234 results about "Dynamic screening" patented technology

The invention discloses a health degree evaluation system and method based on a photovoltaic string, and belongs to the technical field of photovoltaic operation and maintenance intellectualization, and the method comprises the steps: firstly collecting SCADA operation data of a plurality of strings of a photovoltaic power station, including voltage, current, assembly temperature, environment irradiance and power factors; secondly, constructing a whole-station feature reference model, and generating a state vector fusing environment and load characteristics through a nonlinear projection algorithm; aiming at the target group string, extracting a health reference track, calculating a disturbance propagation factor, and identifying a health abnormal state; when the deviation degree exceeds a threshold value, dynamically screening heterogeneous reference group strings, constructing a health score distribution model, and further calculating a confidence score and a prediction residual error; if the score is low and the residual error is unstable, determining that the string is in a sub-health or hidden fault state, and generating an intervention instruction; the method has high accuracy, strong adaptability and good interpretability, and can be widely applied to intelligent diagnosis and refined operation and maintenance management of the photovoltaic power station.

The invention provides a dynamic screening medical image target tracking method and device based on visual Transform, and relates to the technical field of computer vision, and the method comprises the steps: standardizing near-infrared or visible light fundus video frames into uniform resolution, constructing a template-search frame pair, and then jointly mapping the two frames of images into a Token sequence; a dynamic local interaction module is embedded in front of each pruning layer of the whole network, a local context is captured by using depth separable convolution and point convolution, a dynamic convolution kernel generator is driven, and a neighborhood Token is adaptively weighted and aggregated. Next, the Token screening and the compression mechanism TSC are operated in the same pruning layer, only the Top-K key Token is reserved, the redundant Token is cut off, and the original index is recorded; the objective of the invention is to improve the positioning stability and reasoning efficiency of a focus area (such as an optic disc) in a complex operation video.

The invention relates to the technical field of chemical raw material processing, and discloses a petroleum coke particle size grading and batching system which comprises a raw material pretreatment module, a multi-stage dynamic screening module, a particle size real-time monitoring module, an intelligent batching decision module and a closed-loop control execution module. The classification definiteness of materials with different particle sizes is ensured; meanwhile, through a particle size monitoring feedback mechanism, the particle size distribution abnormal state in the screening process is recognized in real time, the material purity within the target particle size range is guaranteed, the particle size grading accuracy is improved, a closed-loop matching decision module is arranged, particle matching weight parameters are dynamically optimized in combination with the electrolytic aluminum process requirement, and the particle size grading accuracy is improved. The deviation degree between the ingredient proportion and the target performance can be judged in real time; the speed ratio of the conveying belt and the opening degree of the stock bin are automatically corrected through the executing mechanism, it is guaranteed that the mixing proportion of coarse particles, medium particles and fine particles is always in the optimal interval matched with electrolytic aluminum anode production, and the quality stability of anode carbon blocks is improved.

The invention provides an irrigation water demand prediction method based on meteorological soil crop multi-source feature fusion, and relates to the field of data prediction, and the method specifically comprises the steps: firstly collecting farmland region multi-source feature data, and constructing an original data set; then, a nonlinear disturbance relation between farmland characteristics is extracted through a high-order disturbance encoder module, and a structure energy tensor is introduced to describe high-order structure relevance; thirdly, a dynamic gating characteristic generator module is introduced, a disturbance gain factor and a time coupling factor are combined to generate a gating signal, and dynamic screening of multi-source characteristics is achieved; then constructing a cross-variable interaction modeling module, and generating a feature embedding vector to model a complex coupling relationship among weather, soil and crops; and finally, outputting an irrigation water demand predicted value by adopting a multi-layer sensing network in combination with channel mapping and a cross memory mechanism, so that efficient dynamic prediction of the farmland water demand is realized, and the water resource utilization efficiency and the accurate regulation and control capability are improved.

The invention provides a switch security protocol interaction method and device, equipment and a storage medium, and the method comprises the steps: carrying out the protocol type recognition of a received network data frame, and generating first intermediate data according to the recognized protocol type; according to the protocol type identifier of the first intermediate data, converting the format of the network data frame through a protocol mapping rule to generate second intermediate data; determining a security processing strategy corresponding to a preset strategy rule set according to the strategy identification field of the second intermediate data; reconstructing the original content of the second intermediate data into a target data frame conforming to a target protocol format according to a security processing strategy; and according to the security level field and the communication direction of the target data frame, performing path screening through the trusted path list, and determining a target interaction path. According to the scheme, the trusted interaction path can be dynamically screened based on the protocol type and the security level field of the target data frame, and controllable data transmission between switches under security protocol cooperation is realized.

The invention discloses an abnormal data monitoring method and device based on artificial intelligence, and the method comprises the steps: 1, dividing an original data stream through a sliding window, extracting statistics, time sequence and change rate features, and dynamically screening features adaptive to data distribution based on an SHAP value; 2, constructing a double-flow model, capturing a global isolated mode by adopting an improved isolated forest in a static flow, capturing time sequence dependence on the basis of LSTM-AE in a dynamic flow, and fusing two-flow scores through performance-driven dynamic weight distribution; 3, combining a density peak value algorithm with historical density attenuation weighting, and dynamically adjusting an abnormal threshold value; 4, realizing low-delay incremental learning through a double-trigger mechanism and experience playback; 5, multi-granularity interpretation is generated, manual annotation feedback is supported, feature engineering and model training are integrated, and a'detection-interpretation-feedback-optimization 'closed loop is formed; high-adaptability anomaly monitoring is realized through dynamic feature screening, double-flow fusion detection, threshold value self-adaption and man-machine collaborative optimization.

The invention relates to the technical field of three-dimensional scene information transmission, and discloses a digital information transmission method and system in a three-dimensional level loading process, and the method comprises the steps: obtaining three-dimensional level scene information, and dividing the three-dimensional level scene information into a plurality of scene region units; executing multi-factor priority evaluation; executing secondary priority evaluation to generate a resource task queue; adopting a multi-stage asynchronous streaming mechanism to complete digital resource transmission based on the loading granularity; and dynamically screening resources in combination with the GPU rendering budget. Compared with the prior art, the problem that loading efficiency and real-time performance cannot be considered in a complex three-dimensional scene is solved, and especially the technical problem that loading and rendering collaborative optimization is difficult to realize when a user view angle is quickly switched or the granularity difference of scene resources is large is solved. Due to the fact that space partitioning, dynamic priority evaluation and frame budget scheduling mechanisms are introduced, the problems of loading delay and resource redundant transmission are avoided, and the real-time loading efficiency and rendering fluency under the large-scale three-dimensional level scene are improved.

The invention discloses a visual question-answering system based on multi-modal knowledge autonomous learning and a construction method, and belongs to the technical field of artificial intelligence. The invention provides a visual question and answer model construction method fusing image information, natural language questions and external knowledge. A knowledge dynamic screening and replacement strategy is constructed by introducing an autonomous evolution mechanism, and whether knowledge is inserted or updated can be judged according to a current training state and knowledge similarity, so that the problem that a traditional static knowledge base cannot adapt to task changes is solved. According to the method, through pseudo data construction and multi-modal triple extraction, structured and high-association-degree'image-question-answer 'knowledge representation is established, so that knowledge has better semantic expression ability, and redundancy and noise are effectively reduced. According to the method, rapid retrieval and active calling of knowledge are achieved by means of the vector indexing technology, key support information can still be obtained in a complex scene, and the expression of the model in a visual question-answer task based on knowledge is remarkably improved.

The invention provides an NL2SQL database table dynamic screening method and system based on multi-strategy fusion. The method comprises the steps that a preliminary candidate set is generated, a semantic encoder is used for encoding a user natural language problem and table and column information in a database into vectors, the semantic similarity is calculated, a preset number of tables are screened out from a whole database table based on the similarity, and a preliminary candidate table set is formed; performing deep analysis and reordering, performing multi-dimensional analysis on the tables in the preliminary candidate table set, and calculating a graph relationship compactness score and a content correlation score of each table; and performing fusion decision, inputting the semantic similarity score, the graph relationship compactness score and the content correlation score of each table into a fusion model to obtain a final comprehensive score of each table, and screening out a final correlation table set based on the final comprehensive score. The method has high accuracy and high recall rate, the generation speed and response efficiency of the NL2SQL system are improved, meanwhile, the robustness is high, and the practicability and generalization ability of the system are improved.

The invention discloses a trajectory planning method based on a multi-stage optimization strategy and a mixed diffusion model. The method comprises the following steps: fusing multi-source sensor data and extracting features; screening and fusing intention anchor point tracks; and optimizing the fine-grained trajectory based on the Transform-Mamba mixed diffusion model. According to the method, an intention anchor point track screening mechanism is provided, the anchor point track matched with the scene is selected from the compact vocabulary through the dynamic screening module and fused with the static prior anchor points, the calculation complexity is reduced, the diversity of the initial track and the scene consistency are guaranteed, and the problem that a traditional fixed anchor point set is insufficient in flexibility is solved. According to the method, a mixed diffusion decoder is designed, space environment dependence is efficiently modeled through a cross attention mechanism, a bidirectional Mama module captures time sequence motion dependence with linear complexity, global perception ability and long sequence modeling efficiency are both considered, and safety and dynamics feasibility of generated tracks are ensured.

The invention belongs to the field of computational chemistry, and particularly relates to a multi-parameter collaborative dynamic screening method for adsorbed species on the surface of a catalyst. Comprising the following steps that an initial crystal structure is constructed and optimized through VESTA software, and a stable crystal structure model is obtained through density functional theory optimization; integrating working condition parameters such as a solvent, pH and potential, and drawing a Bayer diagram to label adsorption species and a mapping relation; and calculating a zero charge point and capacitance to obtain electrochemical parameters, and determining stable adsorption species under actual reaction through constant potential molecular dynamics simulation. According to the scheme, a deep structure-activity relationship between an active site structure and catalytic activity under a working condition is established through integration of all working condition parameters, labeling of adsorption species by a Boubye diagram, constant potential molecular dynamics simulation and the like; the problems that a systematic method combined with actual working condition parameters is lacked in prediction of the adsorbed substances on the surface of the catalyst, and real adsorbed species are difficult to accurately screen are solved, and accurate capture and efficient screening are achieved.

The invention discloses a scientific research theme evolution method based on text embedding vector clustering, and relates to a data processing technology, the method comprises the following steps: collecting scientific research literature data, preprocessing, identifying and extracting field keywords, generating text embedding vectors corresponding to literatures, and constructing a scientific research literature representation vector library; taking a preset time slice as a unit, performing density adaptive clustering on the text embedding vector, generating a scientific research subject cluster in the corresponding time slice, and dynamically extracting Top-N subject keywords of the scientific research subject cluster based on keyword weights and semantic vector center distances of literatures in the cluster; and constructing a dual-verification mechanism of vector space similarity and keyword semantic overlap ratio, dynamically screening a topic evolution relationship between scientific research topic clusters in adjacent time slices, introducing a keyword overlap ratio to filter associated noise, constructing a high-confidence topic evolution chain, and revealing development venation and technical migration rules in the field of scientific research.

The invention relates to the technical field of industrial electric control optimization, and discloses an intelligent industrial electric control optimization system based on a deep learning algorithm. The system comprises a multi-modal sensing module which integrates a temperature sensor, a flow sensor and a pressure sensor to collect multi-modal data flow; a feature extraction module extracts three types of features to construct a multi-modal industrial feature set; the state comparison module generates normal and abnormal operation state comparison characterization by means of a generative adversarial network; the contribution degree evaluation module takes the system load rate as a tool variable, and calculates the contribution degree score of each sensing channel to the system stability through a Bayesian network; the dynamic screening module fuses the two types of indexes, and adjusts threshold screening through a deep belief network to generate optimization target representation; the time sequence grouping module performs phase grouping and weighting according to an operation cycle; and the fusion decision module generates a regulation instruction through hierarchical attention mechanism fusion, so that the operation stability and efficiency of the industrial electronic control system are effectively improved.

The application discloses a computing power scheduling system and method of a computing power network, relates to the technical field of computing power distribution and scheduling, and comprises the following steps: acquiring all computing power nodes in the computing power network and constructing a computing power node interaction network model; acquiring a computing power task initiating node and computing power task attributes; determining a target attribute group cluster of the computing power task based on the computing power task attributes, and recording all computing power nodes in the target attribute group cluster as demand target nodes of the computing power task; determining an execution trust target node of the computing power task based on the computing power task initiating node; finding an intersection of the demand target nodes of the computing power task and the execution trust target node of the computing power task, and recording the intersection as a comprehensive target node; and determining an execution computing power node of the computing power task from the comprehensive target node in combination with balanced load. The application has the advantages that the interaction network model of the cooperative trust relationship and the attribute group cluster is constructed, multi-dimensional node dynamic screening and optimized matching are realized, and the efficiency and reliability of the computing power scheduling are significantly improved.

The invention relates to the technical field of construction risk early warning, in particular to a power grid construction risk early warning method based on multi-source data fusion, and the method comprises the steps: S1, obtaining transient electromagnetic field intensity and ground potential change data of a construction site, carrying out the synchronization processing, obtaining synchronized multi-source data, carrying out the credibility evaluation, and carrying out the early warning of the construction risk; forming a credible input data set with a credibility identifier; s2, according to the credibility identifier, performing dynamic screening or weighting processing on the multi-source data, constructing fusion features, inputting the fusion features into a risk analysis model, and generating a construction risk grade result; and S3, whether an early warning condition is met or not is judged according to the risk level result, when the early warning condition is met, a construction intervention instruction is generated, and the construction intervention instruction comprises switching of an anti-interference monitoring mode, adjustment of a construction process or pause of operation. According to the method, the robustness of multi-source data fusion and the accuracy of subsequent risk analysis in a complex electromagnetic interference scene are improved.

According to the intelligent auxiliary learning method and system based on the model context protocol and the cognitive state modeling, decoupling of a model end and a tool end is achieved through the model context protocol, an error type-tool dependency topological graph is constructed through the cognitive state modeling, tool dynamic screening and pruning based on deterministic rule constraints are achieved, and the method and the system have the advantages that the method and the system are easy to implement. The Token consumption is reduced, and the accuracy of model reasoning is improved at the same time. In the aspect of safety control, semantic firewall middleware is introduced into the system, output streams are monitored in real time, illegal behaviors directly giving code answers are intercepted through text semantic analysis and code abstract syntax tree comparison, and a model is forced to turn to a thought guide mode. The system also includes hierarchical context compression to maintain long term memory, adaptive difficulty knowledge retrieval based on user cognitive states, and a mechanism to utilize code sandbox to assist verification of model reasoning logic correctness. According to the method, the behaviors of the large language model can be effectively regulated and controlled, and safe, efficient and personalized intelligent auxiliary learning is realized.

The invention discloses a fine screening and impurity removing device for milled rice with embryo, which relates to the technical field of grain processing machinery, performs dynamic self-adaptive regulation and control on the screening state of the milled rice with embryo in the screening process by virtue of a flexible screen surface assembly and a low-amplitude vibration assembly, and can be effectively different from a passive screening mode of fixed amplitude of a traditional rigid screen surface. Firstly, the dynamic form of the catenary screen surface changes the curvature in a self-adaptive mode along with the weight of the materials, active diffusion and self-cleaning of the materials are achieved through elastic opening and closing of the screen seams, and secondly, damping buffering of the carrier roller set and low-amplitude high-frequency vibration parameter optimization are matched, so that the screening efficiency is improved. The stacking state and the impact stress requirement of materials can be dynamically met, and the dynamic screening requirement of the milled rice with embryo caused by the water content difference, the variety characteristics and the particle size deviation is accurately matched in this way, so that efficient separation of impurities can be achieved in the screening process; germ damage and high-humidity material blockage caused by rigid impact can be avoided.

The invention discloses a data error rate reduction method for a memory, and relates to the technical field of memorizers, the method comprises the following steps: classifying a single event causal chain into three types of error codes of a physical degradation type, a crosstalk type and an environmental interference type, and screening a core feature subset through frequency statistics and mutual information analysis; utilizing the sparse feature capsule network to predict the bit error rate, generating a pulse parameter dynamic model in combination with the two-dimensional emergency degree matrix, and adjusting the pulse amplitude and width; and delimiting an error code field through physical connection relevance, and implementing dual-stage repair. According to the method, accurate error code classification, dynamic feature screening and targeted repair are realized, and the reliability of the memory is improved.

The invention provides a low-altitude unmanned aerial vehicle visual target tracking method based on dynamic sparse Sinkhorn attention, and belongs to the technical field of unmanned aerial vehicle intelligent perception. According to the method, an unmanned aerial vehicle carries imaging equipment to obtain video images of a target template area and a search area, and visual and language tokens are respectively generated in combination with natural language description; a dynamic sparse Sinkhorn attention module is embedded behind a multi-head self-attention layer of a visual encoder, an approximate double-random transmission matrix is constructed through similarity calculation and iteration, and differential sorting between tokens is achieved; a previous visual token most related to target semantics is dynamically screened and reserved according to the correlation score, background interference and redundant information are inhibited, and airborne calculation burden is reduced; and fusing the screened tokens in a prediction network, and outputting a target existence probability and a bounding box to realize stable real-time tracking. According to the method, the calculation complexity is reduced on the premise of ensuring the precision, and a low-power-consumption embedded platform is adapted; the robustness in a complex low-altitude environment is enhanced; and the cross-domain generalization ability is improved.

The invention relates to a time sequence electric energy data-oriented comparative distribution matching dataset distillation method, which is applied to the field of data processing, and comprises the following steps: obtaining a to-be-distilled multi-dimensional time sequence electric energy dataset, and in a preset comparative learning framework, respectively carrying out data enhancement and random sampling on the to-be-distilled multi-dimensional time sequence electric energy dataset to obtain a to-be-distilled multi-dimensional time sequence electric energy data set; obtaining to-be-distilled multi-dimensional time sequence electric energy positive and negative comparison data; and a dynamic adjustment temperature coefficient is set in the contrast learning framework. And extracting electric energy time sequence perception characteristics in the to-be-distilled multi-dimensional time sequence electric energy positive and negative comparison data. And according to a preset dynamic screening strategy and a feature enhancement strategy, determining a target time sequence electric energy data set containing the target abnormal time sequence feature in the to-be-distilled multi-dimensional time sequence electric energy data set based on the electric energy time sequence sensing feature. And performing data set distillation on the target time sequence electric energy data set according to a preset distribution matching strategy. According to the method, the problem that the analysis effect of a traditional distillation method on time sequence electric energy data is poor is solved.

The invention provides a grain multi-group decoupling prediction method, a modeling method, equipment and a storage medium. The modeling method comprises the following steps: providing a training database; extracting a feature library; training a secondary loop model, and outputting a first component prediction result; training a main loop model; and iteratively training the secondary loop model and the main loop model to obtain a grain multi-group decoupling prediction model. The method not only corrects a spectral reflection value, but also constructs a closed-loop prediction mechanism based on a first component state, dynamically screens wavebands, establishes an internal and external double-closed-loop cascade architecture of a main loop and an auxiliary loop, continuously adjusts parameters and updates a feature library, thereby realizing systematic suppression of interference signals and enhancement of intrinsic features of a second component. An interpretable feature screening mechanism is introduced, the importance of each wave band is evaluated in real time, the combination of a data-driven model and spectral physical prior is realized, and the problem that the generalization performance is reduced due to the fact that the model depends on shortcut features in the training process is avoided.

The invention discloses a water treatment optimization system based on multispectral technology combination and a regulation and control method, and belongs to the technical field of software. The invention aims to solve the problems of serious redundancy, noise amplification or information idleness caused by lack of a systematic decision-making process in state vector construction in a multispectral combination scene in the prior art. Therefore, the invention provides a water treatment optimization system and a regulation and control method, and the method comprises the steps: constructing a candidate state set containing multispectral analysis indexes and operation parameters; establishing a prediction contribution degree, signal stability and control sensitivity three-dimensional evaluation system; key state quantities are dynamically selected through a layered screening strategy; and generating a refined and information-rich target state vector for a control decision. Compared with the prior art, the method has the advantages that refining and information enrichment of state vectors are realized through dynamic screening and derivation construction mechanisms, the control precision and the calculation efficiency are remarkably improved, and the noise robustness is enhanced.

The invention discloses a transformer substation video fusion monitoring method and system based on 3DGS, and relates to the technical field of three-dimensional modeling. The method comprises the following steps: acquiring a laser radar point cloud and a multi-view camera image of a transformer substation, and constructing a three-dimensional Gaussian splashing model; establishing a mapping relation library based on P source view angle and Q target view angle transformer substation 3DGS samples; based on the library, using a picture splicing boundary confidence degree adjusting pointer and a space coordinate mapping deviation correction pointer to carry out fusion training; and when a user clicks a target device in the model, calling an improved line-of-sight cone analysis algorithm to dynamically screen an optimal camera, receiving a video stream of the optimal camera, and determining a device state recognition result. The technical problem of poor accuracy of image splicing and equipment positioning caused by multi-view and view range limitation in substation equipment monitoring is solved, and the technical effect of improving the monitoring precision and efficiency of substation equipment by dynamically screening the optimal camera and equipment state recognition of space coordinates is achieved.

The invention provides a Python bytecode obfuscation method and system based on dynamic screening and random replacement, and the method comprises the steps: calculating a plurality of safety indexes and a comprehensive safety index SDI of bytecode files, and screening out the bytecode files which have high safety requirements and need to be obfuscated; a linear congruence generator is constructed, a pseudo-random number sequence is provided for obfuscation operation of byte codes, and operation code dynamic replacement obfuscation operation is performed in byte code instructions by generating random seeds, constructing random dynamic operation code mapping tables and applying the independent operation code mapping tables to different byte code files needing to be obfuscated; and performing obfuscation state detection on the obfuscated bytecode file, if the detection result is the obfuscated bytecode file, modifying bytecode analysis and instruction scheduling logic, increasing a runtime de-obfuscation process, embedding an Opcode reflection mechanism, re-compiling to generate a customized Python interpreter supporting the execution of the obfuscated bytecode, and executing the obfuscated bytecode file. According to the invention, byte codes can be confused.

The invention relates to the technical field of event-driven control, in particular to an event-driven control method for a rescue robot, which comprises the following steps of: extracting event triggering conditions, triggering time sequence data and state parameters based on identification and serial number management of various events in a rescue environment; event type classification, spatial region partitioning and triggering probability modeling are carried out in combination with an SPFA algorithm, the corresponding relation between an event and a response state can be accurately mapped in the space-time dimension, the structural expression and state scheduling capability of a complex environment is enhanced, path optimization is carried out in combination with weights, and the path optimization efficiency is improved. Quantitative allocation of response resources and quantitative evaluation of response execution efficiency can be realized, a response path cost model is constructed by adopting a Dijkstra algorithm, and a mapping relation between path suitability and a triggering probability is compared, so that dynamic screening of an optimal path can be realized under an environment constraint condition; and the execution accuracy of the response path and the task completion collaboration are improved.

The invention discloses an enzyme activity optimization and dynamic screening method based on collaboration of multiple computing servers, belongs to the field of computational biology and enzyme engineering, and realizes collaborative design of multiple platforms by integrating technical advantages of online and local computing servers such as FireProt 2.0, PROSS, Rosetta and FoldX 4.0. Mutation design is carried out from multiple dimensions of structure prediction, evolution conservative property, energy calculation and the like, and a potential mutant library is constructed; a multi-condition screening strategy is used, and a mutant with high potential is screened out through visual inspection, conservative analysis and multi-method cross validation, so that invalid design is avoided; molecular dynamics simulation and MM / GBSA binding free energy decomposition are further combined for dynamic binding analysis, the influence of mutation on the substrate binding strength is quantified, and the stability and reliability of mutation effect prediction are remarkably improved.

The invention discloses a power system power flow self-adaptive regulation and control method based on deep transfer learning, and the method comprises the steps: 1, constructing a power system source domain power flow analysis model, and constructing a loss function for a power flow error in combination with the power flow data characteristics of a source domain and a target domain; step 2, designing a network parameter initialization method suitable for power flow analysis, realizing transferable representation from source domain data to a target domain, and constructing an initialized target domain power flow analysis model; 3, introducing a sample selection and incremental learning mechanism, and carrying out dynamic screening and retraining on key power flow samples in a target domain; and 4, carrying out load flow calculation and optimization control on the target power system by adopting the trained target domain load flow analysis model, and realizing self-adaptive optimization of the system operation state through iterative correction and parameter updating. According to the method, rapid calculation and dynamic optimization of the power flow of the power system are realized, and the intelligent level, stability and calculation reliability of power grid operation analysis are remarkably improved.

The invention relates to the technical field of computing power resource allocation, and discloses an AI computing power resource dynamic allocation method and system, and the method comprises the steps: analyzing a resource state according to pre-obtained computing power resource state data, recognizing discretely distributed fragment resources, calculating a corresponding resource value, and constructing a fragment resource map; in response to a preset calculation task, configuring a fragment combination mechanism to obtain a fragment resource block; configuring a resource screening mechanism, dynamically screening out computing power resources which can meet the computing task after being combined with the fragment resource blocks, and obtaining a resource allocation scheme; scheduling the computing power resources based on the resource allocation scheme; by effectively utilizing the fragment resources, the overall resource utilization rate of the computing power resources can be increased, the task response speed can be increased, the fragment resources are actively combined to complete the corresponding computing tasks, and the task queuing and execution time is shortened, so that the dynamic scheduling efficiency and the task execution efficiency of the computing power resources are improved.

The invention discloses a quantitative evaluation method for the cavitation effect of an emulsifying machine, and the method comprises the steps: obtaining the multi-source heterogeneous time series data of pressure, acoustics, current, temperature and the like in real time through various types of sensors, carrying out the synchronous calibration, signal standardization and efficient denoising, extracting the multi-dimensional and multi-scale statistics and frequency domain features, and carrying out the quantitative evaluation of the cavitation effect of the emulsifying machine. A working condition sensitive heterogeneous feature pool is formed through normalization and dynamic screening, feature weight distribution is achieved in combination with an emulsifying machine working condition label, the cavitation intensity and the influence index are dynamically calculated based on a self-adaptive multi-task cooperation model, and the cavitation monitoring accuracy, adaptability and long-term reliability of the emulsifying machine in a complex environment are effectively improved.

This invention discloses a machine vision-based method and system for dynamic screening and size detection of glass fiber particles, specifically relating to the field of glass fiber particle size detection. The method includes real-time monitoring of particle adhesion and dynamic adjustment of the vibration dispersing mechanism, further integrating vibration signals and image features for analysis, and correcting motion blur and refractive index distortion in the image to accurately measure the physical size of the particles. The machine vision-based method and system for dynamic screening and size detection of glass fiber particles achieves coordinated control of vibration and imaging by dynamically adjusting vibration parameters through the adhesion index, effectively solving the problem of lack of coordination between the vibration dispersing mechanism and the imaging mechanism. Through distortion correction algorithms, it compensates for motion blur caused by vibration and optical distortion caused by differences in glass fiber refractive index, reducing measurement distortion. Finally, through a reinforcement learning model, it minimizes size measurement errors while ensuring high dispersion.