228results about How to "Efficient identification" patented technology

The invention discloses a key value storage system indexing method oriented to an NVM-NVMe SSD (Non-Volatile Memory-Non-Volatile Memory Express Solid State Disk) hybrid architecture, which comprises the following steps of: constructing a heterogeneous storage architecture taking cold and hot data perception as a core driving mechanism, coordinating and managing two types of storage media, namely a non-volatile memory NVM and a solid state disk NVMe SSD, and realizing efficient identification, layered writing and dynamic migration of cold and hot data. According to the method, the characteristics of low delay, durability and byte addressing of the NVM are utilized, the frequency and I / O overhead of Flush and Compaction operations are reduced in a data write-in path, meanwhile, the problem of mixed storage of cold and hot data is avoided, and therefore the overall performance and storage efficiency of a system are remarkably improved. In addition, by introducing an asynchronous migration module, the method can dynamically adapt to the change of data popularity along with time evolution, and effectively support high performance and high availability of the key value system in long-term operation.

The invention provides a system frequency response estimation method and system based on a fault position, and belongs to the technical field of new energy power grids, and the method comprises the steps: constructing a new energy multi-operation-state model; solving the node voltage of the state model after the fault based on the superposition principle and iterative calculation, establishing a mapping model of the fault position and the low-cross cluster, obtaining the initial value of the active power vacancy, and comprehensively considering the difference between the initial value of the active power vacancy and the active recovery rate between stations at different fault positions. Calculating the system-level total active power vacancy considering the station low-voltage crossing characteristic difference; and constructing a system frequency response model based on the frequency modulation capability difference caused by the fault position, and estimating the system frequency response by adopting the response model. According to the invention, quantitative description of active power vacancy and system equivalent frequency modulation capability loss after disturbance at different fault positions is realized, so that the accuracy of frequency response analysis is improved.

This invention discloses a multi-scale icing detection method and system for transmission line insulators under extreme weather conditions. The method includes: dividing a visible light image into pixel blocks and converting them into feature vectors with position encoding; inputting these vectors into an improved Swin Transformer integrating a Dynamic Deep Sample Generation (DDSGM) module; applying DDSGM to enhance features in a four-stage feature extraction phase; additionally fusing polarized light features in the first two stages; and outputting a multi-scale feature map. The multi-scale feature map is then input into a deformable attention unit containing DDSGM, undergoing secondary enhancement and adaptive sampling (fusing physical field information to correct for offset) to obtain enhanced features. A prediction head predicts the icing category, bounding box, and thickness; Hungarian matching is used to associate the target; and an early warning system outputs the results and a warning. This invention is the first to integrate DDSGM, optimize sampling, and combine a multi-physical quantity strategy to improve robustness to low-quality images and transparent ice layers, achieving accurate detection and early warning under extreme weather conditions.

This invention belongs to the field of impurity detection technology for ethanol solutions, and particularly relates to a method for detecting trace impurities in ethanol solutions based on near-infrared spectroscopy, comprising the following steps: (1) examining the standard deviation (RSD) value and calculating the required sample size N; (2) acquiring samples; (3) outlier removal and spectral preprocessing; (4) dividing the dataset; (5) using the difference spectrum method to screen feature signals; (6) establishing a "background silence" discrimination criterion; and (7) using an external prediction set for discrimination. The detection method model described in this invention has a wide range of applications. It establishes a model based on the background system, effectively identifying normal samples. For impurity samples or abnormal situations that are inconsistent with the near-infrared absorption characteristics of the background system, as long as they cause changes in the background system, they can be effectively detected. Compared with traditional PLS, PLSDA and other modeling methods, background silence discrimination only requires calculation using relevant statistical methods, resulting in a smaller overall computational load and more convenient data processing.

This application provides a water pressure test method and system for evaluating grouting effectiveness, relating to the field of mine water hazard prevention technology, to solve the problem of difficulty in effectively identifying weak areas in grouting. The water pressure test method for evaluating grouting effectiveness includes dividing the grouting hole into multiple isolated independent test sections; conducting a water pressure loading test on each independent test section; conducting a pressure recovery test on each independent test section; calculating segmented hydraulic parameters; obtaining grouting sealing evaluation indicators based on the segmented hydraulic parameters; and identifying weak grouting sealing areas based on the grouting sealing evaluation indicators. This method can more accurately identify weak grouting areas.

The invention is applicable to the technical field of clinical examination and auxiliary diagnosis, and provides an extreme behavior risk auxiliary diagnosis biomarker combination, a kit and a detection method, the biomarker combination is composed of IL-4, IL-12 and IFN gamma, and the extreme behavior risk can be objectively identified by detecting the expression level of the combination in peripheral blood of a subject. The matched kit comprises specific antibodies aiming at the three markers, a standard substance, a buffer solution and other reagents, and according to the standardized detection method, through peripheral blood sample treatment, incubation, fluorescent staining, flow detection and result analysis, precise evaluation is achieved in combination with a clear judgment threshold value. According to the method, the defect that suicide risk assessment depends on a subjective scale and doctor experience in the prior art is overcome, the operation is convenient, the result is objective, the combined detection AUC of the suicide risk assessment, the subjective scale and the doctor experience can reach 0.726, a reliable extreme behavior risk auxiliary diagnosis means is provided for clinic, and the method is particularly suitable for risk identification of depression patients and has important clinical application value and conversion prospect.

The invention relates to an HPMG-Transform-based false data defense method for a vehicle network interactive information physical system, and belongs to the field of information security of the vehicle network interactive information physical system. The method comprises the following steps: decomposing long-term trend and short-term fluctuation of measurement data by using an HP filter; a Transform model fused with multi-scale Gaussian prior is constructed, the extraction capability of data local features is enhanced, and high-precision state prediction is realized; the suspicious state quantity is preliminarily locked through consistency detection, and then concealment detection is carried out by combining KL divergence and a K neighbor weighting algorithm to identify false data; finally, the predicted value is used for replacing polluted data, and a state estimation result is corrected. The method effectively improves the data prediction precision and attack detection accuracy in a vehicle network interaction complex dynamic scene, has the characteristics of high detection efficiency, high practicability and the like, and can provide reliable data support for power grid dispatching.

This invention discloses a multi-source electromagnetic noise suppression method based on noise classification and deep learning, belonging to the field of geophysical electromagnetic exploration technology. The method includes: segmenting and preprocessing the original electromagnetic observation sequence; using an improved U-Net network with an encoder embedded in a Mamba time-series modeling module for low-frequency noise suppression; identifying strong noise types in the data segments using a ROCKET classifier; based on the classification results, calling a second improved U-Net network trained for the corresponding noise type for class-based targeted denoising; and finally, splicing the data segments to obtain complete, high-quality data. This invention, through a phased processing framework of "low-frequency pre-suppression—noise classification—class-based denoising," combined with the strong time-series modeling capabilities of the Mamba module and the efficient classification performance of ROCKET, significantly improves the suppression accuracy and signal fidelity for complex, multi-source electromagnetic noise, and is particularly suitable for processing ground and airborne electromagnetic exploration data.

The application discloses a kind of laser radar point cloud DTM reconstruction methods for geological disaster survey, to solve the problem of microrelief distortion and feature recognition accuracy caused by vegetation obstruction.The method comprises the following steps: data acquisition, obtaining laser radar point cloud data and auxiliary data such as orthophoto, slope vegetation data; data preprocessing, original data solving, point cloud denoising, retaining intensity and normal vector information, and exporting.las format; point cloud elevation flip constructs "inverted mold", extracts ground feature points through mirror transformation and rasterization, and generates flip DSM; DTM reconstruction and optimization, flip DSM restores true geographic reference, and high-precision DTM is obtained after optimization.The present application can effectively solve the problem of important microrelief distortion and feature recognition accuracy caused by vegetation obstruction when establishing terrain model by using the present point cloud classification technology.

The application discloses a cyanide-containing slag tailing pond groundwater environment risk early warning system, a groundwater monitoring well subsystem comprises a plurality of groundwater monitoring wells; a timing sampling control subsystem is used for controlling the pumping time, flow and frequency of a water pump in each groundwater monitoring well; a color reaction subsystem comprises a color developing device and is used for monitoring the color developing reaction of the groundwater sample; a color sensing subsystem comprises a light source and a color sensor, the light source is irradiated towards the color developing device, and the color sensor is used for collecting the emergent light signal of the color developing device after the color developing reaction; a central control platform comprises an early warning subsystem and an emergency disposal response subsystem, the early warning subsystem receives the color sensor signal and divides the risk grade according to the obtained signal; the emergency disposal response subsystem executes corresponding detection and repair responses according to the risk grade made by the early warning subsystem; the application realizes the rapid detection of the cyanide content in the groundwater around the cyanide-containing slag tailing pond and the feedback and early warning of the environmental pollution risk grade.

The invention relates to the technical field of Internet of Things communication, and particularly provides an intelligent analysis method for a power terminal protocol, which comprises the following steps of: inputting communication flow data into a detection model under the condition that a feature extraction model fails to extract semantic features of the communication flow data to be analyzed to obtain a corresponding active detection message; under the condition of security verification, sending the active detection message to a terminal device of communication flow data, obtaining physical response data, and comparing the physical response data with a preset physical semantic mapping relationship to determine a corresponding semantic feature; and taking the semantic features as analysis results of the communication flow data under the condition that verification of the semantic features is passed through a simulation operation mode, updating historical communication data, and carrying out intensified training on the detection model. The technical problem that effective analysis is often difficult to realize during semantic analysis due to the fact that a frame structure of a non-standard power terminal protocol is different from a frame structure of a standard terminal protocol in the prior art is solved.

The embodiment of the application discloses an industrial internet threat identification method and system, the method comprises the following steps: collecting multi-source data in the industrial internet scene in real time through a bypass deployed collection terminal; pre-processing and feature extraction are performed on the multi-source data to obtain first features and second features; the first features and the second features are fused to obtain fused features; the fused features are identified by using a threat identification engine to obtain a threat identification result. The application has the following advantages: 1. The single-dimensional detection limitation is broken, the coordinated attack can be effectively identified, and the false positive rate and the false negative rate are reduced; 2. The known threat and the unknown threat are fully covered, and the high safety requirement of the industrial system can be met; 3. The attack chain correlation analysis across domains and time is supported, the attack process can be completely restored, the attack influence range is clear, and accurate basis is provided for threat disposal; 4. The industrial production network is not invaded, the normal communication process is not disturbed, and the core requirement of the industrial scene, that is, low invasion and high availability, is met.

This application provides a method and apparatus for open-set skeleton action recognition based on outlier prototype learning. The method includes: constructing a neural network model; preprocessing human skeleton data by obtaining initial features through a multi-branch feature extraction network; processing the initial features using a classifier and a hypersphere feature mapper to obtain logical prediction values ​​for action categories and branch features; training the neural network model based on a training set; after the first iteration, optimizing the feature space of in-distribution samples using multi-class loss; synthesizing virtual outliers in the optimized feature space of in-distribution samples; optimizing the energy boundary by combining in-distribution samples and virtual outliers; after the second iteration, selecting the optimal model weights based on the combined performance of open-set recognition and closed-set classification on the validation set; and determining the action type and triggering the corresponding robot operation based on the energy score output by the trained neural network model and the action classification result.

This invention discloses a method and system for predicting RNA binding protein binding sites on RNA without targeting, relating to the field of bioinformatics. The method includes the following steps: acquiring raw RNA structure data; expanding and filtering the preprocessed data; training an RNA analysis model using the filtered data; processing the RNA to be analyzed using the trained RNA analysis model, calculating the probability that each nucleotide is single-stranded, and obtaining a structure score; identifying low-accessibility regions based on the difference between the structure score output by the RNA analysis model and the structure score calculated by the original method; and identifying potential RNA binding protein binding sites by comparing the signal difference with that of the original method in a single RNA structure detection experiment. This invention utilizes the STONE method to improve the accuracy of RNA structure analysis in low-accessibility regions, and uses the improved difference to infer RBP binding sites, thereby improving the accuracy of RBP binding site identification.

The application provides a bolt recognition method and system and an electronic device, and relates to the technical field of computer-aided engineering. The application determines the connection relationship features between the candidate sub-regions and adjacent virtual components by acquiring the geometric features of the candidate regions in the assembly model, combining the connection topological information of the candidate sub-regions in the assembly model, and inputting the connection relationship features into a multi-modal learning model for bolt recognition, so that efficient and accurate recognition of the bolt is realized, and the fault tolerance and recognition accuracy are higher in complex assembly structures.

The invention discloses background layer updating during coding. A method performed by an image processing apparatus for updating a background layer during encoding of a scene is provided. A scene is encoded based on classifying objects depicted in the scene as foreground or background. The background is divided into ranked background layers, wherein each background layer is associated with a respective depth model. The method comprises detecting (S102) a change in an image portion of a background layer. The method comprises calculating (S104) a difference between the image portion and a corresponding image portion of a background layer ranked behind the one background layer. The method comprises selecting (S106a) the background layer ranked behind the one background layer to represent the image portion when the difference is less than a threshold.

The invention provides a program identification method and device, a server and a computer readable storage medium. The method comprises the following steps: acquiring program running information corresponding to an initial unintercepted program sent by terminal equipment; according to the program running information, determining target feature information of an initial program which is not intercepted; and according to the target feature information, identifying whether the initial unintercepted program is a target program, and obtaining a target identification result. According to the program identification method provided by the invention, the program operation information which is sent by the terminal equipment and corresponds to the program which is not intercepted initially can be firstly obtained; afterwards, according to the obtained corresponding program operation information, relevant target feature information of the program which is not intercepted initially can be determined, and whether the program needs to be intercepted or not is identified according to the target feature information. Therefore, the terminal equipment can be helped to effectively identify the programs needing to be intercepted and the programs not needing to be intercepted.

This application discloses a method for training a recognition model, a recognition method, a computer device, and a computer-readable storage medium. The training method includes the following steps: controlling a single-photon lidar to emit photon pulses towards a target scene and receiving one-dimensional single-photon echoes reflected from the target scene, where the target scene includes at least one object to be recognized; constructing a training sample set based on the one-dimensional single-photon echoes; constructing an initial model and inputting the training sample set into the initial model for iterative training; validating the initial model after each iteration, marking the validated initial model as the recognition model, and outputting it. The application method involves inputting the echo signal to be recognized into the trained recognition model, and obtaining the recognition result based on the processing of the recognition model. The recognition result includes the pose and type of the object in the target scene. Therefore, this application can achieve high-precision recognition of long-distance targets.

The application discloses a kind of textile flaw detection methods based on machine vision, and the textile RGB image is obtained by image acquisition and preprocessing, and size adjustment, normalization operation is carried out, core detection process includes: feature extraction is carried out using ResNet50 network, and abnormal area recognition is realized in pixel level in combination with SPADE algorithm and k nearest neighbor, on this basis, using transfer learning strategy, specific textile flaw data set is fine-tuned and optimized, to build high-precision detection model, CFLOW-AD model is used for multi-scale feature extraction and the generation of abnormal score chart, accurately identify potential flaws, and integrated real-time feedback mechanism, system according to preliminary detection result, dynamically adjust detection parameter, and suspected area is re-detected, through TP, FP, ROC curve index carries out model evaluation, generates detailed flaw report, the method effectively improves the automation level and accuracy of textile flaw detection.

The invention provides a log analysis method, device and equipment based on a time sequence database and sequence modeling, and relates to the technical field of network security, the method comprises the following steps: obtaining original historical alarm logs from a plurality of heterogeneous security devices; analyzing and aggregating the attack index to obtain log feature data which takes the attack index as a main key and is aggregated according to a time window, and storing the log feature data into a time sequence database; extracting data of the same attack index in a continuous time window from a time sequence database to construct a spatial-temporal feature sequence of the data; inputting the spatial-temporal feature sequence into a sequence modeling model to obtain a sequence feature representation representing the comprehensive behavior pattern of the attack index; and inputting the sequence feature representation into a classification model, and outputting the threat category of the attack index. According to the method, the multi-source logs are efficiently aggregated through the time sequence database, and the time-space correlation characteristics of the attack indexes are deeply mined by using the sequence modeling technology, so that the recognition accuracy and efficiency of threat intelligence are improved.

The application provides application of dimethyl carbonate as a chronic obstructive pulmonary disease diagnostic marker and belongs to the technical field of biomedical detection. The application first finds that the detection rate of dimethyl carbonate in exhaled air is significantly related to chronic obstructive pulmonary disease, and the detection rate of dimethyl carbonate in exhaled air of a chronic obstructive pulmonary disease patient is significantly higher than that of a healthy person. The ROC curve analysis result shows that dimethyl carbonate has extremely high specificity as a biomarker for diagnosing chronic obstructive pulmonary disease, and therefore can be used as a detection target for the diagnosis of chronic obstructive pulmonary disease patients, and has good practical application value.

The invention provides an intelligent analysis system based on multi-source data fusion, and the system comprises a data preprocessing and model building module, an intelligent data fusion and event matching module, and an event monitoring and analysis module. Identifying and predicting key events in the cross-platform data; the intelligent data fusion and event matching module realizes accurate fusion and event matching of multi-source data through deep text analysis so as to reveal internal association among cross-platform data; and the event monitoring and analyzing module is used for screening important items from high-matching-degree events screened from the system. According to the intelligent analysis system based on multi-source data fusion, data from a plurality of information channels are deeply integrated, a natural language processing technology is applied, and efficient recognition of cross-channel common concerns and potential trends is achieved. According to the method, themes displayed in different information channels can be accurately captured, and dynamic changes can be perceived in real time by means of precise text analysis.

This invention relates to a method and system for cross-domain open set person re-identification based on unsupervised domain adaptation. The method includes: constructing a cross-domain feature adversarial alignment module including a feature extractor and a cross-domain discriminator; constructing an integrated binary classification bias module including multiple binary classification bias modules; alternately optimizing the cross-domain feature adversarial alignment module and the integrated binary classification bias module in an adversarial manner by combining multi-class cross-entropy loss, difference loss, and bias loss in the source and target domains; after the model converges, determining whether the person in the current test image belongs to a known object or an unknown object based on the re-identification result and bias score output by the model; this invention uses a joint training method to simultaneously perform cross-domain feature distribution adversarial alignment and unknown object identification, ensuring efficient re-identification of known objects while also identifying unknown objects with high accuracy.

The invention discloses a dynamic authentication network information acquisition security management and control system and a control method, and relates to the technical field of network information security. The method comprises the following steps: in response to an access request initiated by a user, silently collecting minimized multi-dimensional context information before authentication, inputting the collected multi-dimensional context information into a preset comprehensive risk calculation model, calculating to obtain a real-time risk score representing a current request risk level, and according to a preset threshold interval in which the real-time risk score is located, determining the risk level of the current request according to the real-time risk score. Dynamically deciding and triggering an authentication strategy matched with the risk level, executing the authentication strategy, granting corresponding access authority if the authentication is successful, continuously monitoring user behaviors in subsequent sessions, dynamically adjusting session authority or triggering re-authentication according to abnormal behaviors, and executing the authentication strategy if the authentication is successful. And collecting log data of the authentication process for updating the user behavior baseline and optimizing the comprehensive risk calculation model.

The embodiment of the invention discloses an incidence relation graph adjusting method and device, equipment, a medium and a program product. A specific embodiment of the method comprises the steps of obtaining a target association relation graph; utilizing a preset object identification rule model to generate first object identification information corresponding to each piece of equipment information in the target association relation graph; according to the first object identification information, determining whether to use a pre-trained object identification information generation model to identify the same natural object; in response to determination of use, generating second object identification information corresponding to each piece of equipment information by using an object identification information generation model; and according to the second object identification information, performing graph adjustment on the target association graph to obtain an adjusted association graph. The implementation mode is related to artificial intelligence, and the same natural object identification can be accurately and effectively carried out on each piece of equipment information in the target association relation graph so as to determine each piece of natural object information corresponding to the target association relation graph.

The application discloses a network rumor intelligent identification and rumor-busting announcement automatic generation method and system, and belongs to the technical field of artificial intelligence. The system comprises a data acquisition module, a rumor identification module, a rumor classification module and a rumor-busting announcement generation module. The data acquisition module acquires network rumors and rumor-busting information in real time and stores the information into a case library. The rumor identification module judges whether the information is a rumor according to text features. The rumor classification module classifies high-risk rumors and suspected rumors that are manually audited by adopting a TextCNN model and outputs the rumors to a plurality of preset labels. The rumor-busting announcement generation module is used for providing a targeted rumor-busting strategy according to rumor categories and automatically generating a rumor-busting announcement according to user input and the rumor-busting strategy. The application reduces the time and effort of manually writing announcements and significantly improves the authenticity and timeliness of rumor-busting.

The application provides a vehicle hibernation method, medium, vehicle and product, relates to the technical field of vehicle control, and the method comprises the following steps: generating a hibernation preparation instruction and sending the hibernation preparation instruction to an electronic control unit on the vehicle under the condition that the vehicle meets hibernation conditions; receiving confirmation information fed back by the electronic control unit based on the hibernation preparation instruction, performing state detection on the vehicle based on the confirmation information, and obtaining a detection result; wherein the confirmation information is used to indicate that the electronic control unit has completed a hibernation preparation task indicated by the hibernation preparation instruction; and if the detection result indicates that the vehicle is in a standby state, performing a whole-vehicle hibernation operation. The technical scheme of the application sends a hibernation preparation instruction to the electronic control unit first, so that the electronic control unit has sufficient time to complete the hibernation preparation task, effectively avoids the task execution risk caused by direct forced hibernation, improves the reliability and robustness of the vehicle hibernation process, and guarantees the integrity and safety of the vehicle function.

The application provides a kind of based on eye diagram radio frequency fingerprint identification method and system, it is related to wireless communication technical field.The method comprises: the demodulation processing of transmitter signal to be identified, the separation of carrier is realized, the Gaussian baseband waveform of transmitter signal is obtained;Calculate eye diagram UI, with eye diagram UI as period to the Gaussian baseband waveform is intercepted section, all eye diagram trajectory frames intercepted are superimposed to obtain eye diagram;Calculate the probability density of eye diagram;The BT value of GMSK signal is obtained by backstepping using the probability density of eye diagram;The difference between the measured signal eye diagram and modeling eye diagram under the condition of same BT value is obtained, the subtle difference between them is obtained, the statistical characteristics of eye diagram difference are further extracted, and the radio frequency fingerprint feature vector is constructed by combining BT value characteristics;Radio frequency fingerprint feature vector is input to support vector machine to do the identity authentication of equipment.The authentication method of the application can utilize the hardware subtle difference between different wireless devices of modulation mode GMSK to accurately identify the identity of the same.

The invention relates to a knowledge-enhanced big language model reasoning method based on experience knowledge. Comprising the following steps: constructing a knowledge base; obtaining reasoning problems and corresponding problem solving experience; extracting the problem solving experience into structured knowledge entries; a reasoning structure of the problem is extracted in the structure abstraction process, and experience is refined into a reusable solution, judgment and correction mode in the mode generalization process; retrieval of knowledge: for a new problem, related items are retrieved by using a reasoning structure of the new problem; the diversity reordering step then selects a final set of entries, providing a strategically diverse context for reasoning. The problem-based core reasoning structure adopts a diversity perception retrieval mechanism, so that the LLM reasoning accuracy is improved, and the LLM can learn and evolve from practice under the condition that re-training is not needed.