25results about How to "Improve fusion efficiency" patented technology

This invention relates to a method for fusion of heterogeneous knowledge in the power industry, comprising: obtaining first triples containing entity relations from the ontology layers of at least two knowledge graphs; classifying the entity relations in the first triples into synonym relations and inclusion relations according to relation aggregation prompts by a language model; merging entity relations belonging to synonym relations; obtaining second triples containing entity attributes or relation attributes from the ontology layers of at least two knowledge graphs; classifying the attributes in the second triples into synonym attributes and non-synonymous attributes according to attribute aggregation prompts by a language model; merging entity attributes or relation attributes belonging to synonym attributes; obtaining entities from the instance layers of at least two knowledge graphs and dividing them into several data blocks; determining whether there are duplicate entities according to entity alignment prompts by a language model; and merging duplicate entities.

The application discloses a kind of industrial production safety early warning analysis methods for multi-modal data, belong to industrial production safety monitoring technical field.S1, numerical text, audio and video, time series multi-modal data in industrial production are preprocessed, and unified format multi-modal feature data are obtained;S2, the data obtained in S1 are executed time, space, semantic alignment processing, and the fusion multi-modal feature data of space-time semantic consistency are obtained;S3, the data obtained in S2 are integrated processing of neural symbolization data, and integrated feature vector is output;S4, multi-model collaborative early warning analysis is carried out based on integrated feature vector, and the final early warning conclusion is generated;S5, four-level early warning grading response operation is executed;S6, dynamic optimization is carried out by comparing early warning result and actual disposal effect, and model parameters and strategy are updated.The above method is used, multi-modal data are fused, early warning is accurate, response is fast, adaptability is strong, can be dynamically optimized, and industrial production safety is comprehensively guaranteed.

The utility model discloses a micro-thread implant, which relates to the technical field of implants and comprises a cylinder, a rubber sleeve is arranged at the lower end of the cylinder, straight threads are arranged on the outer side of the lower end of the cylinder, and an inner hexagonal groove is arranged on the top surface of the cylinder; according to the utility model, through the matching of the rubber sleeve and the cylinder, the cylinder can conveniently extrude the rubber sleeve, so that a hydroxyapatite material in the rubber sleeve is extruded out, and the fusion efficiency of the cylinder and new tissues is improved; through the cooperation of the rubber sleeve and the hydroxyapatite, gaps around the cylinder can be conveniently filled, and the fusion effect of the cylinder and new tissues is improved; and through the matching of the straight thread and the fine thread groove, the tooth body can be better mounted, and the tooth body mounting effect is improved.

The application discloses an environmental monitoring data processing method and system based on pollution discharge permission execution report, and belongs to the field of environmental monitoring data processing.The method comprises the following steps: obtaining multi-source environmental data flow for preprocessing and space-time alignment to generate a first data set; extracting a constraint condition field, and instantiating a dynamic quality control rule library based on the constraint condition field and working condition state information; adopting double-layer discrimination logic to process the first data set to generate a second data set after quality control, so as to distinguish between false abnormalities and real emission over-standard caused by equipment drift; taking effective data in the second data set after quality control as input, performing emission accounting operation, and outputting a third data set containing a compliance rate and a compliance margin rate; performing memory mapping corresponding to the third data set and an execution report template, and executing a three-level verification control flow to generate a correction suggestion table based on the verification result.The application reduces the misjudgment rate of abnormal data and improves the accuracy of quality control.

This invention provides a deep learning-based intelligent analysis method and system for multimodal heterogeneous data, relating to the field of multimodal data processing technology. The method includes: acquiring multimodal raw data during urban rail transit train operation; performing timestamp synchronization alignment and dimension normalization on the raw multimodal data to obtain a standardized multimodal data sequence; inputting the standardized multimodal data sequence into corresponding modality-specific feature extraction networks to obtain modality feature vector sets with quality assessment labels; inputting the modality feature vector sets with quality assessment labels into a signal-to-noise ratio adaptive cross-modal feature alignment network; and spatially projecting and mapping the data onto the train's key physical structures to locate three multidimensional sensing anchors: the lateral stiffness node of the bogie side beam, the center of the wheelset tread contact patch, and the traction motor rotor air gap monitoring point. This invention enables accurate identification and graded early warning of bogie faults, wheelset wear, and power supply anomalies.

The invention discloses a large language model adaptive quantitative transaction system based on parallel game and strategy evolution. The large language model adaptive quantitative transaction system comprises an acquisition module, an evaluation module and an execution module. The acquisition module preprocesses the multi-source heterogeneous financial data to generate a structured market state; the evaluation module generates candidate transaction strategies in parallel based on a large language model, and selects a target execution strategy through comprehensive score function evaluation in combination with a disturbance mechanism; the execution module is converted into a transaction instruction for execution after being verified by a preset risk control rule, target strategy parameters are updated by adopting a strategy gradient algorithm based on a transaction result reward signal, and meanwhile candidate strategies are evolved through an evolutionary algorithm and the evaluation step is cycled. According to the system, the strategy generation efficiency and diversity are improved, and the dynamic adaptability of transaction decisions is enhanced.

This invention discloses a method and apparatus for dynamic configuration of a fusion tree based on a virtual interface, belonging to the field of multi-sensor information integration. The method includes: Step 1, performing an uncertain mapping of sensor measurement parameters using a virtual sensor interface, where the sensor measurement parameters include target motion feature parameters and target recognition feature parameters; Step 2, setting fusion nodes after the virtual sensor interface mapping, and dynamically configuring the fusion tree using processor-run program code; Step 3, storing the dynamically constructed fusion tree in memory using a neighbor table, where each fusion tree is treated as a directed graph, and any fusion tree structure is described by a unique neighbor table; the neighbor table specifies the input sensor combination used by each fusion node and the order of these combinations, thus completing the fusion tree configuration based on the virtual sensor interface for different task and multi-situational requirements. This invention's configuration method is unaffected by changes in the sensor interface and has high fusion efficiency.

The application discloses a kind of ship intelligent video monitoring systems, belong to computer technical field.System includes: data acquisition unit, for obtaining the multiple modal data of ship;Data fusion unit is used to determine reference information, based on reference information in multiple modal data self-adaptation determines the target modal data to be fused, target modal data is fused, and multiple modal fusion data are obtained;Data processing unit is used to determine abnormal information based on multiple modal fusion data, pre-constructed ship safety rule base and abnormal detection model, and abnormal information includes ship safety equipment abnormal information, ship personnel action abnormal information, sensor abnormal information and ship driving trajectory abnormal information;Alarm unit is used to process abnormal information, and the abnormal level of abnormal information is obtained, and alarm information is determined using abnormal level and abnormal information.The application can improve the safety and standardization in the process of ship navigation.

This invention relates to the field of data processing technology, and in particular to a public safety risk assessment and early warning method that integrates multi-source heterogeneous data. The method includes: acquiring multi-source heterogeneous data of a target area; extracting multiple features from the multi-source heterogeneous data; and mapping the multiple features to corresponding nodes in a knowledge graph; optimizing knowledge graph reasoning using a graph attention network to obtain interpretable reasoning paths and risk assessments, including: generating a representation vector containing a global graph structure context for each entity in the knowledge graph based on the graph attention network; vectorizing the prior risk logic rules and constructing an energy function; performing probabilistic reasoning using the representation vector, the vectorized representation of the prior risk logic rules, and the energy function; and outputting the risk assessment results.

This invention provides an ultrasonic hydrogen mixing and dissolving device. The tank contains a closed space for hydrogen mixing and dissolving. An inlet pipe is connected to the top of the tank and feeds a measured amount of drinking water into the closed space. An inlet pipe is connected to the bottom of the tank and feeds a measured amount of hydrogen gas into the closed space. An outlet pipe is connected to the bottom of the tank and is equipped with an outlet valve, allowing high-concentration hydrogen-rich water to be discharged from the closed space after hydrogen mixing and dissolving. A circulation and fusion unit is located within the closed space and includes an ultrasonic mixer, configured to drive the drinking water and hydrogen gas to circulate and mix within the closed space, and to perform ultrasonic treatment during the circulation. A central control unit controls the circulation drive unit to drive the drinking water and hydrogen gas to circulate and mix within the closed space; and controls the ultrasonic treatment unit to perform ultrasonic treatment on the drinking water and hydrogen gas passing through the flow channel, effectively enhancing the mixing rate and uniformity of hydrogen gas in the drinking water.

The invention discloses a dynamic metadata updating management system based on a lake-warehouse integrated framework, and relates to the technical field of metadata processing. In order to solve the problems of insufficient collaborative linkage and poor coordination consistency of metadata between a data lake and a data warehouse under a lake-warehouse integrated architecture, the system comprises the following steps: firstly, receiving and managing heterogeneous data streams, capturing a dynamic change state of the metadata of the heterogeneous data streams by combining initial metadata in the data lake, and completing lake entry processing; and then writing the metadata entering the lake into a data lake, correlating metadata records after metadata registration to perform semantic analysis, synchronously performing data warehouse structure updating and metadata adaptation management, and then positioning the position change of the corresponding metadata in a metadata directory according to each unified metadata record generated after adaptation management, and finally, dynamically updating the metadata catalog according to a positioning result to realize full-process management and control of the metadata from change capture, in-lake adaptation to catalog synchronization, and ensuring consistent metadata states and real-time catalog update.

The utility model discloses formaldehyde disinfectant adding equipment for shrimp seed breeding, which comprises a culture pond, and the right side of the culture pond is fixedly connected with a box body. A conveying pump is matched with a suction plate, water on the left side in a culture pond is conveyed into a mixing bin, a metering pump is started, formaldehyde disinfectant in a storage bin is conveyed into the mixing bin through the metering pump, the metering pump stops working after conveying the formaldehyde disinfectant with the preset amount, a motor is started, the motor drives a stirring rod to rotate, and the stirring rod is driven by the motor to rotate. The stirring rod fully mixes water in the mixing bin with the formaldehyde disinfectant, the control valve is started, the control valve conveys the water mixed with the formaldehyde disinfectant to the right side of the interior of the culture pond, the process is repeated, the formaldehyde disinfectant can be evenly distributed in the culture pond, the advantage that quantitative and even adding can be achieved is achieved, the adding amount can be accurately controlled, and meanwhile the adding efficiency is improved. The shrimp seeds can be quickly and uniformly diffused into the whole water body, so that the shrimp seed generation efficiency is improved.

This application discloses an image fusion method, apparatus, storage medium, and electronic device. The method includes: performing image transformations in multiple directions on at least one image to be fused to obtain a low-frequency image group and a high-frequency image group corresponding to the at least one image, wherein the low-frequency image group includes low-frequency images corresponding to each image, and the high-frequency image group includes high-frequency images corresponding to each image; dividing the images in the high-frequency image group into blocks according to a preset size to obtain multiple image block pairs corresponding to the high-frequency image group; performing image block fusion on the image blocks within each of the multiple image block pairs to obtain multiple fused image blocks; stitching the multiple fused image blocks together to obtain a high-frequency fused image; performing image fusion on the images in the low-frequency image group to obtain a low-frequency fused image; and performing an inverse image transformation on the high-frequency fused image and the low-frequency fused image to obtain a fused image of at least one image.

The invention discloses a multi-service converged communication system and method based on soft switching, and relates to the technical field of communication. According to the system, by constructing a soft switch core network with a layered architecture, fused transmission and processing of multiple services such as voice, data, video and the Internet of Things are realized. The system comprises a service access layer, a soft switch core layer, a service support layer and an application service layer which work cooperatively through standardized interfaces, and the service expansibility and the resource utilization rate of the communication system are improved. Meanwhile, a multi-service priority scheduling algorithm and a dynamic bandwidth allocation method provided by the invention solve the QoS guarantee problem when multiple services are concurrently. Experimental data show that the transmission delay of the system in a multi-service fusion scene is reduced by 15%-20%, the bandwidth utilization rate is improved by more than 25%, and the system has remarkable technical advantages.

The invention relates to the technical field of operation and maintenance of new energy equipment, in particular to a fault processing method for new energy station equipment and related equipment, and the method comprises the steps: obtaining multi-source operation information of a station, and carrying out semantic recognition and standardization processing to form unified semantic expression; based on the expression, knowledge reasoning is carried out by using a pre-constructed knowledge model, and a fault causal relationship is derived; and generating a fault handling strategy in combination with a preset procedure and a constraint condition. And executing the strategy under the collaborative architecture of the edge side and the cloud side, and feeding back and updating the knowledge model according to an execution result to form a closed-loop optimization mechanism. The method realizes semantic unification of fault information, intelligent derivation of a causal relationship and dynamic generation of a disposal strategy, improves the fault response speed and disposal precision through edge-cloud cooperation, guarantees safe and stable operation of a new energy station, has a self-optimization capability, and adapts to complex working condition requirements.

The invention relates to the technical field of submarine cable inspection, in particular to a submarine cable inspection method and system based on an underwater inspection robot. According to the method, the physical characteristics of the metal properties of the cable are used as the fusion key, the corresponding visual image data pairs can be accurately found through consistency verification of the physical properties even if a plurality of interferents with similar geometrical shapes exist or errors exist in pose calculation, and the fusion reliability is remarkably improved. Meanwhile, the use of the acoustic trusted area mask provided by the embodiment provides powerful prior information for the search and verification of the visual image, avoids blind search in massive visual image data, and improves the efficiency of data fusion.

This invention relates to the field of medical assistance technology, specifically to a pneumonia medical assistance system based on an improved VMamba model. The system includes: dividing the lung image to be diagnosed into raster cells based on a raster cell partitioning algorithm; calculating the lesion representation coefficient of each raster cell within each management time period; calculating the lesion representation index of the lesion trend path within the management period based on the raster cells along the lesion trend path; determining whether the raster cells covered by the lesion trend path are lesion regions based on the lesion representation index; inputting the lung field feature map and semantic feature vector into the improved VMamba model after channel-level concatenation, and outputting the pneumonia classification diagnosis result; and linking the pneumonia classification diagnosis result with a pneumonia knowledge graph to generate a visual auxiliary diagnosis report, significantly improving the accuracy and stability of lung medical image processing and achieving full-process automation of intelligent image analysis.

The embodiment of the invention provides an image classification method and device, electronic equipment and a storage medium, and belongs to the technical field of image classification. The method comprises the following steps: determining single-mode image data as target image data, and calling a preset image classification model to perform high-frequency feature fusion on image high-frequency features of the target image data and a plurality of single-mode image data to obtain image high-frequency fusion features; calling a preset image classification model to perform low-frequency feature fusion on the image low-frequency features of the target image data and the multiple pieces of single-mode image data to obtain image low-frequency fusion features; calling a preset image classification model to perform multi-modal feature fusion on the image high-frequency fusion features and the image low-frequency fusion features of the plurality of single-modal image data to obtain multi-modal image fusion data; and calling a preset image classification model to perform image classification on the multi-modal image fusion data to obtain image classification result information. According to the embodiment of the invention, the accuracy of classifying the objects in the image can be improved.

The invention relates to the technical field of distributed computing, in particular to an intelligent iterative knowledge base management system and method for on-site technical services, and the system comprises an edge layer which is used for collecting multi-modal data of on-site equipment in the operation process; the data acquisition layer is used for collecting multi-modal data and preprocessing the multi-modal data to obtain preprocessed multi-modal data; the distributed storage layer is used for carrying out distributed storage and deduplication storage on the preprocessed multi-modal data; the intelligent analysis layer is used for processing the preprocessed multi-modal data and constructing a technical service iteration knowledge graph; performing causal reasoning on the technical service iteration knowledge graph, and updating a pre-constructed technical iteration tree; the application service layer is used for providing field technical services for technicians according to the technical service iteration knowledge graph and the technical iteration tree, and the problems that multi-modal data fusion is low in efficiency, technical iteration manual dependence is high, knowledge and operation collaboration is poor and the like can be solved.

The invention discloses a multi-modal data deep fusion method, system and device and a medium, and the method comprises the steps: obtaining the multi-source fault associated data of a power distribution network, and verifying and unifying space-time labels; constructing a space-time grid coding system, and forming an index structure containing time, space and data source identifiers; key features such as temperature change rate and vibration frequency are extracted, and initialization classification is carried out; feature matching of fault features and grid attributes is realized by dynamically adjusting grids and weights; and through normalizing multi-modal data, reconstructing spatial-temporal distribution, evaluating a fusion effect, and outputting a deep fusion result. According to the method, the deep data fusion effect is remarkably improved, powerful support is provided for fault prediction and health management of the power distribution network, and the reliability and operation efficiency of the power grid are remarkably improved.

The application discloses a multi-modal three-dimensional target detection method based on structural feature and semantic feature fusion, and belongs to the technical field of target detection.The application solves the problems of low precision, poor real-time performance and insufficient robustness of the existing method.The application constructs an explicit matching fusion mechanism of low-layer structure guided initialization, high-layer semantic optimization and time sequence expansion, projects 3D boundary boxes of a laser radar and a camera to multi-view image planes, calculates geometric similarity and category consistency constraints of 2D projection regions, and constructs an explicit matching graph in combination with low-layer structure information and high-layer semantic features.The application guides weighted aggregation of cross-modal target features with high confidence through a sparse matching graph, aligns historical frame targets to a current frame coordinate system to construct a space-time matching graph, aggregates historical information through expansion to a time sequence dimension, and realizes efficient and accurate three-dimensional target detection of the laser radar and the visual multi-view camera.The application method can be applied to multi-modal three-dimensional target detection.

The utility model discloses a chemical reaction kettle structure for generating nanofluid, which belongs to the technical field of reaction kettles and comprises a base, a heating resistance wire is arranged in an inner cavity of the base, a reaction kettle body is fixedly connected to the top of the base, a sealing box is fixedly connected to the bottom of an inner cavity of the reaction kettle body, and the sealing box is fixedly connected to the bottom of the inner cavity of the reaction kettle body. A motor is fixedly mounted at the top of the reaction kettle body, and the output end of the motor is fixedly connected with a main stirring rod. The motor drives the main stirring rod to rotate, the main stirring rod drives the main stirring blade and the main gear to rotate, the main stirring blade drives the scraper to rotate, the main gear drives the auxiliary gear to rotate, and the auxiliary gear drives the auxiliary stirring rod to rotate. The auxiliary stirring blades are driven by the auxiliary stirring rod to rotate, and in the rotating process of the auxiliary stirring blades, the stirring track can be changed, so that collision among the raw materials is more frequent, and the material fusion efficiency is further improved.

This application discloses a training-free multi-image fusion method, device, and medium based on a diffusion model, relating to the field of computer vision technology. The method includes: acquiring multiple source images to be fused; simultaneously inputting the multiple source images into a diffusion model; performing a preset diffusion step on each source image to obtain a noisy image of the same size as the source image; and simultaneously performing a denoising process on the noisy image corresponding to each diffusion step until the noisy image corresponding to the 0th diffusion step is determined as the fused image of the multiple source images. This application aims to solve a series of problems in existing technologies when performing multi-image fusion, such as poor fusion effect, long processing time, and low efficiency, achieving fast and efficient multi-image fusion.

Embodiments of the present application relate to the technical field of electronic map, and disclose a POI data fusion method, an electronic device and a computer readable storage medium. The POI data fusion method comprises: obtaining a first data set and a second data set to be fused, wherein the first data set and the second data set each comprise a plurality of POI data; searching for POI data in the first data set and the second data set located in a same geohash nine-square grid according to geohash encodings of the POI data in the first data set and the second data set; performing a Cartesian product on the POI data in the first data set and the second data set located in the same geohash nine-square grid to form different data groups; calculating a similarity of two POI data in each data group, and determining whether the similarity exceeds a threshold value. The POI data fusion method, the electronic device and the computer readable storage medium provided by the embodiments of the present application can improve the fusion efficiency of POI data.

The application discloses a brain TV listens to a kind of music generation method for fine-grained emotion recognition based on multi-modal fusion, comprising: collecting multi-modal physiological and audio-visual signals, and performing millisecond-level time alignment;Feature extraction is carried out on the electroencephalogram mode and audio-visual mode;Emotion features are fused through cross-modal attention, and emotion state output is calculated;Emotion-music mapping is constructed to generate initial music parameters;Personalized music strategy is generated using reinforcement learning, and music is generated according to the personalized music strategy;Play music and collect physiological feedback in real time, calculate emotion offset;Adjust the music parameters according to the emotion offset until the final personalized music generation strategy is determined and the music generation result is output.The application can continuously improve the effectiveness of the personalized music emotion regulation method in real use scenarios, so that the system becomes more and more suitable for the emotion characteristics and reaction patterns of specific users in long-time and multiple interactions, and finally realizes fine, stable and self-adaptive emotion regulation effect.