force monitoring system and method based on ultrasonic waves. The system comprises an ultrasonic probe, a data acquisition and processing analyzer and the like, a probe array is uniformly mounted on the surface of a bolt, an improved depth residual network is arranged in the data acquisition and processing analyzer and can process a flight time difference signal, and an audible and visual alarm can trigger a three-level alarm according to pre-tightening force deviation. The data remote transmission device provides data for the digital twin system, the display unit displays a pre-tightening force distribution atlas in real time, and a digital twin system platform generates a three-dimensional visual dynamic model and the like. According to the method, bolt pre-tightening force monitoring is achieved through the system. The system is high in measurement precision, convenient to install, capable of achieving real-time early warning and intelligent operation and maintenance, and suitable for various industrial fields.

The invention provides a rapid intelligent identification method and device for tunnel diseases and a medium, and relates to the field of target classification, machine vision and intelligent identification, and the method comprises the steps: constructing a sample library through obtained unlabeled ground penetrating radar images; constructing a self-supervised contrast learning network and a disease detection network; training a self-supervised contrast learning network through the sample library, and extracting tunnel ground penetrating radar wave features of the ground penetrating radar image in the sample library to obtain feature extraction network parameters; the feature extraction network parameters are migrated to a backbone feature module of the disease detection network; screening and marking disease radar response data in the sample library, and constructing a tunnel disease ground penetrating radar image set; training the disease detection network according to the tunnel disease ground penetrating radar image set; and obtaining a to-be-detected tunnel disease ground penetrating radar image, inputting the to-be-detected tunnel disease ground penetrating radar image into the trained disease detection network, and identifying a tunnel disease result in the to-be-detected tunnel disease ground penetrating radar image.

The invention discloses an iron tower foundation crack detection method based on pixel focusing attention, and relates to the field of crack detection. The invention aims to solve the problem that the existing iron tower foundation crack detection accuracy is low. The method comprises the following steps: acquiring a training set by using an iron tower image data set, and training an iron tower foundation crack detection network by using the training set; inputting a to-be-detected iron tower foundation crack image into the trained ResNet network, and obtaining a to-be-detected iron tower foundation crack image feature map; and forming a to-be-detected query sequence by using the crack prediction type and the crack target prediction frame position of the iron tower image data in the training set, and inputting the to-be-detected query sequence and the to-be-detected iron tower foundation crack image feature map into the trained iron tower foundation crack detection network. And obtaining a crack prediction type and a corresponding probability in the to-be-detected iron tower foundation crack image and a prediction position of the to-be-detected iron tower foundation crack target detection frame. The method is used for detecting the iron tower foundation crack type.

The present invention discloses an intelligent customer service method for adaptively adjusting interaction strategies using reinforcement learning, comprising the following steps: S1, obtaining natural language text input by a user, generating a multidimensional state vector; S2, applying a real-time dynamic clustering algorithm to group users, and updating clustering results; S3, generating an action space of a specific state, and combining a rule-based strategy with a learning-based strategy to enrich the action space; S4, aggregating neighbor node features through a graph convolutional network, and extracting local features; S5, applying a graph attention network to weight neighbor node features, and dynamically assigning attention weights; S6, using a convoluted ladder-type deep exploration network to optimize interaction strategies and optimize action selection; S7, combining online learning with an experience playback mechanism, continuously updating interaction strategies, and responding quickly to dynamically changing user needs. The present invention combines reinforcement learning, graph neural networks, and sentiment analysis technology to achieve adaptive adjustment and optimization of intelligent customer service interaction strategies.

The invention belongs to the technical field of cargo transportation systems, and particularly relates to a waybill-free express cargo identification system based on data labels, which comprises a data acquisition module and a model training module, according to the invention, through combination of the image and the sensor data, real waybill-free operation is realized, and a user does not need to manually fill in a paper waybill. The design not only improves the operation efficiency, but also conforms to the digital and paperless development trend of the modern logistics industry. A traditional system often depends on paper waybills or needs to manually input a large amount of information, and operation is tedious. A user interface and a feedback module are designed in the system, a user is allowed to confirm and correct an identification result, and feedback data are used for subsequent model retraining and optimization. The self-improvement mechanism enables the system to continuously learn and improve the recognition accuracy. However, many existing systems lack an effective user feedback mechanism and are difficult to realize continuous improvement.

The invention discloses an intelligent electronic medical record management method and system based on big data, and the method comprises the steps: obtaining medical image data of a patient, eliminating image noise in the medical image data through a Gaussian filter, and carrying out the feature extraction of a medical image after noise elimination through a convolutional neural network; detecting the medical image after feature extraction to obtain a focus area, generating annotation data according to the focus area, and associating with a medical record database through the annotation data; according to the association relationship between the annotation data and the medical record database, summarizing the annotation data and the patient information in the medical record database, filling a report template, converting the report template into report characters through a natural language processing model, issuing the report characters, and updating the medical record of the patient; according to the method, the Gaussian pyramid is established, the focus area in the image is detected, the annotation data is generated, and the annotation data is associated with the medical record database, so that the image processing efficiency and the focus detection precision are improved, and the electronic medical record management efficiency is enhanced.

The invention discloses a feature extraction method based on a directional space and a multi-scale feature extraction network and a detection system, relates to the technical field of cloth flaw detection, and solves the problem of low feature extraction precision when feature extraction is performed on fine flaws with different shapes in the prior art. The method comprises the steps that a textile cloth image is acquired, feature extraction is conducted on the textile cloth image through a directional space and multi-scale feature extraction module, and the directional space and multi-scale feature extraction module carries out feature extraction on the textile cloth image; the directional space and multi-scale feature extraction module adopts a multi-branch parallel gradient stacking mode to arrange an information flow of network internal feature extraction, and a re-parameterization convolution module is adopted as a branch feature extraction component in the directional space and multi-scale feature extraction module, so that spatial features of warp and weft flaws can be extracted more effectively; therefore, the feature extraction is more accurate, and the defect features from the global to the local and from the small scale to the large scale can be captured.

This invention relates to a method for NAION prediction and acute phase differentiation based on deformable convolution and multi-site OCTA, belonging to the field of ophthalmic disease diagnosis and medical image processing technology. This invention acquires OCTA images of ocular samples and extracts data from key sites. Using the OCTA deformable convolution feature extraction module as the core feature extraction backbone network, its optimal performance is verified through single-site experiments, and a multi-site joint diagnostic framework is constructed. Radiomics features are introduced, and deep fusion of deformable convolution features and radiomics features is achieved through a bidirectional cyclic feature interaction module. Finally, a hybrid expert module is added to the three-site joint model to decouple NAION prediction from the acute phase differentiation task of NAION / ON, improving diagnostic accuracy. This invention effectively solves the problems of difficult accurate differentiation between the acute phases of NAION and ON and insufficient early prediction of NAION, providing reliable technical support for ophthalmic clinical diagnosis.

The present invention provides a surface morphology reconstruction method and system integrating machine learning and spatial frequency analysis, including: step 1: using the analytical formula of the spatial frequency method to generate a large number of two-dimensional arrays representing rough surfaces; step 2: by analyzing the height distribution and frequency energy proportion of the generated surface, combined with the machine learning method, extracting the characteristic variables of the analytical formula; step 3: combining the test or simulation method to establish the relationship between the characteristic parameters and the surface performance, which is used for performance prediction and surface reconstruction in specific scenarios. The present invention can realize the surface morphology feature extraction based on the spatial frequency analysis method, reconstruct the rough surface suitable for current carrying performance prediction, and thus improve the practical significance of the surface reconstruction method in performance prediction.

The present invention provides a micro-expression fitting method and system based on multiple feature points. The method comprises: obtaining a reference image and an emotional image, wherein the reference image refers to an image without a stimulus source, and the emotional image refers to an image with a predetermined stimulus source; inputting the reference image and the emotional image into a facial detection neural network for facial detection to obtain a reference facial image and an emotional facial image respectively; inputting the reference facial image and the emotional facial image into a facial feature point marking network for facial feature point extraction to obtain reference feature points and emotional feature points respectively; calculating a facial micro-feature vector according to the position information of the reference feature points and the position information of the emotional feature points, and outputting the facial micro-feature vector that meets a predetermined threshold as a facial motion unit; and fitting the facial motion unit using a recursive neural network model to obtain predicted emotional features.

The invention relates to the technical field of image display, in particular to a metal metallographic structure image classification display method and system. The method comprises the following steps: preprocessing a metal metallographic structure image, calculating and evaluating an image feature index Tzls, quantifying the overall quality of the metal metallographic structure image, automatically screening low-quality and repeated images, and reducing stored and processed invalid data; classifying the metal metallographic structure images by using two characteristics of structural factors and defect factors of the metal metallographic structure; the metal metallographic structure image is classified and graded by calculating and evaluating a structure characteristic index Jtzs and a defect characteristic index Qtzs; and all grade division results are transmitted to the image classification display module, so that timely and accurate data transmission is ensured, a user searches and filters according to a structure grade or a defect grade, a required image is quickly found, and the working efficiency is improved.

The invention provides a canteen evaluation sentiment analysis method, system and device and a storage medium, and relates to the field of text data analysis. The method comprises the following steps: acquiring first text data containing dining feedback information of a canteen; performing emotional polarity labeling on the first text data to obtain second text data; performing vectorization processing on the second text data by utilizing a pre-training language model to obtain text vector data; performing feature extraction on the text vector data by using a bidirectional gating loop unit network, and calculating the weight of each sentence and character for sentiment classification in combination with a double-layer attention mechanism to obtain text feature data; and performing sentiment classification on the text feature data by utilizing the classifier, and generating a sentiment analysis report according to a sentiment classification result. According to the scheme, canteen evaluation sentiment analysis can be flexibly and accurately carried out without the help of a large number of labeled data sets.

The invention relates to the field of image processing, and discloses a machine vision image generation method, which comprises the following steps of: firstly, scanning a three-dimensional object target according to a task requirement of a machine, and establishing a knowledge base; establishing a machine vision detection model based on the knowledge base; depicting a plurality of target calibration points on the three-dimensional object target, and performing calculation to obtain a mapping conversion relation; generating a target image according to the machine vision detection model and the corresponding mapping conversion relation; updating the knowledge base in real time, and maintaining and correcting the target image by using the knowledge base; the method solves the problems that in the machine vision imaging process, a linear image is prone to geometric distortion, feature extraction is more accurate through the image processing technology, knowledge base data is converted into an ordered tree structure, a selection algorithm is used for replacing a traditional traversal algorithm, and a fuzzy classification algorithm is combined to achieve rapid recognition of a three-dimensional object.

The invention discloses a method and a device for generating a long-term random scene in new energy output, and belongs to the field of power grid operation scheduling. The new energy output medium-and-long-term random scene generation method comprises the steps of performing dimension reduction on a historical day scene to extract significant features of the historical day scene; clustering the salient feature set of the historical day scene by adopting an affinity propagation clustering algorithm, and dividing the generated day scene into a corresponding cluster; then simulating evolution characteristics among different categories of day scenes based on a Markov chain; and generating a medium and long term random scene through Monte Carlo simulation. According to the method, probability distribution, time sequence characteristics and spatial correlation of medium and long-term scenes of new energy output can be accurately fitted, and effective data support can be provided for medium and long-term power transaction and operation mode formulation.

The invention relates to a spatio-temporal trajectory data character representation and feature extraction method, and belongs to the technical field of trajectory data processing. Through the steps of preprocessing, standard vectorization, unit circle character processing and feature extraction, efficient characterization and depth feature mining of spatio-temporal trajectory data are realized. According to the method, data missing is effectively filled, data processing efficiency is improved, data interpretability is enhanced through character representation, meanwhile, the feature extraction process is optimized, calculation complexity is reduced, and data analysis precision is improved. The method is suitable for various spatio-temporal trajectory data, has wide applicability and high flexibility, is easy to implement, and provides a novel and effective technical means for the fields of trajectory data analysis, clustering, pattern recognition and the like.

The present invention discloses a method, device, equipment and storage medium for identifying components in a power grid wiring diagram. The method includes: creating a ledger for recording component information in the power grid wiring diagram, and setting multiple feature categories for the components in the ledger; based on the feature categories, using a pre-trained convolutional neural network to extract multi-dimensional semantic features of each component in the power grid wiring diagram, and obtaining a semantic feature matrix of each component; for each component, adopting a multi-criteria evaluation method, designing weights and reliability for each semantic feature of the component, and using weights and reliability to construct the evidence credibility of each semantic feature of each component; for each semantic feature of each component, inputting the evidence credibility and the semantic feature matrix into a pre-trained evidence reasoning classifier for the component, and using the evidence reasoning classifier to perform multiple binary classifications on the component, obtaining the classification results of the component and the feature category, and saving the classification results to the feature category of the ledger.

The invention relates to the technical field of power transformers, and provides a transformer voiceprint detection method and system based on edge intelligence and deep learning, and the method comprises the steps: S1, collecting an original voiceprint signal, and carrying out the preprocessing and voice enhancement, and obtaining a voiceprint signal; s2, carrying out real-time analysis and feature extraction on the voiceprint signal by the edge computing node by utilizing a preliminary analysis deep learning model, and judging whether voiceprint abnormity exists or not; s3, when the edge computing node judges that abnormity exists, the voiceprint signal and related feature data are uploaded to a cloud big data platform; s4, collecting massive voiceprint data, constructing and continuously updating a standard voiceprint library, and optimizing to generate a deep analysis model; s5, issuing the optimized depth analysis model or model parameters to an edge computing node, and updating the preliminary analysis model; and S6, identifying the type and severity of the defect, carrying out fault positioning, and triggering an early warning and decision instruction. According to the invention, the voiceprint of the transformer can be well detected.

A difference guided multi-directional complementary remote sensing land use classification mapping method comprises the following steps: S1, acquiring a public remote sensing image dataset and performing cropping; S2, constructing a multi-class pixel-level classification model for land use classification and mapping, and training the model by using the cropped remote sensing image dataset; S3, processing a remote sensing image to be classified and mapped by using the trained multi-class pixel-level classification model to obtain a land use classification map; the model adjusts the attention distribution of aggregated features by using difference information to enhance the difference between classes with similar textures; meanwhile, four sub-bands obtained by wavelet decomposition are used to extract edge features in different directions, and texture information is used to position edge details to retain the contour and boundary information of small-area ground objects. The present application can perform pixel-level classification on multi-class ground objects in a remote sensing image and generate a land use classification map.

The present invention relates to the technical field of blood oxygen saturation, and specifically to a method for detecting local brain tissue blood oxygen saturation, comprising the following steps: S1, based on the original spectral data of the brain tissue, using high-pass and low-pass filter methods to perform signal preprocessing, remove low-frequency and high-frequency noise in the signal, and use discrete Fourier transform and wavelet transform to perform frequency domain analysis and filtering processing to generate preprocessed spectral data. Based on the original spectral data of the brain tissue, the present invention uses high-pass and low-pass filters to perform signal preprocessing, effectively removes low-frequency and high-frequency noise, and uses discrete Fourier transform and wavelet transform to perform frequency domain analysis and filtering processing to generate high-quality preprocessed spectral data.