30results about How to "Guaranteed generalization ability" patented technology

The invention discloses a robust federated learning method for processing heterogeneous noise and non-independent identically distributed data, and belongs to the technical field of federated learning. The method provides a robust learning framework of two-stage client quality perception. The method comprises the following steps of: 1, constructing a category-level loss vector and clustering by using a Gaussian mixture model, and accurately dividing a clean and noise client set; stage 2, performing differential training: performing standard training on the clean client; dual-network cooperative training, dynamic sample screening and exchange, and a self-distillation and entropy regularization mechanism are introduced to a noise client, so that robust learning is realized; in the global aggregation stage, a distance sensing weighting strategy is further adopted to dynamically suppress the influence of a noise client; according to the method, original data does not need to be shared, the robustness and generalization performance of the federated learning model in the coexistence environment of heterogeneous noise and non-independent identically distributed data can be effectively improved, and the method has wide application value in the fields of medical images, financial risk control and the like.

The invention discloses a horizontal well multi-cluster fracturing crack propagation prediction and parameter optimization method and device, and the method comprises the steps: inputting the target input data of a target fracturing section into a fracturing crack propagation prediction model, and outputting a crack propagation prediction result; the target input data comprises preprocessed geological parameters and engineering parameters; the fracturing crack propagation prediction model is obtained by training according to a preset training set; the preset training set is a data set obtained by performing data enhancement on the input sample data and the corrected crack expansion sample data; based on the fracturing fracture propagation prediction model, determining the contribution degree of each geological parameter and engineering parameter in the target input data to a fracture propagation prediction result, and screening key parameters influencing multi-cluster fracture equilibrium propagation; and adjusting engineering parameters in the key parameters to serve as optimization engineering parameters by taking a crack propagation prediction result meeting a multi-cluster crack balanced propagation requirement as a target. According to the method, accurate and rapid prediction of crack propagation and targeted optimization of engineering parameters can be realized.

The application discloses a tunnel or mine gushing water space-time prediction method and system coupled with a water power numerical model, and the method comprises the following steps: based on the identified and verified underground water numerical model, outputting multi-source data, complementing the missing measured data, quantifying the difference between fault and normal stratum permeability characteristics and coupling to the data system, and incorporating the tunnel or mine excavation space data. An LSTM-isolation forest-K nearest neighbor regression coupled model is constructed, and a multifunctional module is configured to realize common training of multi-scene data. The pretreated multivariate time series data is divided into a training set and a test set, hidden features are extracted through the coupled model, abnormal detection results are fused, a residual correction model is trained synchronously, and the hyperparameters and weights are adaptively optimized according to the multi-project prediction error feedback. Based on the trained coupled model, a window rolling strategy is adopted to carry out synchronous gushing water space-time prediction, residual correction is combined, and prediction data meeting the engineering precision is output. Reliable technical support is provided for engineering construction safety control.

The invention relates to the technical field of artificial intelligence image recognition and data mining, and discloses an image data mining processing system and method combining a visual large model and a visual small model. The method aims at solving the problems that in the prior art, the image recognition recall rate and accuracy are difficult to consider, the calculation cost and the processing speed are unbalanced, the data management efficiency is low, and the model generalization ability is insufficient. The system comprises a data input module, a frame extraction deduplication module, a picture library construction module, a visual large model judgment module, a visual small model judgment module, a label preprocessing module, a manual inspection module, a training data generation module, a model training module and a data security module. According to the method, the generalization ability is guaranteed through preliminary screening of an LPM (Deic) large model, missing detection is complemented through secondary verification in cooperation with a YOLO small model, false alarms are eliminated in combination with manual verification, meanwhile, double models are optimized through targeted sample training, the pain point of missing detection or misjudgment of a single model is effectively solved, and accurate and comprehensive consideration of target recognition is achieved.

The intelligent automobile lane-changing decision method and device, the electronic device and the storage medium provided by the embodiments of the present disclosure comprise: constructing a virtual lane-changing scene by using a simulation method; constructing a decision model for determining a mapping function between a state and a reward, obtaining a Q table for evaluating the lane-changing feasibility of an intelligent automobile based on the mapping function, and obtaining a lane-changing decision, i.e., "lane-changing" or "not lane-changing"; based on a greedy algorithm, making the intelligent automobile produce a random lane-changing behavior in the constructed virtual lane-changing scene, storing the state and the reward of the current lane-changing behavior in an experience replay pool, when the number of state-reward pairs in the experience replay pool reaches a minimum sample number, randomly sampling a plurality of state-reward pairs from the experience replay pool for single-step training of the decision model, storing a new state-reward pair generated in the training process in the experience replay pool, repeatedly performing the above training process until a maximum training number is reached, and obtaining a lane-changing decision model. The lane-changing decision generated by the present disclosure is high in safety and wide in applicability.

The invention discloses a complex scene traffic perception method and system based on fuzzy logic and data enhancement, and relates to the technical field of automatic driving environment perception. The method comprises the steps of obtaining an RGB image and point cloud data, and performing preprocessing operation to obtain a sample set; expanding the sample set by using a forced enhancement mechanism to obtain an expanded sample set; performing exclusive feature enhancement operation on the expanded sample set according to the scene complexity to obtain enhanced features; performing dynamic weight calculation and feature fusion on the enhanced features based on factors influencing traffic perception by using a fuzzy logic algorithm to obtain fusion features; and performing perceptual prediction on the fused features by using a perceptual prediction model to obtain a perceptual prediction result. The method combines forced weather enhancement and fuzzy logic fusion technologies, adapts to complex scenes, and solves the problem that multi-modal sensing robustness and generalization are insufficient in severe weather.

This invention relates to the field of scene text segmentation, specifically a scene text segmentation method based on an improved SAM visual segmentation large model. Based on the SAM visual large model, this invention extracts text content perception features through an image content perception module and text edge perception features through a text edge perception module. Furthermore, the text feature fusion module extracts and calculates text edge perception feature maps, which are then added to the vectors requiring attention calculation before each self-attention calculation in the SAM encoder. This improves the accuracy of SAM in text segmentation and shortens the model training time while maintaining generalization.

This invention discloses a method for constructing a parameter extraction model for active ingredients in traditional Chinese and Tibetan medicines. The method involves acquiring multi-source sample data of traditional Chinese and Tibetan medicines and constructing key derived features from the preprocessed multi-source sample data. A random forest model is used to generate a corresponding importance score for each key derived feature, forming a feature importance vector. The feature importance vector and the key derived features are weighted and summed to obtain a fusion feature, and the variance and correlation coefficient matrices corresponding to the fusion feature are calculated. The average values ​​of the variance and correlation coefficient matrices are weighted and summed to obtain a comprehensive score. The depth and width of the parameter extraction model to be trained are determined based on the comprehensive score. The number of fully connected modules and neurons in the parameter extraction model is determined by the depth and width to obtain the corresponding target architecture. The parameter extraction model under the target architecture is trained using the fusion features and extracted parameter data to obtain a trained parameter extraction model.

The application discloses a social media-based interpretable dynamic graph network benefit prediction model training method, relates to the technical field of text analysis and market prediction, and obtains market data and related social media data of a financial asset to divide the data by weeks; for the social media data of each week, a topic is taken as a node, and similarity between topics is taken as an edge weight, a topic correlation graph of each week is constructed to obtain weekly graph data; a prediction model performs time sequence updating and market prediction according to the weekly graph data; a text time sequence memory unit obtains a current week topic memory vector according to a current week topic text embedding vector and in combination with a last week topic memory vector; a multi-layer graph attention network captures the correlation features between nodes through an attention mechanism and adopts attention weights for weighted summation to obtain a current week global graph embedding; and a classifier outputs a next week price change prediction result. The application solves the problem that the prior art cannot effectively capture the time sequence evolution of topic sentiment and the correlation between topics.

This invention discloses an improved machine learning-based intelligent prediction method for stratigraphic models and related equipment. By fusing depth-aligned high-frequency active seismic source data and low-frequency passive seismic source data, a multi-dimensional complementary source feature vector is constructed, significantly improving the vertical resolution and deep penetration capability of stratigraphic prediction. A genetic algorithm is used to globally search and optimize the network parameters of the convolutional neural network. Combining binary encoding and genetic operations effectively avoids the pitfalls of traditional gradient descent methods, which are prone to getting trapped in local optima, thus improving the globality and stability of model parameter optimization. Through iterative fitness verification and threshold determination, the optimal network parameters are adaptively determined without manual intervention, significantly reducing the cost of hyperparameter tuning while effectively ensuring the generalization performance of the prediction model. The final prediction model can perform fast and accurate stratigraphic prediction based on real-time source data and can be widely applied in the field of data processing technology.

The invention relates to a GNSS deception signal detection method based on composite machine learning. The detection method comprises the steps that detection parameters are acquired; performing standardization processing on the obtained detection parameters, constructing to obtain a detection feature vector, and making a data set according to the detection feature vector; building a deep learning model based on a double-layer GRU structure and an SVM, and training the deep learning model through the data set; and the detection feature vector is input into the trained deep learning model again to calculate the deception confidence coefficient, and the closer the confidence coefficient is to 1, the higher the possibility that the current signal is a deception signal is. Through multi-parameter time domain feature extraction and detection, the detection algorithm can effectively detect different types of deception jamming in a complex environment, multiple parameters ensure the generalization ability of the algorithm for different deception, and the time domain feature can help the algorithm to better distinguish the multipath effect and deception jamming.

The application discloses an electrochemical noise corrosion state identification and early warning method and system, and belongs to the technical field of corrosion monitoring and early warning. The method collects the potential noise signal and the current noise signal of the target component through an electrochemical workstation and calculates the noise resistance, adopts db4 wavelet for multi-layer wavelet decomposition and threshold denoising processing, extracts the kurtosis feature, the skewness feature and the wavelet coefficient feature, and constructs a corrosion state identification and early warning data set after feature screening; a fusion model including an LSTM time sequence feature extraction layer and an Attention attention weight distribution layer is constructed for training, and a corrosion state identification model is obtained; the real-time collected and processed feature parameters are input into the model for identification, the proportion of a specific corrosion state in a preset time period is counted, and a blue, orange or red graded early warning is triggered. The application realizes high-precision identification and accurate graded early warning of the corrosion state, the identification accuracy is above 85%, and reliable technical support is provided for industrial component corrosion protection.

The invention relates to a multi-target parameter self-adaptive noninvasive tibial nerve stimulation method and system, and aims to solve the problem that an existing nerve regulation device cannot perform on-demand treatment according to dynamic change of symptoms of a patient. The induced foot sole neural signals are collected, and the F wave peak value amplitude and the incubation period meeting the stability condition are extracted in a preset time window to serve as real-time feedback signals. And establishing an individualized target interval based on the F-wave characteristic, calculating a dynamic deviation of a current F-wave peak value amplitude relative to the interval, and iteratively adjusting parameters such as stimulation intensity, frequency, pulse width and a double-target stimulation intensity proportion by taking a mean square error of the dynamic deviation as a loss function and adopting a gradient descent method until the F-wave characteristic is converged into the target interval. The dynamic change of the bladder function is self-adapted, and multi-target closed-loop self-adaptive stimulation of the tibial nerve is realized.

The invention provides a low alloy steel flow stress prediction method and system based on fusion of constitutive parameters and a GA-BP neural network. The method comprises the following steps: firstly, acquiring experimental data of the low alloy steel under different thermal deformation conditions and preprocessing the experimental data; then, based on an Arrhenius constitutive relation model and a dynamic material model theory, constitutive characteristic parameters of the low alloy steel under different thermal deformation conditions are obtained through calculation; combining the preprocessed experimental data with the constitutive characteristic parameters obtained by calculation, and constructing a training data set containing intrinsic parameters, process parameters and constitutive characteristic parameters; using the training data set to train a GA-BP neural network model of a back propagation neural network optimized by a genetic algorithm to obtain a flow stress prediction model; and finally, inputting parameters under a to-be-predicted working condition into the flow stress prediction model, and outputting to obtain a predicted flow stress value.

The application provides a kind of time sequence feature and causal decoupling light weight ViT industrial hidden danger detection method in the technical field of industrial safety detection, comprising the following steps: first, obtaining industrial inspection image and preprocessing into standardized image, then constructing embedded local feature enhancement module and light weight improved Vision Transformer network, extracting enhanced local feature map;Further introduce the mechanism of physical guided causal decoupling, decompose the feature into real hidden danger, environmental interference and background residual, compare the similarity by differentiable counterfactual intervention, and output the interference discrimination reason;At the same time, cross-frame time sequence feature fusion is adopted to align continuous multi-frame features to distinguish hidden dangers and transient interference;Finally, combined with the discrimination reason, the light weight network outputs the industrial hidden danger detection result.The application has the advantages that: it greatly improves the anti-environmental interference ability, weak target perception accuracy, real hidden danger and transient interference time sequence causal discrimination ability, model explainability and end-side real-time deployment efficiency of industrial hidden danger detection.

The invention relates to the technical field of biomedical signal processing, and discloses a calcium ion imaging signal peak inference method and system based on deep learning, and the method comprises the following steps: segmenting an original noisy calcium ion signal time sequence through interval sampling to generate a recovery sequence, and training a one-dimensional U-Net neural network model; processing the historical noisy calcium ion signal time sequence by using a one-dimensional U-Net neural network model to obtain a historical de-noised signal, and training a one-dimensional convolutional neural network model by using the historical de-noised signal; and denoising the original noisy calcium ion signal time sequence to be processed, and inputting the denoised original noisy calcium ion signal time sequence into the one-dimensional convolutional neural network model to deduce a neuron absolute spike rate sequence. According to the method, noise can be suppressed through self-supervised denoising without external truth value data, the signal-to-noise ratio is improved by using a two-stage framework, a neuron discrete discharge event is accurately solved, and the problems that the prior art depends on a truth value database and the inference precision is low under the low signal-to-noise ratio are solved.

This invention discloses an intelligent optimization method for a multi-type well-fracture joint control fine-grained injection-production mode, relating to the field of oil and gas field development technology. This invention sets up a well-fracture joint control fine-grained injection-production mode, establishes a reservoir numerical simulation model in reservoir numerical simulation software, obtains multiple reservoir injection-production schemes based on the Latin hypercube sampling method, simulates each reservoir injection-production scheme using the reservoir numerical simulation model, generates multiple sample data to establish a sample database, and establishes a deep learning agent model. After training the deep learning agent model using the sample database, a particle swarm optimization algorithm is used for single-objective pre-search global optimization to obtain the optimal baseline strategy. A reinforcement learning dynamic decision model is established and trained based on the PPO proximal policy optimization algorithm to obtain a reinforcement learning agent. The reinforcement learning agent is used to obtain the optimal injection-production development scheme for the reservoir, realizing rapid optimization and decision support for reservoir injection-production schemes under a new multi-type well-fracture joint control mode.

The application discloses a method for monitoring the capacity of a vanadium redox flow energy storage system online, comprising the following steps: S1, acquiring multi-dimensional characteristic data collected by a multi-source data acquisition channel; S2, preprocessing the collected data to obtain a standardized characteristic vector; S3, inputting the standardized characteristic vector into a pre-trained kernel extreme learning machine (KELM) model to output a preliminary capacity estimation value; S4, adaptively weighting and fusing the preliminary capacity estimation value and the last valid capacity storage value read from a system historical database to obtain a final capacity monitoring value at the current moment; and S5, updating the system historical database and outputting the final capacity monitoring value. The application adaptively weights and fuses the preliminary capacity estimation value output by the KELM model and the historical valid capacity storage value, and dynamically adjusts the fusion weight according to the relationship between the characteristic frequency offset and the preset sensitivity threshold, so that the mechanism can fully exert the rapid response capability of the data-driven model when the electrolyte conductivity changes significantly.

The invention provides an EEG signal feature extraction method and system, a storage medium and a terminal. The method comprises the following steps: preprocessing an EEG signal; constructing a multi-scale reference window and a multi-scale test window, and calculating a multi-scale spectrum structure change score of the preprocessed EEG signal; detecting candidate segment boundary points of the preprocessed EEG signal based on the multi-scale spectrum structure change score; on the basis of the candidate segmentation boundary points, EEG adaptive segmentation is obtained; performing Hermite function decomposition on the EEG signal in the EEG adaptive segment to obtain EEG segment level features; and constructing an EEG feature vector based on the EEG segment-level features and the EEG global statistics. According to the EEG signal feature extraction method and system, the storage medium and the terminal, EEG signal features with the characteristics of high event sensitivity, high adaptive ability, excellent stability and the like can be extracted.

The invention discloses a single-station photovoltaic power prediction method based on vision and language multi-mode fusion, and relates to the technical field of new energy power generation power prediction, and the method comprises the following steps: S1, collecting historical power data, numerical weather forecast data, satellite cloud picture data at a corresponding moment, and weather text description data of a target photovoltaic power station; s2, performing normalization processing on the historical power data and the numerical weather forecast data to form numerical data; according to the method, four kinds of modal information including historical power data, numerical weather forecast data, satellite cloud picture visual data and weather text semantic data are fully fused, the limitation of a single data source of a traditional method is broken through, multiple factors influencing photovoltaic power can be comprehensively captured, richer information support is provided for prediction, and prediction precision is remarkably improved.

This invention proposes a test-time optimized method for super-resolution reconstruction of UAV aerial infrared images, belonging to the fields of image processing and computer vision. A test-time optimization framework based on unfamiliar infrared degradation perception is constructed: a teacher-student model fusion mechanism is introduced to synthesize degraded reference images, and the optimization amplitude is dynamically controlled to prevent catastrophic forgetting; a frequency domain degradation estimation network calculates the blur kernel and noise parameters to generate pseudo-label images to guide model optimization; a multi-scale linear layer parameter optimization strategy is designed, updating only linear layer parameters to achieve rapid adaptation; and a joint constraint optimization process using reconstruction loss and feature consistency loss ensures pixel-level reconstruction accuracy and the preservation of pre-trained feature extraction capabilities. This method effectively solves the problem of super-resolution quality degradation caused by unknown degradation conditions in UAV aerial photography scenarios, significantly improves generalization ability under complex degradation migration conditions, and is suitable for applications such as nighttime search and rescue, security patrol, and agricultural assessment.

The invention provides a logistics cargo volume prediction method, electronic equipment and a program product, and relates to the technical field of logistics. The method comprises the following steps: acquiring first historical logistics cargo volume time sequence data of a to-be-predicted object, and a first auxiliary feature related to the first historical logistics cargo volume time sequence data; the first auxiliary feature represents an influence factor related to goods quantity fluctuation of the to-be-predicted object; generating an embedded vector sequence corresponding to the first historical logistics cargo volume time sequence data, and inputting the embedded vector sequence into a time sequence large model based on a Transform structure to obtain predicted cargo volume data; the predicted cargo volume data comprises a corresponding predicted cargo volume in a future period of time; performing linear prediction based on the first auxiliary feature by using a linear auxiliary model to obtain deviation data of the predicted cargo volume; and fusing the predicted cargo volume data and the deviation data of the predicted cargo volume to obtain a final logistics cargo volume prediction result. The method and the device are used for providing an efficient solution for logistics cargo quantity prediction.

The application relates to the technical field of water level prediction, and particularly discloses a basin water level prediction system and method for complex terrains, wherein the method is characterized in that: a feature screening algorithm is used to screen a preliminary model feature set, and effective features which have a significant influence on the transit of a flood peak and the risk of backflow are reserved; the effective features are scientifically divided into a training set and a test set; the training set is used to enable the model to fully learn the mapping relationship between the buffer state of a buffer confluence unit, the flood discharge, the movement of the flood peak and the change of the confluence water level under complex terrains; and the fusion of hydrological data and topographic image data covers key influence factors in the time and space dimensions on the data level; the water flow lagging effect is captured through an upstream water level lagging feature set on the feature level; the discharge position is accurately positioned in combination with the discharge channel and the buffer state of the buffer confluence unit; and the transit path of the flood peak and the peak water level are simulated, so that the accurate simulation of the transit of the flood peak in the complex terrain and the reliable prediction of the backflow risk can be realized.

The invention provides a plateau railway icing early warning method and system based on sequence learning, and is applied to the technical field of meteorological disaster prediction, and the method comprises the steps: obtaining attention points and original meteorological observation data along a railway for several years, extracting the original meteorological observation data corresponding to the attention points, and obtaining meteorological data; preprocessing the meteorological data, generating an icing label corresponding to each concerned point location according to a preset icing standard, performing interval grouping based on longitude and latitude, and performing gridding processing on the meteorological data with the icing labels to obtain aggregated data; dividing the aggregated data into training data and test data according to a preset rule, and inputting the training data and the test data into a constructed space-time fusion deep learning model for rolling training to obtain a model parameter corresponding to each prediction year; and obtaining model parameters corresponding to the year to be predicted, substituting the model parameters into the model, inputting the meteorological data to be predicted into the model to obtain a corresponding icing risk prediction result, and performing early warning.

The application discloses a hot shock cooling rock rotary cutting damage prediction method, solves the technical problem that the prior art cannot realize high-precision and high-generalization prediction of hot shock cooling rock rotary cutting damage parameters while ensuring physical mechanism consistency. The method first acquires hot shock cooling rock rotary cutting sample data under different initial temperature conditions, constructs a hot shock cooling rock rotary cutting failure criterion RCFC considering thermal damage evolution and temperature-dependent fracture mechanics characteristics, and then embeds the RCFC as a hard physical constraint into a physical information neural network model, and outputs the prediction results of key damage parameters such as rotary cutting energy and rotary speed after training. The application realizes the deep fusion of data driving and rock cutting damage physical mechanism, significantly improves the prediction accuracy and physical consistency, and still has good robustness under the condition of sparse data, and can provide technical support for drilling parameter design and construction safety control of deep high-temperature rock mass engineering.

The invention discloses a large model accelerated training method based on staged learning, and relates to the technical field of large model training. The method comprises the following steps: step a, dividing a training process of a model into a continuous core structure learning period, a detail feature rich period and a final fine tuning period; b, in the core structure learning period, a first information bottleneck constraint is injected behind a first middle layer of the network, and a loss function of the first information bottleneck constraint is divergence obtained through calculation based on output features of the first middle layer and variational prior distribution. According to the method, a three-stage progressive training framework from core feature learning to detail feature enrichment to final fine adjustment is constructed, differentiated information bottleneck constraints are applied to each stage, the model is guided to follow a feature learning rule from macroscopic to microscopic, calculation waste of redundant features at the initial stage of training is effectively avoided, and the training efficiency is improved. The overall convergence process of the model is accelerated, and the accuracy is improved.

The invention relates to a tobacco production data feature screening technology, in particular to a method and a system for characterizing cigarette end shred falling quantity based on prior process knowledge and a storage medium, and the method comprises the following steps: obtaining cigarette production process data and carrying out dynamic region division; performing multi-dimensional feature extraction on the data after region division to obtain a multi-dimensional feature set; performing edge feature pre-screening according to the multi-dimensional feature set to obtain an effective feature set; obtaining a comprehensive score feature set according to the effective feature set; performing intra-region feature screening on the comprehensive score feature set to obtain an initial feature set and a candidate feature set; and performing improved sequence forward selection fine screening based on the initial feature set and the candidate feature set to obtain an optimal feature subset. According to the characterization method for the tobacco rod end portion shred falling amount, higher characterization precision, higher model robustness, better process interpretability and wider production scene adaptability are achieved.

The present application relates to the technical field of detection of Dendrobium polysaccharide content, and discloses a system and method for rapid nondestructive detection of Dendrobium polysaccharide content, which comprises the following steps: collecting spectral reflection data in the 600 nm-1700 nm band by using a FLA6800 spectrometer, and performing data preprocessing, feature extraction, machine learning model training and prediction to rapidly and accurately predict the Dendrobium polysaccharide content, and finally generating a visual report and performing grading. The present application combines near-infrared spectroscopy technology with a machine learning model to realize efficient, accurate and nondestructive detection of the Dendrobium polysaccharide content, significantly improves the detection speed and accuracy, and provides real-time and reliable quality control basis.

This invention discloses a data forgetting method for dynamically allocating pruning ratios for heterogeneous resources in industrial equipment. The method includes: when a server detects a target device failure, the server allocates an adaptive pruning ratio to each industrial device based on its resource information; each industrial device prunes its local model according to the allocated pruning ratio, removing parameters related to the abnormal data of the target device, thus forgetting the target data; other industrial devices update their pruned local models, obtaining updated local models and uploading them to the server; the server aggregates the updated local models to obtain a global model that guarantees generalization performance and forgets the abnormal data of the target device. This invention can adaptively allocate appropriate pruning ratios to heterogeneous computing and communication resources, thereby reducing the waiting time of all devices, improving system resource utilization, and enhancing the execution efficiency of the forgetting task.

This invention discloses a method for intraday dynamic prediction of cyanobacterial blooms based on meteorological and water quality synergy, relating to the field of water environment monitoring. This method, based on satellite data, identifies hourly cyanobacterial bloom area results images using a multi-index consensus method and calculates the actual coverage area of ​​the cyanobacterial bloom. Using joint data of meteorological factors and water quality parameters as input and the actual coverage area of ​​the cyanobacterial bloom as output, a BP neural network model is constructed and optimized. The optimized model is then used to predict the estimated coverage area of ​​the cyanobacterial bloom in the future hourly, and the efficiency coefficient between the estimated and actual coverage areas of the cyanobacterial bloom is calculated. This invention fully utilizes the high temporal resolution advantage of geostationary meteorological satellites to identify cyanobacterial blooms and establish a prediction model, achieving intraday prediction of the cyanobacterial bloom area.