90 results about "Probability mapping" patented technology

The invention relates to the technical field of medical information, in particular to an intelligent prediction model and method for postoperative complications of anesthetized patients, and the method comprises the steps: collecting preoperative to postoperative complete-cycle clinical data of a patient through a medical data interface; analyzing operation codes to generate risk features, extracting vital sign dynamic features, and establishing a complication probability mapping relation through a multi-modal fusion network; combining the complication probability and pharmacokinetic parameters to construct an optimization model, and solving an individualized anesthetic dosage interval by using a gradient descent algorithm; vital signs are dynamically monitored in the operation, a dose re-optimization mechanism is triggered, the infusion rate is adjusted, and a closed-loop control link is formed; and generating a visual decision report. According to the method, through deep integration of complete-cycle clinical data and multi-modal feature modeling, preoperative physiological parameters, operation coding semantic information and intraoperative vital sign dynamic modes are subjected to fusion analysis, a nonlinear mapping relation between dosage and complication probability is constructed, and the risk prediction precision and individualized adaptability are remarkably improved.

The invention relates to the technical field of simulation model optimization, and discloses an aero-engine model Bayesian optimization method for quantizing uncertainty, and the method comprises the steps: building a probability mapping relation from a component index to an output response through constructing a Bayesian neural network agent model based on a probability weight coefficient; and by taking the difference between the output response and the corresponding complete machine test data as a multi-objective loss function and taking the minimization of the multi-objective loss function as an optimization objective, optimizing the component indexes by adopting a Bayesian optimization method based on a Gaussian process to obtain an optimal component index combination. Not only is a nonlinear relationship between high-dimensional parameters and simulation-test deviation accurately modeled through a neural network, but also efficient search of a parameter space is realized through a Gaussian process. The technical problems that when a traditional optimization method is used for processing the high-dimensional, strong-nonlinearity and multi-parameter coupling complex optimization problem of the aero-engine, the calculation efficiency is low, local optimum is prone to occurring, and result uncertainty cannot be quantified are solved.

The invention belongs to the technical field of brain-computer interfaces, and particularly relates to a brain-computer interface driven emotion recognition and self-adaptive regulation and control system which comprises a multichannel nerve-peripheral coupling module, an emotion intensity probability mapping module and a closed-loop nerve regulation and control current module. The multi-channel nerve-peripheral coupling module is used for realizing overall quantification of central and peripheral emotional physiology; the emotion intensity probability mapping module is used for generating continuous emotion probabilities ranging from 0 to 1 through normalization and nonlinear mapping by utilizing emotion energy and combining eye movement fatigue and electroencephalogram entropy; and the closed-loop nerve regulation and control current module is used for dynamically adjusting the transcranial electrical stimulation intensity within the safety current upper limit according to the difference value between the emotion probability and the expected target. According to the invention, the recognition precision, the response speed and the use comfort are obviously improved.

The invention provides a code dynamic completion method for a large language model. The method comprises the following steps: constructing a mixed training data set; the method comprises the following steps of: initializing a Transform model architecture fused with probability space topological transformation; based on the mixed training data set, executing model training of collaborative optimization of the main target and the auxiliary target; in each step of decoding, the model calculates the probability distribution of the next token according to the current context; the generated probability distribution is restrained in an effective probability subspace through spherical probability mapping, and grammar error candidates are filtered in real time and the probability distribution is adjusted in combination with a tabu table mechanism; and dynamically selecting a sampling strategy according to the current decoding depth, sampling from the adjusted probability distribution to obtain a next code snippet, and finally generating a code completion suggestion conforming to abstract syntax tree rules and semantic constraints. According to the method, the probability distribution constraint and the taboo table mechanism are combined, the model can dynamically adjust the generated probability distribution, the more appropriate candidate token can obtain the higher probability, and therefore the accuracy of code completion is improved.

The invention relates to the technical field of electric power market price prediction, in particular to an electric power spot price prediction method and system based on similar day probability correction, and the method comprises the following steps: S1, electric power data collection: obtaining historical market information of a day-ahead spot market and day-ahead electric power spot price data; s2, feature engineering construction: performing feature engineering construction on the collected power data; s3, data preprocessing: forming a data set; s4, performing model training: performing training on the LightGBM model; s5, similar day matching: screening k historical days with the highest similarity with the prediction day; s6, performing interval probability distribution modeling of the bidding space and the day-ahead electricity price: establishing a probability mapping matrix of the bidding space interval and the day-ahead electricity price interval; and S7, prediction result correction and final electricity price prediction: generating a final day-ahead spot electricity price prediction sequence. According to the invention, the method is closer to the actual market situation when the prediction logic is constructed, and the stability and adaptability of electricity price trend modeling are enhanced.

The invention provides a method and system for inverting metro train wheel out-of-smoothness through tunnel wall vibration source intensity data, and belongs to the technical field of accurate quantitative inversion of train wheel out-of-smoothness, and the method comprises the steps: obtaining actual measurement data of wheel out-of-smoothness, track irregularity and tunnel wall vibration source intensity; on the basis of the probability mapping relation between the out-of-roundness of the wheels and the source strength of the tunnel wall, actually measured source strength data of the tunnel wall are input into a model by utilizing a pre-trained probability machine learning model, so that probability statistics of the out-of-roundness of the train wheels is output, and accurate inversion of the out-of-roundness state of the wheels is realized. According to the method, random excitation of a wheel track system and a dynamic coupling mechanism of a vehicle-track-tunnel-stratum system are considered; meanwhile, wheel out-of-roundness probability statistics of a group of whole-line trains is quantitatively inverted by considering a probability mapping relation between wheel out-of-roundness excitation and tunnel wall source intensity response.

The invention provides a large-scale three-dimensional tunnel surrounding rock parameter random field model construction method and system, and relates to the technical field of tunnel surrounding rock modeling, and the method comprises the steps: carrying out the fitting construction of a probability distribution model reflecting surrounding rock parameters; selecting a plurality of surrounding rock parameters as clustering features, quantitatively classifying global geologic structures in tunnel surrounding rocks, and performing global decomposition on a tunnel model into continuous sub-domains containing overlapped buffer areas; for a first sub-domain from any direction of the whole domain of the tunnel model, performing covariance matrix spectral decomposition by adopting a KL decomposition method to generate a first sub-domain random field, and extracting random field values of overlapped regions among the sub-domains as boundary condition data of a recursive generation process; on the basis of the boundary condition data, conditional random fields of subsequent sub-domains are generated layer by layer; and seamlessly splicing the conditional random fields generated by recursion of the sub-fields into a global parameter random field, converting the global parameter random field into probability distribution by adopting a probability mapping method, and performing instruction encapsulation by developing a cross-platform data interface engine to realize a construction process of a parameter random field model.

The invention discloses a welding quality on-line detection method based on machine vision, and particularly relates to the field of welding quality detection.The welding quality on-line detection method comprises the steps that multi-dimensional basic data collection is conducted, the two-dimensional form, heat distribution, micro texture, three-dimensional topological data and welding process parameters of a welding seam are obtained, and a full-dimensional original data set with time-space synchronization is constructed; performing feature decoupling on the data, respectively constructing geometric morphology, thermal, texture and three-dimensional topological feature functions, and converting unstructured data into structured feature values; integrating multi-dimensional features through a nonlinear deep fusion model, introducing thermal probability mapping, texture attention weighting and a three-dimensional residual compensation mechanism, and outputting a weld comprehensive quality value; and finally, positioning dominant defect dimensions and identifying defect types in combination with a dynamic threshold model to finish quality grading judgment. Multi-dimensional accurate detection and dynamic grading of the welding quality are achieved, different scene requirements are met, and the detection efficiency and the judgment accuracy are improved.

The invention discloses an automobile casting multi-equipment detection data fusion judgment method, relates to the technical field of intelligent detection, and solves the problems that in an existing detection method, confidence output of multi-modal data is incomparable, misinformation is easily caused by evidence conflicts, and positioning offset is generated due to the fact that multi-modal registration is limited by a rigid model. According to the technical scheme, the method comprises the steps that surface images, point cloud and X-ray projection data are preprocessed, a detection response matrix is established, probability mapping is conducted on each modal score to form likelihood distribution and uncertainty representation, weighted fusion is completed in combination with a redundancy coefficient matrix, and a conflict judgment mark is generated; on the basis, hierarchical registration is executed, and a final judgment and recheck label is output; based on the technical scheme, the judgment accuracy and stability of multi-source detection data fusion of the automobile casting are remarkably improved.

The invention discloses a deep learning-driven inspection result automatic auditing and abnormity early warning system and method. The system is composed of a data acquisition and preprocessing module, a feature extraction module, a knowledge graph module, an abnormity simulation and enhancement module, a deep learning auditing module, a dynamic abnormity scoring and probability mapping module, an early warning and feedback module and the like. According to the system, through deep coupling of a knowledge graph in feature weight, anomaly simulation and scoring calculation, cooperation of rule knowledge and a data-driven model is achieved. In the training stage, an abnormal sample expansion training set conforming to physical constraints is generated by utilizing the generative model, and joint training is carried out in the multi-modal deep learning model. In the operation stage, a dynamic anomaly scoring formula based on a Gaussian kernel and a micromappable function are adopted to calculate the anomaly probability, and early warning is carried out according to the probability. According to the method, the accuracy, robustness and interpretability of examination result auditing are remarkably improved through multi-module cooperation and a simulation generation technology.

The invention discloses an underwater Internet of Things destroy-resistant link optimization method combining self-guided graph representation and AMF. The method comprises the following steps: 1) generating a scale-free topology and a candidate link set according to node positions, residual energy and underwater acoustic channel quality; 2) inputting the topology into a self-guided graph representation learning model, and applying contrast constraint through unsupervised training to obtain robust node embedding representation; 3) on the basis of candidate link two-end embedding, combining node degree, betweenness centrality, bridge edge coefficient and link success rate to construct feature representation, and generating an existence probability matrix through a probability mapping function; 4) inputting the matrix into adaptive multi-layer filtering (AMF), reserving edges according to Top-k to form a connected skeleton, setting a hierarchical threshold according to node hierarchy and load, and maintaining approximate power law degree distribution; and 5) applying a global connectivity constraint, and outputting an optimized link structure. According to the method, link optimization can be realized without supervision labels, network connectivity and survivability are improved in an attack environment, and the method is suitable for ocean monitoring and security scenes.

The invention provides a grounding nuclear explosion damage range calculation method. The method comprises the following steps: decomposing nuclear explosion energy into four main components: shock wave energy flow, thermal radiation energy flow, seismic wave energy flow and electromagnetic pulse energy flow; a heterogeneous medium model containing an atmosphere model and a geologic model is constructed, terrain and geologic data are fused to generate a joint map, and spatial correlation of the terrain and the stratum is described. Shock wave and thermal radiation propagation is calculated through a nonlinear wave equation and a dynamic atmospheric transmittance model, and the influence of seismic waves and electromagnetic pulses is evaluated. And establishing a damage probability mapping relation by adopting a fuzzy logic algorithm, drawing a nuclear explosion damage range graph, and performing dynamic updating through a Bayesian learning module. According to the invention, through multi-physics field coupling analysis and a dynamic learning mechanism, the accuracy of nuclear explosion damage assessment is improved.

The invention discloses a fracture network three-dimensional reconstruction and connectivity evaluation method and system based on Monte Carlo simulation, and belongs to the technical field of geotechnical engineering geological exploration and digital rock mass modeling. The method comprises the following steps: firstly, acquiring a two-dimensional fracture image of a drill hole, extracting geometric parameters, and acquiring three-dimensional plane initial parameters through a dimension conversion model; introducing Monte Carlo simulation, extending the deterministic plane model into a probability distribution model, and generating a large number of plane instances; calculating statistical characteristics of distances and included angles between fracture planes, and constructing a connectivity probability mapping function to evaluate a comprehensive connectivity probability; and finally, generating a global uncertainty cloud picture based on analysis results of all fracture pairs to identify a high-uncertainty region, and proposing an optimization and supplementation drilling scheme according to the high-uncertainty region. According to the method, the fracture network is transformed from deterministic reconstruction to probabilistic evaluation, connectivity can be quantified, uncertainty can be visualized, and a scientific basis is provided for engineering exploration decision making.

The invention discloses a gallium oxide crystal storage environment monitoring system and method based on artificial intelligence, and relates to the technical field of intelligent industrial monitoring, and the method comprises the steps: inputting a multi-source coupling data set into a space-time hypergraph model, carrying out the multi-source data coupling of a hypergraph construction layer, and carrying out the spatial dependence capture and time sequence feature extraction of a space-time convolution layer, the method comprises the steps of forming an environment parameter prediction matrix, performing damage quantification and probability mapping on a gallium oxide crystal damage image, obtaining a reference damage distribution diagram, performing coupling association on the environment prediction matrix and the reference damage distribution diagram, outputting an environment-damage weight matrix, and performing defect positioning on the reference damage distribution diagram according to the environment-damage weight matrix. And obtaining a damage sensitive area. According to the invention, through the space-time hypergraph model, the DBSCAN clustering analysis and the risk grade division strategy, the accuracy and the real-time performance of the monitoring scheme are enhanced.

The invention provides a prediction scene scoring method, electronic equipment and a storage medium, and relates to the technical field of intelligent driving, and the method comprises the steps: obtaining a trajectory prediction result, carrying out the shape remodeling of the trajectory prediction result, and obtaining input data, the input data comprises multi-dimensional trajectory data, modal data and dynamic agent time step joint data; extracting features of the input data to obtain a feature map; performing feature pooling on the feature map to obtain a pooling result; and performing score mapping on the pooling result to obtain a modal probability score which is used for improving the accuracy of self-vehicle scene and probability mapping.

The invention provides a power distribution network early fault section identification method and system based on multi-dimensional feature probability mapping, and belongs to the technical field of power system power distribution network fault detection. The method comprises the following steps: collecting zero-sequence voltage of a power distribution network bus and zero-sequence current signals of head ends of all feeder lines, and obtaining derivative signals of the zero-sequence voltage and the zero-sequence current signals; calculating the third harmonic energy of the zero-sequence current of each feeder line, and when the third harmonic energy exceeds a set threshold value, judging that the feeder line has an early fault; processing the zero-sequence current signal, and extracting an IMF1 component; calculating a correlation characteristic quantity and a dot product characteristic quantity for each section formed by the current difference values of the adjacent measuring devices based on the IMF1 component; and respectively carrying out standardized transformation on the correlation characteristic quantity and the dot product characteristic quantity, substituting the correlation characteristic quantity and the dot product characteristic quantity into a Laplacian cumulative distribution function, calculating a probability value of a fault in each section, and judging a fault section according to a set probability threshold value. The method can effectively and reliably identify the early fault section of the power distribution network.

The invention discloses a generative model and digital twinborn fusion aerospace equipment structure reliability test method, which comprises the following steps of: firstly, establishing a physical entity model of aerospace equipment, installing a sensor at a key position of the physical entity model, and acquiring stress, strain and vibration data of the equipment in real time; secondly, establishing a digital twinborn model corresponding to a physical entity, and determining a load-response relationship of equipment; processing the actually acquired data to form a standardized load-response data set; establishing and training a probability mapping relation between load and response by utilizing a diffusion generation model, and realizing the generation of equipment response data; then defining a limit state function of reliability analysis of the equipment structure, and carrying out reliability simulation evaluation to obtain a failure probability and a reliability index of the equipment structure; and finally, acquiring equipment operation response data in real time, and dynamically updating parameters of the diffusion generation model and the digital twin model. According to the invention, the accuracy and efficiency of aerospace equipment structure reliability testing are effectively improved.

A method, apparatus, and computer-readable storage medium providing data processing on at least one graphics processing unit (GPU). The method includes: acquiring sparse features of a target sample comprising first and second sparse features of different types; acquiring embeddings corresponding to first sparse features from stored embeddings of full sparse features of the first type; acquiring embeddings corresponding to second sparse features from a second GPU based on querying embeddings of full sparse features of the second type; forming embeddings corresponding to the sparse features of the target sample; performing probability mapping on the formed embeddings; generating an update instruction based on the probability mapping; and updating the embeddings of the full sparse features of the second type.

The invention discloses a health assessment method, device and equipment for a rotating part of a vehicle running gear, and a medium, and relates to the technical field of vehicle health management, and the method comprises the steps: obtaining a target sensitive fault feature of a to-be-assessed rotating part in the vehicle running gear in a current time period; wherein the target sensitive fault feature is a feature associated with degradation of the rotating part to be evaluated; processing the spliced target sensitive fault features by using a target state level probability mapping model to obtain target probabilities of the to-be-evaluated rotating part under each degradation state level; wherein the target state level probability mapping model is a deep neural network model constructed based on a deep belief network; and quantifying the degradation state of the to-be-evaluated rotating part according to each target probability to obtain a health index of the to-be-evaluated rotating part. According to the scheme, the degradation state of the rotating part of the vehicle running gear can be accurately evaluated.

The invention relates to a long straight control cable space radiation field effect calculation and evaluation method, system and device and a medium. The method comprises the following steps: constructing a partial differential equation constraint model based on a cable geometric structure and electromagnetic transient characteristics; fusing an actually measured magnetic field and an additional physical rule through a physical information neural network, and generating a complete electromagnetic data set to compensate for electric field data missing; a multi-dimensional interference factor is calculated by combining time-frequency distribution characteristics, equipment sensitive parameters and cable equivalent inductance, and the radiation field effect complexity is objectively quantified; establishing a probability mapping model of interference indexes and effect states by adopting multi-dimensional Gaussian process regression to realize risk probability distribution prediction; and finally, generating a visual probability curved surface of an electric field-magnetic field plane, and identifying a high-risk area to guide electromagnetic protection optimization. According to the method, the evaluation precision is improved in a data limited scene, and a technical closed loop from multi-dimensional feature fusion to probabilistic risk early warning is realized.

The invention discloses a big data security access control method. The method comprises the following steps: collecting multi-dimensional information of an access request, including time, place, equipment, network environment and historical behaviors, and generating a risk score in combination with a weight; and converting the risk score into a risk probability by using a probability mapping model, and dividing risk levels according to a preset threshold. And the system forms a dynamically adjustable permission set according to the basic permission of the user role and the limit permission corresponding to the risk level, and calculates a security strength index. And further combining the risk score with the security strength to generate a comprehensive risk coefficient and recording the comprehensive risk coefficient in an audit log. And when the risk coefficient exceeds a threshold value and exceeds the limit continuously for multiple times, triggering a security response mechanism, including account freezing, session blocking, alarm pushing and global strategy tightening. According to the method, real-time evaluation, fine-grained control and linkage response of big data access are realized, and the dynamic defense capability and the overall security of the system are improved.

The invention discloses an interpretable lightweight automatic driving lane changing decision method and system based on fuzzy structure adaptive configuration, and relates to the technical field of automatic driving lane changing decision, and the method comprises the steps: obtaining a six-dimensional state of a unified aperture, building an adaptive fuzzy structure for each dimension, automatically determining the precondition division according to the data distribution, and carrying out the self-adaptive configuration of a fuzzy structure. Underfitting or redundancy caused by manual setting is avoided; rules are activated and normalized, and a first-order TS linear consequent is matched to form traceable segmented approximation for a continuous state space, so that a lane changing probability source is clear, and an explanation path is clear; end-to-end learning is carried out through probability mapping and a consistent training target, so that the judgment accuracy and the output stability are improved; after training is completed, based on joint evaluation of rule activation and category deviation, only effective rules are reserved for online reasoning, the calculation scale and memory occupation are remarkably compressed, and therefore a decision link is controlled within vehicle-mounted time delay and resource constraint.

The invention relates to the technical field of data processing, and relates to an intelligent pre-declaration risk management and control method for cross-border e-commerce customs clearance. Through the multi-dimensional data dynamic verification and automatic declaration mechanism, real-time compliance verification of customs clearance pre-declaration data is realized, the rejection rate caused by data errors is reduced, and the customs clearance period is shortened; by constructing a multi-dimensional risk scoring model of commodity compliance, transportation abnormality, document completeness and flow tax compliance and combining a risk score-rejection probability mapping table, the rejection probability of a commodity to which a target cross-border e-commerce belongs is determined, the customs clearance risk prediction accuracy is improved, and high-risk declaration can be avoided in advance; by automatically triggering data correction and resubmission for high-risk declaration and quickly releasing low-risk declaration, intelligent allocation of customs clearance resources is realized, and reduction of enterprise declaration labor cost is facilitated.

The invention relates to the technical field of artificial intelligence medical aid decision making, and particularly discloses a sepsis risk prediction method based on a health index. The method comprises the following steps: collecting health index data to construct a standardized feature vector, and training a cumulative health index scoring model; constructing an ordered logistic regression model based on the health index, and outputting sepsis grade probability distribution; introducing a teacher-student model to carry out probability transfer learning to obtain a first sepsis risk prediction result; and explaining and analyzing by adopting an SHAP method, and outputting a second prediction result. Compared with the technical problems that in the prior art, sensitivity is low mainly depending on a regular scoring system such as SIRS or SOFA, and particularly risk early warning is difficult to achieve under the condition that early clinical indexes are slightly abnormal, due to the fact that accumulative scoring modeling and ordered logic probability mapping are introduced, sepsis risk level continuous early warning and explanation are achieved, and the risk early warning efficiency is improved. And the early sepsis risk prediction effect is improved.

The application provides a new view synthesis method based on Gaussian probability distribution and feature regularization, relates to the technical field of new view synthesis, and comprises the following steps: extracting a feature tensor from a preprocessed input image, performing Gaussian probability calculation based on the mean and standard deviation of the feature tensor; introducing an elastic net regularization loss to constrain the feature tensor, performing mean calculation on the Gaussian probability of the feature tensor, substituting the calculation result into a loss function, minimizing the loss function, obtaining an optimized Gaussian point set, performing multi-scale densification and pruning, and generating a new view. The application solves the technical problem that, due to overfitting in a sparse scene, the number of Gaussian points is too small and the transparency is low, which further affects the accuracy of view synthesis, and improves the modeling effect and the accuracy of view synthesis in a sparse view angle by introducing Gaussian probability mapping and feature regularization, thereby increasing efficiency and improving precision.

The application discloses a mountainous city park CES space probability mapping method considering population differences, and steps include: constructing a space feature classification system; dividing population categories according to population difference characteristics and obtaining space perception data; generating CES value space vector data and frequency data through data cleaning; generating environment layer grid data set and park boundary; performing space density interpolation to output CES value density distribution map; training and verifying using a space probability prediction model, performing probability prediction taking population categories as grouping variables, converting and standardizing mapping probability data and maximum value index to generate a CES space distribution map considering population differences; and statistically analyzing to generate a frequency feature map, a correlation diagram and a flow direction diagram. The technical scheme realizes precise mapping of CES space distribution of population differences, reveals the correlation mechanism of landscape elements and CES value, and provides scientific support for mountainous park renewal design.

This application relates to a training method and apparatus for improving the information acquisition capability of an artificial intelligence model. The method includes: determining the probability of performing different semantic enhancement modes on any first training sample based at least on first training samples and processing parameters corresponding to the artificial intelligence model; performing random probability mapping on at least one first training sample according to the probabilities corresponding to each semantic enhancement mode; determining the semantic enhancement mode corresponding to each first training sample based on the random probability mapping results; performing corresponding reconstruction processing on at least a portion of each first training sample according to the semantic enhancement mode corresponding to each first training sample to obtain second training samples; and training the artificial intelligence model based on at least one second training sample. This method can improve the model's ability to output accurate answers.

The invention relates to the technical field of man-machine interaction and intelligent control, in particular to a flexible somatosensory interaction system and method based on intention probability mapping. Through setting a flexible somatosensory acquisition module, an intention modeling module, a flexible-rigid mapping field construction module, a dynamic intention screening module and an execution control module, multi-modal somatosensory information of an operator is acquired, a probability distribution model of an action intention is established, and a flexible-rigid mapping field is constructed based on an information retention and minimum semantic loss principle. And the continuous action space of the human body is converted into a limited control instruction set which can be executed by the mechanical arm. The system dynamically screens an optimal control instruction by using a confidence coefficient driving mechanism under multiple candidate intentions, and realizes accurate response of a mechanical arm to the action of an operator through closed-loop feedback. The problems that in the prior art, flexible action and rigid control freedom degrees are not matched, and intention recognition is fuzzy are effectively solved.

The application discloses a complex system coupling coordination degree optimization method and device, and the method comprises the following steps: acquiring a multilevel coupling parameter set in a complex system; screening a to-be-optimized parameter variable from the multilevel coupling parameter set to obtain a to-be-optimized parameter variable set; constructing a Bayesian statistical model; taking the Bayesian statistical model as a proxy model for describing a probability mapping relationship between a target function and a parameter variable; determining a sampling function; and iteratively optimizing the to-be-optimized parameter variable by using the Bayesian statistical model and the sampling function to obtain an optimized parameter. The application can improve the coupling coordination degree of the complex system and meet the requirements of stable operation and intelligent control of the complex system.

The application discloses a low-altitude airspace operation supervision method and system for highway maintenance inspection, and relates to the field of traffic road maintenance inspection control.The application iteratively clusters and merges new and old crack influencing factor data, continuously optimizes mapping model parameters, and makes crack probability prediction more in line with actual road conditions.A crack probability model and an accident level model form a "prevention-emergency" double link, the former predicts crack risks, and the latter controls vehicle speed in real time, covering the whole maintenance cycle.According to crack probability mapping results, the height of unmanned aerial vehicle (UAV) inspection is automatically adjusted, high-risk road sections are preferentially investigated, and inspection efficiency is improved.The accident level model fuses vehicle dynamic speed and crack static characteristics to generate graded alarm instructions, avoiding single factor misjudgment.According to the contribution degree of influencing factors, the crack source is repaired in a targeted manner, and invalid construction is reduced.Through accurate inspection and repair priority sorting, the flight time of the UAV and the frequency of manual inspection are reduced, and maintenance costs are reduced.