73results about How to "Fine granularity" patented technology

The invention relates to the technical field of multi-agent systems, and discloses a multi-agent and multi-mode alignment interaction method and system for a review scene, and the method comprises the steps: mapping real-time multi-mode data into a local belief vector through a heterogeneous agent cluster, and calculating the spatial divergence to determine a global cognitive entropy; interaction time sequence gating logic based on entropy flow differential is established, and active alignment is triggered when the global cognitive entropy is larger than a convergence threshold value and a first-order derivative is smaller than a damping threshold value; according to the method, through entropy flow differential control and an active semantic disturbance closed loop, decision oscillation of the multi-agent system in the non-cooperative environment is eliminated, and the reliability of the multi-agent system in the non-cooperative environment is improved. Mathematical convergence of the belief state in an adversarial disguise scene is realized, and the system identification precision in a complex interactive game is improved.

The application discloses a kind of emotion evaluation and intervention method and system based on multi-modal data.Through terminal synchronous acquisition voice, face, screen tremor and so on multi-modal data;Reaction mode vector is generated based on context information and data verification fusion;The composite emotional state of user is determined using emotional emergence model.At the same time, personalized emotional causal diagram is constructed and updated, and the enhanced state vector is generated by extracting key causal path features.The vector is input into the intervention agent to output intervention actions;After execution, the emotional state is evaluated again and the reward is calculated, and the system model is updated online.The application realizes accurate emotion evaluation and closed-loop adaptive intervention that strikes the root.

The utility model discloses a vanadium removal device for a sodium aluminate solution, which belongs to the technical field of vanadium removal, and comprises a dissolving tank, a solution pump, a settling tank and a stirring component, the stirring component is arranged in the dissolving tank, the top of the dissolving tank is respectively connected with a feeding pipe for supplying the sodium aluminate solution and a feeding pipe for feeding dry vanadium, a bottom outlet of the dissolving tank is connected with a settling tank through a conveying pipe, and a solution pump is mounted on the conveying pipe. The device has the beneficial effects that a sodium aluminate solution to be treated is fed into the dissolving tank and can be stirred by the stirring assembly after being added with dry vanadium, and then the fully stirred solution can be fed into the settling tank under the action of the solution pump to crystallize and settle the vanadium, so that the vanadium is removed, and meanwhile, the quality of the sodium aluminate solution is improved. Dry vanadium can be added into the dissolving tank for dissolving and mixing, so that the particle size and quality of vanadium ions in the solution are improved, vanadium crystals can be better separated out by the settling tank, and the vanadium removal effect of the device on the sodium aluminate solution is further improved.

The application belongs to the technical field of expressway, and particularly relates to an expressway traffic congestion prediction method and system based on deep learning. The method comprises the following steps: acquiring traffic flow characteristic information, identifying a traffic bottleneck node based on the traffic flow characteristic information, summarizing reference characteristic values of the traffic bottleneck node to form an overall traffic flow characteristic data set, performing discrete degree analysis on the data set to obtain a characteristic index, extracting a traffic congestion concentrated interval according to the characteristic index, identifying a risk time point in the interval, and outputting a predicted expressway congestion period and a corresponding congestion area according to the risk time point and the traffic congestion concentrated interval. The application can improve the accuracy of congestion root identification by identifying a traffic bottleneck node and quantitatively analyzing the evolution trend of the bottleneck state, and can output a region to be regulated and a regulation period, thereby providing decision support for traffic relief.

This invention discloses a method and system for diagnosing and assessing teacher competence based on multimodal data fusion, applied in the field of data analysis. The method includes acquiring teaching video data of the target teacher and its corresponding lesson plan text data; then adaptively segmenting the lesson plan text into multiple text blocks, with the segmentation boundaries determined based on semantic density distribution; fusing corresponding sentiment enhancement vectors into each text block and extracting audio track prosodic features from the teaching video; establishing a precise correspondence between each text block and audio track segment using spatiotemporal alignment technology, and then calculating the first and second matching degrees; adjusting the contribution weight of semantic density to each matching degree to generate the final diagnostic assessment result of teacher teaching competence. The teacher competence diagnostic assessment method based on multimodal data fusion provided by this invention effectively improves the objectivity, granularity, and educational guidance value of teacher teaching competence assessment.

This invention relates to the field of chip inspection technology, and more particularly to a chip non-destructive testing system based on embedded current scanning. The system includes a current scanning module, an information acquisition module, a data analysis module, an abnormal scanning region generation module, and a scanning path generation module. Instead of simply performing a higher-density secondary scan of the entire chip, this invention intelligently filters out abnormal scanning regions requiring special attention based on the initial scan results and calculates their scanning priority. This allows valuable inspection time and computing resources to be concentrated on the most suspicious and severe areas, avoiding unnecessary repeated scanning in normal areas. The system dynamically selects the optimal scanning path based on the regularity of the region shape; for example, parallel lines are used for regular regions, and radial lines are used for irregular regions. This makes the scanning path more closely match the geometric features of the defect, obtaining a more complete region characterization with fewer scanning points, resulting in higher path planning efficiency.

The present application belongs to the technical field of sintering mixed material batching, and relates to a method for reducing particle size segregation of mixed material in a multi-stage flat laying manner. According to the required batching ratio, each kind of material is divided into 3-4 stages according to particle size, the first stage is for the material with the largest particle size and the most segregation, the second stage is for the dust removal ash with fine particle size and large dust, and the third and fourth stages are for the material with uniform particle size. Before taking the material, the head and tail cutting operation is implemented, and the mixed material in this area is returned to the top of the new pile. The present application adopts the segmented flat laying manner, solves the particle size segregation, adds the dust removal ash in the second stage, and solves the problem of dust diffusion. Each bin can be used alternately, solving the problem of the number of bins. At the same time, small materials can be added in a certain stage, realizing the recycling of iron-containing resources.

This disclosure provides a method and apparatus for modulation pattern recognition of orthogonal frequency division multiplexing (OFDM) signals based on adaptive modulation, which can be applied to the field of wireless communication technology. The method includes: acquiring an OFDM signal, which comprises N symbols and K subcarriers; performing a fast Fourier transform on the OFDM signal to obtain an NxK-dimensional frequency domain matrix; based on the subcarrier dimension and according to the identifiers of the K subcarriers, extracting multiple frequency domain symbol vectors associated with each subcarrier from the frequency domain matrix to obtain K target data sequences; performing at least one operation, such as clustering or accumulation, on any of the K target data sequences to obtain multiple modulation pattern recognition features associated with the target data sequences; and determining the target modulation pattern based on the multiple modulation pattern recognition features and a preset modulation pattern mapping relationship.

A multi-stage adsorption synergistic dust control system for silicon carbide laser processing includes a suction hood, a vibrating grid module, a porous laminar flow adsorption module, a boiling water mist module, a condensation pipe, a mixing condensation module, a drying module, and a fan, arranged sequentially and connected along the airflow direction. The vibrating grid module includes an alloy grid and a vibration device. The porous laminar flow adsorption module includes an adsorption layer disposed in the airflow channel. The boiling water mist module includes an adsorption chamber and a water mist generator to form boiling water mist to capture dust in the airflow. The condensation pipe performs preliminary condensation of the airflow to capture dust. The mixing condensation module includes an inclined condensation channel that utilizes dust in the airflow as condensation nuclei to remove dust. The inner wall of the condensation channel is alternately arranged with superhydrophilic and superhydrophobic regions. This invention can achieve stepwise dust removal from the micron to the nanometer scale, with a capture efficiency of up to 99.9%, low overall energy consumption, and recyclable materials.

This invention discloses a traffic sign recognition method and apparatus based on few-shot learning, comprising: collecting traffic sign data; training a preset target detection model based on the traffic sign image data to obtain a trained preset target detection model; training a preset few-shot image classification model based on the target detection annotation file and the auxiliary labels corresponding to each traffic sign to obtain a trained few-shot image classification model; identifying target bounding boxes of traffic signs contained in the image data to be identified using the trained preset target detection model, and identifying the specific category of the traffic signs contained in the image data to be identified using the trained few-shot image classification model based on the target bounding boxes. This invention can improve the efficiency and accuracy of few-shot traffic sign detection and recognition.

The invention relates to the technical field of traditional Chinese medicine preparations, and discloses a pear syrup-containing medicinal and edible wall-broken paste for treating heart-spleen deficiency, which is prepared from the following raw materials in parts by mass: 24 to 30 parts of compound medicine, 8 to 12 parts of lily, 100 to 120 parts of pear syrup, 12 to 18 parts of Chinese yam, 6 to 8 parts of semen euryales and 6 to 8 parts of semen coicis. Potential risks of non-edible raw materials are avoided, the pear syrup conforming to the national standard achieves dual effects of nourishing yin, moistening dryness and improving the taste, the industrial pain point that powder and a syrup body of a traditional syrup formula are not uniformly mixed is solved, and uniform distribution of components is ensured. A circulating cooling water and nitrogen double-control system is adopted in the whole process, the crushing temperature is strictly controlled to be 50 DEG C or below, and volatile effective components are reserved to the maximum extent.

This application discloses a method for preparing highly soluble boron citrate, comprising the following steps: S1. Preparing a borate solution and a citrate source solution separately; the borate solution also contains hydrochloric acid; S2. Under microwave heating reaction conditions, slowly adding one solution to the other solution, adjusting the pH of the solution to 5-6 after the first reaction time, and then continuing the reaction for a second time; S3. Heating and concentrating; S4. Spray drying to obtain the boron citrate product. The method provided in this application utilizes the direct reaction of soluble citrate source and soluble borate, which is a simple process. A stable boron citrate solution is prepared by adding hydrochloric acid and adjusting the pH of the solution. After heating, concentration, and spray drying, a highly soluble boron citrate powder product with fine particle size and low heavy metal content is obtained.

Provided is a laxative tablet which includes magnesium oxide as a main constituent, wherein (1) the particle size by volume of 50% (D50) of particles which emerge when the tablet is suspended in water is 70 µm or less as determined by laser diffractometry, (2) the particle size by volume of 90% (D90) of the particles which emerge when the tablet is suspended in water is 130 µm or less as determined by laser diffractometry, and (3) the dissolution time of the tablet when the tablet is suspended in water according to the Japanese Pharmacopoeia - general testing methods - dissolution testing methods is 10 seconds or less.

The application provides an entity recognition method, device and equipment and a computer readable storage medium; is applied to video processing and text processing and other application scenarios; the entity recognition method comprises: obtaining a to-be-processed text corresponding to a target account; performing feature extraction on the to-be-processed text to obtain a character feature corresponding to each character; obtaining each candidate entity string in the to-be-processed text; based on the character feature, identifying each candidate entity string to obtain an entity sentiment category corresponding to each entity string, wherein the entity sentiment category is a category to which a sentiment of the target account for each entity string belongs, and the entity string is the candidate entity string belonging to an entity. Through the application, the entity recognition effect can be improved.

The invention discloses a non-magnetic hard alloy and a preparation method and application thereof.The non-magnetic hard alloy comprises a high-entropy binding phase and a hard phase, the component of the hard phase comprises WC, the component of the high-entropy binding phase comprises at least six of first transition metal elements and second transition metal elements, the first transition metal elements comprise Co, Ni and Fe, and the second transition metal elements comprise Co, Ni and Fe; the second transition metal elements comprise at least three or more of Cr, Re, Ru, Rh, Pd, Os, Ir and Pt, the mass percent of the high-entropy binding phase is 6-25%, and the balance is the hard phase, the hard alloy has very high comprehensive mechanical performance, and the problem that an existing non-magnetic hard alloy is insufficient in wear resistance is effectively solved.

The invention discloses a store stock quantity suggestion generation method and device, equipment and a storage medium, and is used for solving the technical problems of hysteresis quality, coarse granularity and dependence on subjective experience of an existing stock guidance method. The method comprises the following steps: collecting check-in associated data of all target hotels in each city; calculating price fluctuation indexes of different areas of each city in preset holidays and festivals according to the check-in associated data; and according to the price fluctuation index, generating a store stock quantity guidance suggestion.

The present application relates to the technical field of urban intelligent traffic control, and discloses a method for calculating the three-level utilization rate of reserved bus lanes and identifying space-time resources, aiming to solve the problems of rough resource quantification, fuzzy identification of residual space-time resources, and lack of theoretical basis for reservation of road rights. The steps of the present application are as follows: S1, based on the virtual occupation space of buses, discretize the bus lane into three levels of space, i.e. grid, road section and route, and construct a digital road network; S2, fuse bus GPS trajectory, signal timing and traffic flow data to analyze the space-time law of vehicle operation; S3, calculate the utilization rate of bus lanes step by step; S4, identify the residual space-time resources according to the utilization rate to provide support for dynamic allocation of reservation road rights. The present application realizes fine quantification and dynamic identification of lane resources, guarantees bus priority while improving lane utilization rate, provides data support for reservation of road rights, and is suitable for intelligent control and road right optimization of urban reserved bus priority lanes.

The application belongs to the technical field of inorganic material preparation, and particularly relates to a twice calcined tungsten oxide and a preparation method and application thereof, the preparation method comprising: primary calcination: APT raw materials are added into an industrial rotary kiln with an inclination angle of 5-6 DEG at a feeding speed of 100-500 g / min for calcination to obtain primary calcined tungsten oxide; secondary calcination: the primary calcined tungsten oxide is added into an industrial rotary kiln with an inclination angle of 2-3 DEG at a feeding speed of 800-1000 g / min for calcination to obtain secondary calcined tungsten oxide. The preparation method uses simple equipment, and raw materials are easy to obtain, and the method is easy to be industrialized. By controlling the feeding rate, the inclination angle of the rotary kiln, the rotating speed and the calcination temperature, the obtained tungsten oxide has fine particle size, narrow bulk specific gravity and stable screening performance, and can be applied in ultra-fine crystal rods, petroleum catalysts and photocatalyst materials.

The application discloses a method for supporting multi-terminal touch screen information collection and restoration. A data collection terminal collects touch screen data on different signing terminals. A handwriting data conversion module in a data processing layer pre-processes the touch screen data obtained from the different signing terminals, converts multi-terminal handwriting data into standard handwriting data packets in a screen coordinate system, determines key frames of the touch screen data, fills in the blank between adjacent point positions of the touch screen data by using a tweening animation algorithm to ensure that there is no broken line, calls a Bezier curve algorithm to determine the intermediate frame inserted between the two key frames, so that the handwriting is smooth, and the obtained touch screen data is converted into lines close to the natural handwriting habit of a person. A rendering layer determines the size of a handwriting point according to the coordinates of a current point position on a canvas, real-time pressure data and a predetermined thickness, renders a handwriting track, and generates an electronic signature.

The application discloses a non-magnetic hard alloy and a preparation method and application thereof. The non-magnetic hard alloy comprises a high-entropy binder phase and a hard phase. Components of the hard phase comprise WC. Components of the high-entropy binder phase comprise at least six elements of a first transition metal element and a second transition metal element. The first transition metal element comprises Co, Ni and Fe. The second transition metal element comprises at least three or more than three of Cr, Re, Ru, Rh, Pd, Os, Ir and Pt. The mass percentage of the high-entropy binder phase is 6-25%, and the rest is the hard phase. The hard alloy has high comprehensive mechanical properties, and effectively solves the problem of insufficient wear resistance of the existing non-magnetic hard alloy.

The invention provides a pumped storage power station 5G + BIM intelligent interaction management system, and the system comprises a sensing layer which is used for collecting earthwork vehicle transportation data, dam rolling data, underground cavern construction data and monitoring data, and achieves the dynamic data updating of a BIM model through a 5G camera, a sensor and construction data; the network layer transmits dynamic data of the BIM model through a 5G private network, and allocates communication resources according to service requirements by adopting a network slicing strategy; the data layer is used for fusing dynamic data of the BIM model, including BIM geometric data, sensor real-time parameters and construction data, and constructing a dynamic digital twinborn database; and the application layer performs construction state early warning and construction scheduling based on design parameters and sensor parameters and a BIM model, generates a decision instruction through a deep learning algorithm, and realizes data deep fusion, network resource optimization and emergency capability improvement by fusing 5G high-reliability communication, BIM dynamic modeling and AI intelligent decision.

This invention provides diagnostic assistance methods, health assistance methods, diagnostic assistance devices, and health assistance devices, enabling high-granularity judgment of sensor data. The diagnostic assistance methods include: an acquisition step, which acquires sensor data detected by the sensor and lifestyle information related to the user's lifestyle habits, and records them in a time series; an extraction step, which extracts selected lifestyle habit association information associated with the sensor data from the lifestyle habit information; and an output step, which combines the selected lifestyle habit association information with the sensor data.

The present application relates to the field of test questions recommendation, in particular to a mathematics test question recommendation method and system based on learning online teaching evaluation. The method comprises the following steps: collecting student historical behaviors based on an online learning platform, performing data cleaning, and generating learning behavior sequences; identifying answered error test questions based on the learning behavior sequences, classifying and labeling, and generating a set of error question labels; performing error question label sorting analysis and error question similarity identification on the set of error question labels to obtain error cause weak points; performing correlation analysis based on the error cause weak points to generate a weak point correlation network; performing adaptive test question recommendation and progressive difficulty planning based on student ability evaluation portraits, and outputting personalized test question recommendation results. The present application realizes the structured quantification of user online learning state and the accurate matching of adaptive test questions, significantly improving the teaching evaluation accuracy and the efficiency and quality of test question recommendation.

The invention relates to the technical field of integrity verification, in particular to a video and audio file migration integrity verification method, which comprises the following steps: acquiring a synchronous frame extraction pixel sequence, establishing mapping to generate feature data, extracting time, marking a stable section, extracting brightness residual error, judging the direction to be consistent, and analyzing chromaticity variance to generate the stable section. According to the method and the device, the brightness and chroma pixel value sequences are extracted from the synchronous frames in the program packaging file and the broadcast rendering frame, and the mapping relation is established, so that frame-by-frame accurate comparison can be realized, the synchronous state is identified in combination with the time sequence difference trend, the inter-frame comparison precision is improved, and the accuracy of the inter-frame comparison is improved. Through the conjoint analysis of the brightness channel residual direction and the chrominance channel variance distribution, the image anomaly recognition capability is enhanced, and the frame-level consistency result is extracted based on the multi-dimensional residual feature intersection, so that the video and audio integrity verification without depending on a unified decoder is realized, and the real-time performance and fine granularity of quality assurance are enhanced.

The application provides a visual analysis method supporting fine subtype identification of multiple types of immune cells, and the method comprises the following steps: constructing a multi-scale visual ViT model comprising an encoder and a full connection layer; the encoder adopts a multi-scale visual network; blood cell images of each scale are input into the encoder after being divided into blocks, so that image features of each image block are obtained; the image features belonging to the same blood cell image are fused, so that unified image embeddings of each blood cell image are obtained; the unified image embeddings of each blood cell image belonging to the same blood cell image are fused, so that a final feature of the blood cell image is obtained; the final feature is input into the full connection layer for cell type classification, so that prediction probabilities of the blood cell image belonging to each cell category are obtained.

The application discloses a power grid material fault prediction and active defense system and method based on digital twinning and a multi-agent large model, and belongs to the technical field of power system disaster prevention. The system comprises a knowledge graph construction module, a multi-agent collaborative prediction module and a digital twinning simulation pre-play module. By constructing a "disaster-equipment" knowledge graph integrating multi-source data, a structured knowledge base is provided for the system; by using three-level agents of risk perception, probability prediction and strategy generation to work collaboratively, closed-loop decision-making from typhoon warning analysis to executable defense strategy generation is realized; finally, the scheduling scheme is dynamically simulated, deduced and optimized and verified in a digital twinning environment. The application solves the problems of difficult data fusion, disconnection between prediction and decision-making and lack of verification mechanism in traditional methods, and significantly improves the fault prediction capability, emergency response speed and active defense reliability of the power grid in typhoon weather.

The invention relates to the technical field of computer data processing and natural language processing, in particular to a case clue identification and analysis system, which comprises a data access and positioning module for calling a preset multi-source map service interface to concurrently analyze a network address, and determining a target position through voting arbitration; the cascade semantic analysis module is used for identifying attributes by utilizing coarse granularity of a first-level classification model, and routing to a vertical expert agent to generate a fine-granularity label; the element feature extraction module is used for eliminating interference digital features and then generating standard amount features, and analyzing intention intensity of the user; the value quantification and pricing module is used for mapping the heterogeneous features into multi-dimensional scores based on a tree-shaped weight system, and generating clue commercial value scores and pricing results; according to the invention, the problem of value evaluation caused by information non-standardization in a compliance consultation scene is solved.

This invention discloses a gold extraction process and system using whole-sludge cyanidation without pre-concentration, relating to the field of gold beneficiation and smelting technology. It solves the problems of low grinding efficiency, complex processes, difficulty in recovering coarse gold particles, and high safety risks in existing processes. The process includes raw ore crushing, primary grinding, gravity separation, hydrocyclone classification, secondary grinding-leaching, leaching-adsorption, smelting, and tailings treatment; the system includes corresponding crushing, primary grinding, gravity separation, hydrocyclone classification, secondary grinding-leaching, leaching-adsorption, smelting, and tailings treatment units. The key innovation is: gravity separation recovers coarse gold particles after primary grinding; the secondary stage uses an azin mill to achieve synergistic grinding-leaching, eliminating the pre-concentration stage; and the filtrate reagents are recycled. This invention features a simplified process, low energy consumption, high gold recovery rate, and safe and environmentally friendly operation, suitable for gold extraction from gold ores containing coarse gold particles that require fine grinding.

This invention provides a production apparatus for coarse-crystalline sodium metabisulfite, comprising a primary reactor, a secondary reactor, a tertiary reactor, and a sodium carbonate liquid tank connected in series. The primary reactor has a sulfur dioxide inlet on its side wall. A flow pump is installed on the connecting pipe between the secondary reactor and the primary reactor. A gas flow meter is installed at the sulfur dioxide inlet. A pH probe is installed inside the primary reactor. A crystallizer is connected to the bottom of the primary reactor via a conveying pipe equipped with a conveying pump. The crystallizer is installed higher than the primary reactor. An overflow pipe connected to the primary reactor is located at the top of the crystallizer, and a discharge pipe with a discharge valve is connected to the bottom of the crystallizer. The coarse-crystalline sodium metabisulfite production apparatus provided by this invention can increase the particle size of sodium metabisulfite, improve the main content of the product, enhance product quality, and extend the product's agglomeration cycle.

This invention discloses a multimodal named entity recognition method and apparatus, relating to the field of named entity recognition technology. The method includes: inputting the text to be recognized and the image to be recognized into a trained multimodal named entity recognition model to determine the entity recognition result. The multimodal named entity recognition model includes a text extraction module, an image extraction module, a sentence-level soft gating module, a self-attention Transformer module, a cross-modal Transformer module, an MLP expert head, a visual gating module, and a CRF decoder. Based on this scheme, information loss in cross-modal interactions is reduced, and fine-grained alignment and decoding robustness are improved, thereby enhancing entity recognition performance.