67 results about "Spatiotemporal Analysis" patented technology

The invention discloses an urban water supply management data trend analysis method based on space-time analysis, and relates to the field of data processing, and the method comprises the steps: collecting data in real time through an urban water supply pipe network sensor network, building a space-time unified coordinate system, building a space-time Kriging interpolation model based on pipe network topology, and achieving the space-time alignment of multi-source data; dividing an adaptive space-time grid by using a Voronoi diagram and a sliding window mechanism, and calculating multi-dimensional features; constructing a dynamic space-time diagram by taking a grid as a node, performing multi-step prediction in combination with a space-time diagram convolution circulation network, fusing a Kriging interpolation result, and evaluating an abnormal probability and a confidence interval through a Bayesian neural network; a monitoring layer, a prediction layer and a risk layer are overlaid in a three-dimensional GIS, a dynamic thermodynamic diagram is generated, an early warning path is optimized based on a Dijkstra algorithm, and a minimum risk topology path is output. The method has the advantages that through space-time analysis and accurate prediction, the intelligence, stability and emergency response efficiency of urban water supply management are remarkably improved, and powerful support is provided for smart city construction.

The invention relates to a tartary buckwheat pest and disease damage dynamic monitoring method, system and device and a medium, and belongs to the technical field of agricultural intelligent monitoring, the dynamic monitoring method comprises the following steps: periodically collecting environmental parameter data through fixed sensor nodes deployed in a farmland, and obtaining leaf vibration signals and multispectral image data at the same time; performing space-time alignment on the blade vibration signal and the multispectral image data, correcting radiation distortion in the multispectral image data, and outputting a registration data set; according to the registration data set, fusing to generate a multi-modal feature vector, inputting the multi-modal feature vector into a pre-trained space-time analysis model, and outputting a risk level distribution diagram with a geographic coordinate mark; generating a control instruction set according to the risk level distribution map, and triggering execution equipment to execute pest control operation; and optimizing weight parameters of the space-time analysis model through the generative adversarial network based on the execution log of the control instruction set and the historical multi-modal feature vector. The scientificity and timeliness of pest control can be improved.

The invention provides an unmanned aerial vehicle cluster forest fire scene three-dimensional situation construction and updating method and system, and the method comprises the steps: receiving real-time fire scene data from an unmanned aerial vehicle cluster, and building an initial fire scene three-dimensional situation model based on a UNet network and a digital elevation model; the fire scene three-dimensional situation model comprises a fire scene three-dimensional situation map and a pyramid hierarchical structure; when a preset updating period is reached, constructing a pyramid hierarchical structure of the key updating area based on the real-time fire scene data; according to the pyramid hierarchical structure of the key updating area, the pyramid hierarchical structure in the to-be-updated fire scene three-dimensional situation model is compressed based on space-time analysis driving of information entropy, then robust registration fusion is carried out, and the fire scene three-dimensional situation model is updated.

The invention relates to the field of medical image processing and artificial intelligence, in particular to a pulmonary nodule treatment effect AI evaluation system which comprises an image acquisition module, an image registration module, a feature representation module, a multi-scale analysis module, a trajectory analysis module, a response prediction module and a decision support module. According to the system, accurate alignment of CT images before and after treatment is realized through a 4D registration technology, manifold representation of a pulmonary nodule state is constructed based on a differential geometry theory, and nodule features are mapped into points on a high-dimensional manifold; extracting features of different time and space scales by adopting multi-scale space-time analysis, and constructing a manifold trajectory representing a treatment response process; geodesic prediction is realized by using a Riemann geometric framework, and long-term curative effect is predicted from early treatment response; the system not only evaluates the current treatment effect, but also can provide personalized treatment suggestions and optimal follow-up visit plans.

A time-lapse image classification device and method is disclosed that uses a diffractive optical network to classify an optical input, significantly advancing classification accuracy and generalization performance on complex input objects by using the lateral movements of the input objects and / or the diffractive optical network relative to each other. The design space and performance limits of time-lapse diffractive optical networks were numerically tested, revealing a blind testing accuracy of 62.03% on the optical classification of objects from the CIFAR-10 dataset. This constitutes the highest inference accuracy achieved so far using a single diffractive optical network on the CIFAR-10 dataset. Time-lapse diffractive optical networks will be broadly useful for the spatio-temporal analysis of input signals using all-optical processors.

Generating a semantic synopsis is provided. A dataset comprising structured data is received and semantic analysis of the dataset is performed to determine categories of information included in the datasets. A spatio-temporal analysis of the dataset is also performed to determine time periods and locations to which the dataset applies. A static synopsis of the dataset is then generated in natural language sentences that describes semantic information in the dataset and time periods and locations covered by the dataset. The synopsis can be customized according to a user role and search query data.

The invention discloses a real-time storage management method for a large amount of dynamic data, and relates to the technical field of data management, and the method comprises the steps: carrying out the real-time analysis of a dynamic data flow through a storage configuration matrix by using a pulse neural network, generating a data feature vector, and dividing the cold and hot categories of the data through an intelligent classifier, meanwhile, optimization is carried out according to a compression strategy in the storage configuration matrix, a preprocessed data packet is output, intelligent routing distribution is carried out according to a coding label of the preprocessed data packet, hot data is routed to a high-speed storage layer, and cold data is routed to a capacity storage layer; executing data storage and index construction in each storage layer through a plastic queue management mechanism, and generating a storage position mapping table; according to the method, the pulse distribution rate is calculated through the pulse neural network to generate the spatio-temporal characteristics, and the spatio-temporal characteristics are input into the gradient boosting tree model to execute weighted voting to divide cold and hot categories, so that deep spatio-temporal analysis and intelligent classification of dynamic data streams are realized.

The invention discloses a self-adaptive precision polishing control system and method based on multi-modal perception and artificial intelligence decision. According to the system, a multi-mode sensing network is constructed through an acoustic emission sensor, a force / torque sensor, a vibration sensor, a visual sensor and a temperature sensor, and physical signals in the polishing process are collected in real time. The system adopts a deep learning model to carry out feature fusion and space-time analysis on multi-source sensing data, realizes accurate estimation of the material removal rate and comprehensive scoring of the surface quality, further integrates a reinforcement learning optimization engine, takes a process state as input, takes control parameter adjustment as action, and takes a multi-target reward function as guidance, and realizes comprehensive evaluation of the surface quality. And the autonomous optimization decision of the polishing parameters is realized. The invention further comprises a digital twinning system for offline training and online verification, and a transfer learning-based small sample rapid adaptation mechanism, so that the polishing quality consistency, the processing efficiency and the process autonomy can be remarkably improved.

The invention discloses an efficient production scheduling optimization method and an intelligent scheduling system, and belongs to the technical field of production system comprehensive control, and the method comprises the steps: obtaining production data in real time, inputting a time-space analysis model of a fusion graph neural network and a time sequence analysis algorithm, and obtaining an equipment fault probability in a future preset time window; when the fault probability of the equipment is greater than a preset fault threshold value, generating a plurality of candidate scheduling schemes by using a preset quantum heuristic algorithm according to a preset process constraint condition and the production data set, and evaluating a comprehensive performance index; and selecting a candidate scheduling scheme according to an evaluation result, and obtaining a final scheduling scheme to generate an equipment control instruction. According to the method, dynamic fault prediction fusing a graph neural network and time sequence analysis, multi-target optimization scheduling of a quantum heuristic algorithm, adaptive processing parameter adjustment and a high-reliability digital twinning technology are adopted, efficient and intelligent production scheduling can be achieved, and the production efficiency and the equipment utilization rate are improved.

This invention provides a spatiotemporal embodied intelligent system integrating BeiDou and remote sensing, belonging to the interdisciplinary research field of Earth spatiotemporal analysis and decision-making and artificial intelligence technologies. It includes an intelligent sensing module, a data fusion processing module, an intelligent decision-making module, and an execution and feedback module. The intelligent sensing module acquires positioning and remote sensing data; the data fusion processing module fuses the positioning and remote sensing data acquired by the intelligent sensing module to obtain spatiotemporal data of the target area; the intelligent decision-making module generates decision instructions for the target area based on the spatiotemporal data; and the execution and feedback module executes corresponding operations according to the decision instructions and provides feedback results. Furthermore, it enables intelligent equipment terminals such as unmanned aerial vehicles, unmanned agricultural machinery, and unmanned ships to execute corresponding operations based on the decision instructions, and continuously learns, evolves, and optimizes the feedback results.

The application discloses a kind of war destruction global evaluation method based on multi-source remote sensing data, can realize the spatio-temporal continuous monitoring and evaluation of large-scale regional war destruction.It includes the following steps: obtaining and preprocessing land use, NPP-VIIRS night light, multispectral image and surface temperature and other multi-source remote sensing data;Using local optimal threshold method to extract built-up area boundary;Based on total value of night light and light ratio index, using spatio-temporal cumulative calculation method to represent the spatio-temporal evolution characteristics of armed conflict in built-up area;Through principal component analysis method, build non-built-up area war damage index, and reveal the evolution trend and spatial differentiation of non-built-up area armed conflict based on spatio-temporal analysis method.The method of the application can effectively depict the spatio-temporal characteristics of war destruction, has the advantages of global coverage, time sequence integrity, high resolution, and can provide scientific decision support for humanitarian aid and post-war reconstruction planning in conflict area.

The invention relates to a medical image AI auxiliary diagnosis method based on dynamic feature extraction. The method comprises the following steps: carrying out protocol / equipment domain conditioned space-time analysis on medical image data corresponding to a target area to obtain aligned sequence image data; performing dynamic representation construction on the aligned sequence image data to obtain a modeling feature representation sequence; performing hidden state dynamic modeling on the modeling feature representation sequence to obtain hidden state posterior distribution; and carrying out probability diagnosis reasoning analysis on the hidden state posterior distribution to obtain medical image auxiliary diagnosis data. By adopting the method, the diagnosis accuracy and interpretability can be improved.

The invention relates to the technical field of tailing pond safety and geotechnical engineering, and discloses a tailing pond dam slope anti-sliding stability analysis method based on a Sweden segmentation method. The method comprises the steps of fusing dam slope form point cloud, a stratum lithology profile and a multi-stage pore water pressure monitoring sequence through a field domain conversion technology, and constructing a unified three-dimensional space-time analysis model. And intelligently generating a candidate slip trend surface based on the high-precision point cloud data, and carrying out strip division. Rock and soil parameters and pore water pressure which change along with time and space are given to each strip block, and a dynamic strip state set is formed. And carrying out loop iterative calculation by utilizing a Sweden slice method principle, simulating stability evolution under different working conditions, and outputting a stability result of each sliding surface. Finally, the dominant slip plane and the most unfavorable working condition are identified, and an accurate basis is provided for reinforcement decision. According to the method, deep fusion of multi-source data and dynamic simulation of a stability state are realized, and the accuracy of an analysis result and engineering applicability are improved.

The invention discloses a space-time analysis method and device for accumulated snow on a photovoltaic panel in a high-cold and high-altitude area. The method comprises the following steps: partitioning a terrain slope; photovoltaic array slope division extraction is carried out; the unmanned aerial vehicle shoots a panel front image and preprocesses the panel front image; photovoltaic panel detection and accumulated snow detection; calculating the snow coverage rate of the current panel; and predicting a future snow coverage rate. By adopting the technical scheme of the invention, under the condition of paving the photovoltaic panel at high and cold altitudes, the detected accumulated snow on the photovoltaic panel has higher accuracy, and the predicted coverage rate of the accumulated snow on the photovoltaic panel also has higher reliability.

The invention relates to the technical field of power systems, and discloses a Saggob base group extreme turning weather output characteristic and risk assessment method, which comprises the following steps: constructing a multi-source data fusion and extreme scene identification criterion system, and constructing a multi-dimensional meteorological factor matrix; establishing a plurality of extreme meteorological scene criteria; constructing an output factor matrix, setting a low output coefficient threshold value and a ramp rate threshold value, and establishing an output abnormity criterion; identifying a plurality of predefined extreme turning weather scenes; constructing a spatial weight matrix; and performing time autocorrelation function and partial autocorrelation function analysis on the occurrence sequence of each base in each extreme meteorological scene by using the spatial weight matrix and the output factor matrix, and constructing a multi-dimensional risk indicator system. The method has the advantages that a meteorological criterion-output response-space-time analysis-risk assessment framework is constructed, and the problems of inaccurate extreme weather output characteristic capture, insufficient space-time association description depth, single risk assessment dimension and the like are solved.

The application discloses an EEG intelligent agent automatic analysis method based on a large language model, takes the large language model as a strategy engine, autonomously understands user analysis intentions and performs intelligent decomposition of tasks, and then dynamically schedules and cooperatively integrates analysis resources including traditional feature engineering, diversified deep learning models and external knowledge bases. The application can realize end-to-end automatic coordination and execution of complex EEG analysis tasks from signal preprocessing, feature extraction, event positioning to classification diagnosis, emotion recognition, sleep staging, etc., breaks through the limitation of single detection or classification tasks, and enables multi-task, continuous deep reasoning and interpretation of complex EEG data through context perception and flexible space-time analysis capability. The application deeply integrates the general planning and reasoning capability of the large language model with the special analysis model in the EEG field, and significantly improves the automation level, flexibility and clinical application potential of EEG analysis.

The invention discloses a grid merging method and device.The method comprises the steps that the bounding box range of leaf nodes in hierarchical nesting in a target file under a world coordinate system is determined, the target file describes a hierarchical nesting structure and metadata of a three-dimensional tile data set, and the hierarchical nesting structure and the metadata of the three-dimensional tile data set are stored in the bounding box range; the leaf node represents a final rendering unit of the three-dimensional model; performing gridding on the bounding box of each leaf node to obtain a first grid set; and executing a merging operation on the grids in the first grid set. According to the method and the device, the technical problem that space-time analysis and modeling calculation are difficult to carry out on the basis of a unified space unit due to the fact that the three-dimensional tile lacks a gridding representation method of a deep space structure of the three-dimensional tile in the related technology is solved.

The invention relates to the technical field of water and soil conservation, and discloses a disturbance unit division method in a soil loss amount measuring and calculating process, which comprises the following steps of: acquiring a multi-period remote sensing image and a multispectral image of a project area, and performing spatial registration preprocessing on the multispectral image; multi-dimensional disturbance feature extraction is carried out; performing multi-scale image segmentation, and intelligently identifying a disturbance unit in combination with a segmentation result; eliminating classification noise, and extracting a disturbance unit boundary; smoothing the boundary of the disturbance unit; judging the disturbance type of each disturbance unit; carrying out area calculation and attribute labeling based on a judgment result; performing disturbance unit division on the multi-period remote sensing image, and performing change detection according to a division result; performing space-time analysis on a detection result; standardized result output and precision verification are carried out; through multi-source remote sensing data fusion and an intelligent classification algorithm, technicians are liberated from heavy manual sketching work, and the division efficiency is improved compared with that of a traditional manual method.

The invention discloses a wind turbine generator gearbox fault early warning system based on a space-time causal reasoning network, relates to a wind turbine generator early warning technology, and provides a scheme for solving the problem of insufficient early warning capability in the prior art. The causal reasoning module extracts time domain and frequency domain features from original signals collected by a plurality of vibration acceleration sensors of a wind turbine generator, constructs a dependency network among the sensors through causal reasoning, and establishes a dynamic causal atlas. The space-time analysis module captures an abnormal propagation mode through a space-time cooperation mechanism in combination with the stability of a time sequence and the topological characteristics of a spatial causal network; the fault cause and confidence coefficient weighting module identifies a fault root cause through a causal reasoning mechanism, and enhances the confidence coefficient of a detection result based on multi-source evidence; the early fault early warning module realizes intelligent fault detection and accurate positioning based on a causal enhancement probability model, and provides graded early warning suggestions. The method has the advantages that a causal relationship is established in combination with correlation analysis, and a causal graph and comprehensive health indexes are constructed.

The invention discloses a time-varying gravity data probability principal component analysis method considering measurement error covariance. The method comprises the following steps: constructing a measurement equation; introducing a monthly scale and a non-diagonal measurement error covariance matrix provided by a data processing center, and setting a variance component unknown number to adjust the consistency of the variance component unknown number; performing three-decomposition on the total signal by adopting improved PPCA, and realizing parameter maximum likelihood estimation by combining an EM algorithm; selecting the number of principal components of the optimal signal in the alternative set through AIC; and outputting a strip-removed signal reconstruction result and a plurality of space-time orthogonal signal components thereof, and meanwhile, giving out estimation of noise and measurement errors. According to the method, measurement errors such as stripes can be effectively distinguished and suppressed while real geophysical signals are reserved, and the result quality is remarkably improved in the aspects of signal reconstruction and space-time analysis; and meanwhile, the method has the advantages of low calculation amount, high convergence speed and the like, and is suitable for processing GRACE / GFO data with inter-monthly covariance change characteristics.

The invention relates to the technical field of ecological monitoring and prediction, and discloses an ecological restoration effect prediction system based on space-time sequence analysis. According to the system, a prediction process is realized through cooperative processing of an ecological data reconstruction unit, a spatial adaptation judgment unit, a time sequence change mapping unit, an effect anomaly screening unit and a risk prediction output unit. The system firstly collects and processes ecological data, and generates a restoration point ecological suitability data set; identifying spatial consistency and difference sections through spatial proximity sorting; evaluating the ecological response intensity in combination with environment time sequence data; abnormal point locations are screened based on composite conditions of response value anomalies and spatial difference sections; and finally generating risk early warning. The method has the technical effects that through coupling space-time analysis, the defect that the space context is neglected when a traditional method depicts spatial heterogeneity and performs anomaly recognition is overcome, and more accurate prediction and risk positioning on the ecological restoration effect are realized.

This application relates to a dynamic monitoring method and system for urban ground collapse disasters based on multi-source sensor information fusion. It addresses the problems of existing monitoring methods, such as single monitoring dimensions and isolated data, leading to insufficient detection of collapse precursors and low accuracy and timeliness of early warnings. The method includes: collecting multi-source data on pressure, displacement, vibration, and remote sensing; standardizing and spatiotemporally aligning the data; and then performing noise suppression and error compensation. A deep learning model is then used to automatically extract deep features from the multi-source data and perform fusion analysis. The generated feature vectors are input into a risk discrimination model, outputting risk distribution and deformation prediction. Finally, a dynamic risk map is constructed for interactive display, and graded early warnings and emergency response are automatically executed based on the risk level. This application has the following effects: achieving deep fusion and spatiotemporal analysis of multi-source heterogeneous data, improving the accuracy of early collapse risk identification and dynamic early warning capabilities.

This invention relates to the field of data visualization technology, specifically to a fluid data visualization method and system based on spatiotemporal analysis. The method includes the following steps: obtaining the sampling point number, coordinates, and deployment duration; collecting rainfall, temperature, humidity, and pressure data; calculating confidence levels to eliminate low-confidence data; extracting parameter sequences to analyze trends; generating anomaly level labels; constructing a trend offset factor group; identifying high-offset variability grids; and generating a dynamic variability map sequence. This invention utilizes the calculation of sampling confidence levels to eliminate invalid or interfering data. It constructs anomaly level label sets by comparing and classifying trend change indicators with adjacent spatial partitions, enhancing the ability to characterize spatiotemporal trend changes. It uses a dual standard of trend reversal and anomaly density to label the state of anomaly areas, effectively capturing potential anomaly variability areas and assigning them identification codes, thus comprehensively improving the anomaly identification accuracy and area positioning efficiency in the fluid data visualization process.

This application provides a method and apparatus for identifying cloned vehicles. Through an innovative spatiotemporal analysis model, it achieves accurate screening of unreachable vehicles via distance calculation and speed verification. A feature recognition system is constructed, combining image segmentation and tag extraction to establish a reliable mechanism for identifying cloned vehicles. Predictive deployment is introduced, using trajectory analysis and route prediction to ensure the accuracy of strikes. This method effectively addresses the shortcomings of traditional technologies in spatiotemporal analysis, feature recognition, and predictive deployment, providing technical support for the control of cloned vehicles.

This invention discloses a method for simulating the spatial evolution of commercial districts based on spatiotemporal analysis, relating to the field of smart city technology. The method includes: collecting multi-source spatiotemporal data to construct an initial spatiotemporal knowledge graph, and using the initial spatiotemporal knowledge graph to obtain a baseline simulation model; collecting real-time multi-source data streams and incrementally updating the initial spatiotemporal knowledge graph based on the real-time multi-source data streams to form a dynamic spatiotemporal knowledge graph; detecting state changes in the dynamic spatiotemporal knowledge graph, and when any rate of change of key entities or key relationships exceeds a predetermined change threshold, triggering and executing incremental simulation of a local area to obtain a local predicted state graph, and updating the local predicted state graph to the dynamic spatiotemporal knowledge graph to obtain a fused predicted knowledge graph. This invention achieves accurate capture and rapid response to micro-level disturbances.

The invention relates to the technical field of semiconductor manufacturing and detection, and discloses an insulation part defect detection system based on an ion implanter. The system comprises a data reconstruction module for dynamically associating three-dimensional coordinates, time sequence marks, vacuum pressure and thermodynamic parameters and constructing a quality database; the adaptation judgment module is used for carrying out space-time analysis on the quality data and outputting a quality consistency partition map; the rate mapping module quantifies fluctuation and evaluates a defect sensitivity level based on parameter time sequence evolution and thickness distribution; the early warning screening module is used for screening high-sensitivity abnormal instances according to a dynamic threshold value and constructing a defect abnormal instance library; and the risk analysis output module is used for integrating the instance data, marking vacuum leakage risk nodes and outputting defect detection and risk early warning results. According to the system, multi-stage collaborative analysis and real-time risk management and control in the insulating part manufacturing process are realized, and the defect identification precision and the early warning efficiency are effectively improved.

This invention provides an interpretable spatiotemporal analysis method for traffic congestion prediction, extracting key features that trigger congestion events and deep connections between roads from the interpretation. Traditional data mining methods often explore the correlations between traffic spatiotemporal data from a statistical perspective, failing to fully reveal the deep connections and key factors of traffic congestion. Therefore, this invention proposes a spatiotemporal interpretation generation model based on STGCN, leveraging the ability of neural networks to discover hidden features and using deep learning interpretability techniques to extract key input features of interest to the neural network. The model uses a perturbation-based interpretation method to generate a mask and a gradient-based interpretation method to generate the gradient mapping of the mask; furthermore, considering the problem of coarse granularity and poor targeting of spatial masks, a step-by-step masking method is proposed to reduce the interpretation granularity. This increases the effective extraction of hidden information, thereby obtaining more accurate and comprehensive key congestion information.

The invention discloses an urban carbon emission detection method based on an unmanned aerial vehicle, and belongs to the field of carbon emission detection, and the method comprises the steps: obtaining multi-source data from a high-precision carbon emission data set, carrying out the cleaning and formatting of the data through a data preprocessing technology, and obtaining a standardized data set; for the standardized data set, adopting a multi-source data fusion algorithm, and combining weather and traffic data to generate a fusion data set; if the timestamp of the fused data set is later than a preset threshold value, performing incremental updating on the data set through a data real-time updating mechanism to obtain a dynamic data set; according to the dynamic data set, constructing an urban environment model, simulating a complex urban environment through a space-time analysis algorithm, and generating a dynamic environment model; aiming at the dynamic environment model, adopting a dynamic simulation analysis technology and combining with the carbon emission dynamic data set to generate a carbon emission distribution diagram; and obtaining spatial distribution characteristics from the carbon emission distribution diagram, and performing prediction analysis through a machine learning algorithm to obtain a carbon emission change trend.

The invention relates to a rural regional development state comprehensive measurement method and device and a storage medium, the method is applied to the field of computer technology and rural development technology, and the method comprises the following steps: constructing a six-dimensional characteristic index system of rural regional development state measurement by adopting a multi-level fuzzy comprehensive evaluation model, wherein the driving contribution judgment sub-model comprises a natural climate feature dimension, a regional environment feature dimension, a traffic facility feature dimension, an urban and rural associated feature dimension, a resource endowment feature dimension and an economic development feature dimension, and the driving contribution judgment sub-model integrating the six feature dimensions establishes a rural regional development state six-dimensional comprehensive measurement model. A high-precision curved surface modeling method and a GIS space-time analysis method are used to realize space-time simulation of multi-source heterogeneous feature data corresponding to each feature dimension, and a six-dimensional comprehensive measurement model is used to generate a system measurement result of a rural region development state based on six-dimensional feature space-time simulation data. According to the method, the limitation of a single-dimension or single-region measurement result can be avoided.

This application discloses a remote sensing monitoring method, device, equipment, and medium for sulfur dioxide in the atmosphere, relating to the field of remote sensing detection technology. The method includes: indexing high-resolution satellite remote sensing spectra, pre-processed spatiotemporal data, and wavelength data from a target dataset obtained through spatiotemporal analysis of an initial dataset; spatially splitting and pixel labeling the spectral data of the CCD row and column range corresponding to the area to be inverted based on the spatiotemporal data, and processing multiple partitioned data blocks in parallel; spectral splitting of the observed spectra of each partitioned data block based on the wavelength data, and inputting the resulting multiple sub-band data blocks into a radiative transfer model in parallel to obtain the simulated spectrum corresponding to each sub-band data block; fitting and integrating the simulated and observed spectra, and correcting the initial inversion result of the sulfur dioxide concentration to obtain the target inversion result. This method can improve the computational efficiency of sulfur dioxide inversion.